Van Eyck Studies: Papers Presented at the Eighteenth Symposium for the Study of Underdrawing and Technology in Painting, Brussels, 19-21 September 2012 9042938595, 9789042938595, 9789042938601

Table of contents :
Cover
Van Eyck Studies: Papers Presented at the Eighteenth Symposium for the Study of Underdrawing and Technology in Painting, Brussels, 19-21 September 2012
Copyright
Contents
Editors’ Preface
Notes to the Reader
Abbreviations
Bibliography

Citation preview

VAN EYCK STUDIES Papers Presented at the Eighteenth Symposium for the Study of Underdrawing and Technology in Painting

PEETERS

VAN EYCK STUDIES

UNDERDRAWING AND TECHNOLOGY IN PAINTING SYMPOSIUM XVIII

The XVIII Symposium for the Study of Underdrawing and Technology in Painting was organized by

The Royal Institute for Cultural Heritage, Centre for the Study of the Flemish Primitives, Brussels In collaboration with Illuminare – Centre for the Study of Medieval Art, KU Leuven – University of Leuven Université catholique de Louvain-la-Neuve, Laboratoire d’étude des œuvres d’art par les methodes scientifiques

Scientific Committee Christina Ceulemans (KIK-IRPA General Director a.i.), Jacqueline Couvert (UCL), Christina Currie (KIK-IRPA), Anne Dubois (UCL, FRS-FNRS), Bart Fransen (KIK-IRPA), Ron Spronk (Queen’s University, Canada-Radboud University, Holland), Cyriel Stroo (KIK-IRPA), Jan Van der Stock (Professor, KU Leuven), Dominique Vanwijnsberghe (KIK-IRPA), Hélène Verougstraete (Emeritus Professor), Annelies Vogels (KU Leuven), Lieve Watteeuw (KU Leuven)

VAN EYCK STUDIES Papers Presented at the Eighteenth Symposium for the Study of Underdrawing and Technology in Painting, Brussels, 19-21 September 2012

Edited by Christina Currie, Bart Fransen, Valentine Henderiks, Cyriel Stroo, Dominique Vanwijnsberghe

PEETERS PARIS – LEUVEN – BRISTOL, CT

2019

A catalogue record for this book is available from the Library of Congress. © Peeters Leuven, Bondgenotenlaan 153, B-3000 Leuven Revision and editing: Lee Preedy All rights reserved. No part of this book may be reproduced or translated in any form or by any means, by print, photoprint, microfilm, microfiche or any other means without written permission from the Publisher. ISBN: 978-90-429-3859-5 eISBN 978-90-429-3860-1 D/2019/0602/2 This book was published in hardcover in 2017. front cover: Jan and Hubert van Eyck, detail from the Virgin Anunciate panel of the Ghent Altarpiece, 1432, Ghent, Saint Bavo’s Cathedral. © Lukasweb.be-Art in Flanders vzw, photo kik-irpa

Contents Editors’ Preface

IX

Notes to the Reader

XI

Abbreviations

XIII

PART I. THE GHENT ALTARPIECE 1 2

3 4

5

6

7

8

9 10

The Ghent Altarpiece Revisited: 2012-2017 Anne van Grevenstein-Kruse and Hélène Dubois

3

Gems in the Water of Paradise. The Iconography and Reception of Heavenly Stones in the Ghent Altarpiece Marjolijn Bol

35

The Adoration of the Lamb. Philip the Good and Van Eyck’s Just Judges Luc Dequeker

51

‘Revenons à notre Mouton’. Paul Coremans, Erwin Panofsky, Martin Davies and the Mystic Lamb Hélène Dubois, Jana Sanyova and Dominique Vanwijnsberghe

67

Results of Three Campaigns of Dendrochronological Analysis on the Ghent Altarpiece (1986-2013) Pascale Fraiture

77

Research into the Structural Condition and Insights as to the Original Appearance of the Panels and Frames of the Ghent Altarpiece Aline Genbrugge and Jessica Roeders

97

Small Hairs. Meaning and Material of a Multiple Detail in the Ghent Altarpiece’s Adam and Eve Panels Ann-Sophie Lehmann

107

Le rôle du dessin sous-jacent et de l’ébauche préparatoire au lavis dans la genèse des peintures de l’Agneau Mystique Catheline Périer-D’Ieteren

121

Art and Compensation. Joos Vijd and the Programme of the Ghent Altarpiece Bernhard Ridderbos

137

John the Baptist and the Book of Isaiah in the Ghent Altarpiece Patricia Stirnemann, Claudine A. Chavannes-Mazel and Henry Dwarswaard

145

11

12

La présentation de l’Agneau Mystique dans la chapelle Vijd. Le rapprochement progressif de deux retables Hélène Verougstraete

157

The Frames by Schinkel for the Wings of the Ghent Altarpiece and the Copies in Berlin Bettina von Roenne

179

PART II. STUDIES IN VAN EYCK PAINTINGS 13 14 15

15

16 17 18 19

20 21 22 23 24

Van Eyck’s Technique and Materials: Historical Perspectives and Contemporary Context Marika Spring and Rachel Morrison

195

Revelations Regarding the Crucifixion and Last Judgment by Jan van Eyck and Workshop Maryan W. Ainsworth

221

Remarks on Character and Functions in Jan van Eyck’s Underdrawing of Portraits: the Case of Margaret van Eyck Part 1 Rachel Billinge

233

Jan van Eyck’s Underdrawing of Portraits Part 2 Till-Holger Borchert

241

The Speed of Illusion Lorne Campbell

257

Les autoportraits présumés de Jan Van Eyck et la date approximative de sa naissance Pierre Colman

263

Pigments, Media and Varnish Layers on the Portrait of Margaret van Eyck Jill Dunkerton, Rachel Morrison and Ashok Roy

271

New Findings on the Painting Medium of the Washington Annunciation Melanie Gifford, John K. Delaney, Suzanne Quillen Lomax, Rachel Morrison and Marika Spring

281

Jan van Eyck’s Greek, Hebrew and Trilingual Inscriptions Susan Frances Jones

291

Early Texts on Some Portraits by Jan van Eyck Stephan Kemperdick

311

Jan van Eyck and his ‘Stereoscopic’ Approach to Painting Renzo Leonardi

327

The Development Process of the Dresden Triptych. News and Questions Uta Neidhardt

351

Voyager dans les tableaux de Van Eyck Jacques Paviot

367

25

26 27 28

Replications of Exemplary Form. New Evidence on Jan van Eyck’s St Francis Receiving the Stigmata Jamie L. Smith

375

Questioning the Technical Paradigm of the Ghent Altarpiece Noëlle L.W. Streeton

389

Gold-brocaded Velvets in Paintings by Jan van Eyck. Observations on Painting Technique Esther E. van Duijn

403

Surface Effects in Paintings by Jan van Eyck Abbie Vandivere

417

PART III. VAN EYCK IN CONTEXT: OTHER MEDIA, RECEPTION AND LEGACY 29

30

31 32 33 34

35 36

37

Mural Paintings before Jan van Eyck. A Remarkable Discovery from around 1400 in St John’s Church in Mechelen Marjan Buyle and Anna Bergmans

437

The Fishing Party by Jan van Eyck (?). A Technical Analysis Claudine A. Chavannes-Mazel, with Matthias Alfeld, Koen Janssens, Geert Van der Snickt, Peter Klein, Micha Leeflang, Margreet Wolters, Carel Van Tuyll van Serooskerken and André le Prat

455

Van Eyck in Valencia Bart Fransen

469

Jan van Eyck’s Genoese Commissions. The Lost Triptych of Battista Lomellini Maria Clelia Galassi

481

Jan van Eyck, Polychromer (and Designer?) of Statues Ingrid Geelen

495

Les copies de la Vierge à l’Enfant dans une église de Jan van Eyck et le rôle de la version dessinée du Grand Curtius Valentine Henderiks

509

Jan Van Eyck as Illuminator? Hand G of the Turin-Milan Hours Catherine Reynolds

519

The Tomb-Slab of Hubert Van Eyck and Gerard Horenbout. A Tribute to the Great Ghent Master Ronald Van Belle

535

A Note on Gold and Silver in a Metalpoint Drawing by Jan van Eyck Arie Wallert

547

Bibliography Illustration Credits

557 597

Editors’ Preface The work of the Van Eyck brothers, Hubert and Jan, never fails to arouse interest among art historians. If the diversity of topics and aspects addressed in Symposium XVIII for the Study of Underdrawing and Technology in Painting is any guide, the fascination with Eyckian art, with all its dazzling illusionistic effects and iconographic finesse, is every bit as fresh, challenging and current as it was when that art was created six centuries ago. Eminent specialists from Belgium and beyond have shed light on a highly complex matter, evaluating observations made over the last decades, not least in the field of art-historical, conservatorial and technical research. To fathom Van Eyck’s ‘secrets’ is the ultimate aim of such an event. And that aim is achieved, albeit incrementally and with increasing input from modern imaging techniques and scientific analysis. A scientific gathering of this undoubted calibre does not, however, result in a handy guide to Van Eyck, bursting with ready answers. In fact the number of problems has barely decreased. But the wealth of insight has grown, the surprise and amazement has been augmented. It hardly needs saying that solutions are not found by focusing on a single research discipline. We reach consensus only when knowledge and insights from various disciplines are brought to the table and thoughts are exchanged – the very aim of this colloquium. Of course, this kind of interdisciplinary approach has existed for many decades and in that respect the kik-irpa has been a trailblazer from day one. But nowadays so many different disciplines are involved in Eyckian research that one might question whether it is even possible to process all the knowledge we have. On the other hand, extreme specialization comes with the risk that other scientific and scholarly perspectives are lost sight of. If we do not want to lose our way in a vast forest in which no one can see the genuine Eyckian wood for the trees any longer, and if we hope to draw the occasional general conclusion that goes beyond the study of one small specific part, we need to take a couple of precautions. Firstly the various approaches should complement and enrich each other rather than following the divaricating trajectories that extreme specialization tends to: in other words they should be centripetal rather than centrifugal. Those who sit with their noses practically touching a work in order to study its details – the tiny iconographic motifs, micro-historical problems, the molecular structure of the materials, preparatory drawings invisible to the naked eye, minute brush strokes and so forth – would they not find it useful sometimes to stand back a little to look at the paintings in the context of their creation and use? And those who prefer the overview, the history of art with a capital H, would they not be intrigued if they zoomed in now and then, getting ‘closer to Van Eyck’ and giving more attention to the painting as object, man-made with all manner of materials and extraordinary expertise. To offer a platform for discussion, where researchers from different backgrounds with interests and methods that are sometimes hugely diverse can communicate and consult and understand each other, was one of the objectives of this conference. That objective, we feel, has been more than achieved. Secondly, we should encourage the view that works of art are not our property, to be claimed for our own research, but universal cultural heritage that deserves our collective care and study. That the Ghent Altarpiece is for everyone and that science benefits from sharing the new documentation with the world via the Internet was the basic principle behind the construction of the ‘Closer to Van Eyck’ website (http://closertovaneyck.kikirpa.be). For the kik-irpa too this project has contributed to a change in attitude. The open access policy changed the way in which documentation and scientific discoveries were or

X

editor ’ s foreword

were not shared or protected. It was evident from the PowerPoint images shown during the colloquium that the website has proved its worth and we hope that the open access policy will remain and be applied to many future projects. Finally, we would like to thank all our colleagues, partners, speakers and the 256 participants for the great success of the intensive three-day symposium in Brussels in 2012 and for the very valuable contributions to this publication, which with its thirty-seven articles and innumerable illustrations has taken quite some preparation. For their help in the course of that preparation we also thank Bernard Petit, for image processing, and Elisabeth Van Eyck, Julie Van Woensel, Alistair Watkins and Lee Preedy for editing and copy-editing. We wish you all pleasant reading.

The Editorial Team Christina Currie, Bart Fransen, Valentine Henderiks, Cyriel Stroo and Dominique Vanwijnsberghe

Notes to the Reader The Adoration of the Lamb or ‘Ghent Altarpiece’ is a polyptych. In this book the altarpiece as a whole is referred to as the Ghent Altarpiece, in italics; the individual panels are referred to by their descriptive titles (Virgin Enthroned, Singing Angels, Hermits, and so on) in ordinary type. Illustrations are numbered according to chapter: fig. 1.21, fig. 3.2, fig. 16.5, fig. 35.7, and so on. Among the illustrations are numerous details of the Ghent Altarpiece. To avoid repetition in the captions, the altarpiece’s ‘technical data’ appear only in the captions to Plates A and B. ‘Left’ and ‘right’ are from the point of view of the viewer. Dimensions are given in centimetres or millimetres, first height, then width. There are many references in this book to the outstanding ‘Closer to Van Eyck’ website. It can be accessed at http://closertovaneyck.kikirpa.be

Abbreviations belspo

Belgian Federal Science Policy

BnF

Bibliothèque nationale de France

charisma

Cultural Heritage Advanced Research Infrastructures. Synergy for a Multidisciplinary Approach to Conservation-Restoration

kadoc

Documentatie- en Onderzoekscentrum voor Religie, Cultuur en Samenleving | Documentation and Research Centre for Religion, Culture and Society, Leuven

kbr

Koninklijke Bibliotheek van België | Bibliothèque royale de Belgique | Royal Library of Belgium, Brussels

kik-irpa

Koninklijk Instituut voor het Kunstpatrimonium | Institut royal du Patrimoine artistique | Royal Institute for Cultural Heritage

kmkg-mrah

Koninklijke Musea voor Kunst en Geschiedenis | Musées royaux d’Art et d’Histoire | Royal Museums of Art and History

kmska

Koninklijk Museum voor Schone Kunsten Antwerpen | Royal Museum of Fine Arts Antwerp

kmskb-mrbab

Koninklijke Musea voor Schone Kunsten van België | Musées royaux des BeauxArts de Belgique | Royal Museums of Fine Arts of Belgium

rkd

Rijksbureau voor Kunsthistorische Documentatie, Den Haag | The Netherlands Institute for Art History, The Hague

sam

Stadsarchief Mechelen

smb

Staatliche Museen zu Berlin

Plate A Hubert and Jan van Eyck, Ghent Altarpiece (before restoration): open 395 x 530 cm, oil on oak panel Ghent, St Bavo’s Cathedral

Plate B Hubert and Jan van Eyck, Ghent Altarpiece (after restoration): closed 395 x 265 cm, oil on oak panel Ghent, St Bavo’s Cathedral

PART I THE GHENT ALTARPIECE

Fig. 1.1 Hubert van Eyck’s tomb-slab, Ghent, St Bavo’s Abbey, Museum of Stone Objects

1

The Ghent Altarpiece Revisited: 2012-2017 Anne van Grevenstein-Kruse and Hélène Dubois

ABSTRACT: The urgent conservation and technical documentation of the Ghent Altarpiece in 2010 was a moment of scrutiny and evaluation of historical, aesthetic and technical issues that led to recommendations towards a future treatment. The conclusion of this intensive period of international and interdisciplinary collaboration was that the circumstances of the altarpiece’s presentation (climate control, location, safety), its wooden support and frames (conservation, function), and its paint layers and varnishes (structural stability, alterations, aesthetics) should all be reconsidered. Over a period of five years* those issues will be addressed by interdisciplinary research teams, including the conservators in charge of the treatment. During examination and conservationrestoration they will collect data and integrate those findings into their scientific context, bringing into action what Paul Philippot called le lien vivant entre le cerveau et la main – ‘the living link between the brain and the hand’. Interaction between research groups is triggered by questions such as ‘what are the parameters for climate control and presentation in the specific case of St Bavo’s Cathedral and the Ghent Altarpiece?’; ‘can new data be gathered on the genesis and the who-did-what question?’; ‘what is the impact of the history of interventions on the molecular structure of the original paint layers?’; and ‘how can the perception of authenticity and quality be enhanced?’ * Since this paper was written the time allotted to the conservation of the altarpiece has been extended to 2019.

—o— The memory of brothers Hubert and Jan Van Eyck is ever present in the romantic vision of the nineteenth century and in shadows of the past, transmitted by fragile materials and documents. Hubert’s tomb-slab,1 now in the Museum van Stenen Voor-

werpen (Museum of Stone Objects) housed in St Bavo’s Abbey in Ghent (fig. 1.1), could symbolize our search for the information that is crucial for art-historical interpretation and fundamental in the decision-making process in conservation and restoration. The white marble inlay of the corpse and the copper plate bearing texts to the painter’s memory are lost, but their traces – their shadows – remain.2 When the chemist Paul Coremans, director of the Central Laboratory of Belgian Museums, set out to lead a major treatment campaign of the altarpiece in 1950, he was well aware of the need to research its every facet. The project lasted just over a year and focused on the urgent conservation of the fragile paintings, which were suffering the consequences of the veritable odyssey they had made during the Second World War. Laboratory research on the techniques and materials used in their execution continued after the various panels were returned to St Bavo’s Cathedral and the results of the entire project were published in 1953 in L’Agneau Mystique au laboratoire, which would stand as a reference for interdisciplinarity in the conservation of works of art.3 In the introduction, Coremans made a pivotal and inspirational statement by associating the treatment with the collaboration between laboratory specialists, art historians and conservators that was necessary if the insight essential to endorsing the decisions and technical interventions of the restoration was to be gained. This concept is at the heart of the five-year

4

anne van grevenstein-kruse and hélène dubois

conservation and restoration project of the altarpiece that started in September 2012. Evolution Central to the collaboration in 1951 was Albert Philippot, a skilled and experienced restorer who worked alone on the treatment of the varnish and paint layers (figs 1.13, 1.24). A painter by training, he had learned restoration as a craft from his fatherin-law, Josef Van der Veken, who was also known by museum curators and art historians as ‘the Brussels forger’4 (fig. 1.2). Albert Philippot embodied the role of empathic translator of the old master’s original intent. He, the highly accomplished technician, understood the craft and knew by intuition how to, ‘penetrate art to restore the artefact’.5 Coremans oversaw the discussions and chose safety above all, balancing the focus on aesthetics with a pragmatic analytical approach while valuing the interpenetration of scientific method and intuitive knowledge.6 He was the ‘patron’, politely steering Philippot the craftsman. Since then, over three generations, the conservator’s status has evolved considerably. In a growing number of institutions around the world, conservators are now recognized as researchers in their own right. Thanks to the input of the different disciplines, skilled conservators develop a wellinformed insight into the historical and material nature of works of art that is also linked to their manual dexterity and intuition developed through experience. Their academic training provides a methodical and critical approach that was lacking in solely craft-based teaching and facilitates communication and collaboration with other specialists from different fields.7 Since the treatment in 1951, interdisciplinary research has indeed proved to be a fruitful approach to challenging assumptions. Paul Philippot, Albert’s son and the author of influential essays on conservation-restoration theory, ethics, and training, foresaw the danger of the various disciplines segregating rather than integrating around the conservation of the object.8 Though this tendency

is still latent, the interaction between art historians, scientists and conservators has evolved towards the development of new specializations merging the different fields: conservation-restoration, conservation science and technical art history9 have become academic disciplines in their own right, with their own, partly overlapping network of conferences and publications. The study of the Ghent Altarpiece is an iconic example of collaboration between the different fields. The Background to the Project In 2008, Monumentenwacht Vlaanderen, in association with the Royal Institute for Cultural Heritage (kik-irpa), tested the environment of the secured glass case in the Villa Chapel in which the altarpiece has been installed since 1986. The bulletproof case was conceived as a protection against theft and vandalism and was further protected with a concrete ceiling (fig. 1.3). Serious concerns about the climate and the conservation of the altarpiece triggered the formation of an advisory committee comprising representatives of the Churchwardens, the various administrations concerned with the conservation of the altarpiece and cathedral, and international experts.10 The committee instigated the organization of a detailed examination, conservation and documentation of the panels, which took place in the chapel itself between April and November 2010. The chapel was rendered inaccessible to the public who could, however – albeit not without frustration – view the specialists at work through a glass wall. This ‘live’ experience created public awareness of conservation, thereby stimulating the foundation of a local, national and international network for safety and care. The examination and conservation treatment were carried out by the kik-irpa and the Royal Museums of Fine Arts of Belgium (kmskb-mrbab), under the direction of Anne van Grevenstein.11 Ron Spronk led the technical documentation of the paintings, financially supported by the Getty’s Panel Paintings Initiative. The project, labelled ‘Lasting Support’,12 allowed for the training of three

the ghent altarpiece revisited: 2012-2017

Fig. 1.2 Josef Van der Veken puts the finishing touches to his copy of the Just Judges in 1951

5

6

anne van grevenstein-kruse and hélène dubois

Fig. 1.3 Section of the ‘bunker’ built in the Villa Chapel in 1986 to house the Ghent Altarpiece (Monumenten & Landschappen, 1986)

young panel-painting conservators, under the direction of Jean-Albert Glatigny, and the holding of several expert meetings in order to discuss their findings on the panels’ construction techniques.13 The results of this intensive campaign, which also involved examination of the frames and their polychromy by the kik-irpa and Ghent University and the dendrochronological examination of the centre panels by Pascale Fraiture (kik-irpa),14 has been freely accessible on the ‘closer to van eyck’ website (http://closertovaneyck.kikirpa.be) since November 2011. The campaign was also supported by the examination of the centre panels with noninvasive analytical techniques by the molab teams of the European charisma project (fig. 1.4).15 Climate Control and Urgent Conservation Measures The concerns about the climate in the Villa Chapel’s glass case triggered the formation of a working group on preventive conservation.16 Indeed, most modern museums are organized to provide constant climate monitoring and attention to preventive conservation, but for cultural heritage in churches this is not often the case.

Fig. 1.4 A MOLAB project researcher, Dr Claudia Daffara, monitoring the multispectral scanner during the examination of the Adoration of the Lamb panel

Continuity in follow-up depends greatly on the proximity and availability of personnel and this had been less than ideal in previous years. Subsequently, therefore, collaboration with Ghent University was sought so as to form a sound local research support group in the city, in close cooperation with Monumentenwacht Oost-Vlaanderen. International collaboration was stimulated by the Getty’s Panel Paintings Initiative, Klimaatnetwerk Vlaanderen, and Klimaatnetwerk Nederland.17 Until such time as fundamental structural changes can be carried out, the climate in the glass case has been improved by simple, temporary

the ghent altarpiece revisited: 2012-2017

7

measures. The constant relative humidity and temperature monitoring data are transferred to the working group and procedures have been put in place to ensure that immediate measures can be taken should the relative stability of the climate suddenly degrade. Presentation, Location, Safety, Public Access Not everyone approved of the 1986 relocation to the Villa Chapel (fig. 1.5).18 Indeed, the move away from the Vijd Chapel is still unacceptable to those who find the historical significance of the altarpiece has been diminished, even if the original appearance, dimensions and access of the chapel have been extensively altered through the centuries. Since the Ghent Altarpiece is listed as a monument ‘by destination’, which is to say irrevocably bound to the monument in which it resides, that being St Bavo’s Cathedral, the historic link between the architecture and the altarpiece is secured by the legislation on the protection of monuments. In this context the present location in the Villa Chapel is anachronistic. But so too would be all other solutions, other than the relocation in the Vijd Chapel, the ultimate lieu de mémoire, from its foundations to the space of the architecture and the light coming through the stained glass window: contextual and historic authenticity at its best. Van Eyck completed illusionistic effects and rendered light and shadow in accord with the original setting. Awkwardly, the altarpiece can appear too big for the space in the Vijd Chapel where, in its present configuration, it cannot be opened completely (fig. 1.6). But its original relationship to the architecture can also be interpreted in a different way if it is considered in its three-dimensionality. Embraced by the architecture, the altarpiece could be compared with wall paintings, with the halfopen wings as transcendental chapel walls, as TillHolger Borchert has suggested.19 The panels convey different meanings depending on their position. When the altarpiece is closed, we are shown the

Fig. 1.5 The Ghent Altarpiece installed in its glass case in the Villa Chapel, St Bavo’s Cathedral (Monumenten & Landschappen, 1986)

Fig. 1.6 Full-scale photographic reproduction of the Ghent Altarpiece in the Vijd Chapel, St Bavo’s Cathedral

donors, Joos Vijd and his wife Elisabeth Borluut, kneeling humbly in hope of the salvation implicit in the Annunciation depicted above them. When the altarpiece is open, it shows the promise fulfilled. Other forms of presentation have also been proposed: J.R.J. van Asperen de Boer’s suggestion of a half-open installation, based on observation of the original hinges and the frames on nineteenthcentury photographs and on the study of the underdrawing, was recently further explored by Albert Châtelet, who linked the iconography to the movement of the panels.20

8

anne van grevenstein-kruse and hélène dubois

The way the altarpiece was originally presented in the chapel is still the subject of debate. The installation in two clearly separated registers crowned by a canopy, proposed by Elisabeth Dhanens on the basis of archival research and traces in the walls of the chapel, remains a strong hypothesis (fig. 1.7).21 Evidence from the architecture, an archaeological approach to gather facts about the building history, will be re-examined in the coming years, during the restoration of the chapels around the choir. Over the past two years many different options for the post-conservation location of the altarpiece have been discussed with architects, churchwardens, art historians and colleagues responsible for managing the flow of tourists that has steadily increased since the Second World War. A series of interdisciplinary meetings, including experts from the Getty’s Panel Paintings Initiative and the preventive conservation working group, have showed that experts are divided by the varying emphasis given to the factors that have to be taken into account, such as the safety, climate, public access, monuments management and legislation or contextual authenticity of the location. Measured opinion based on research is the best way to reach a consensus. Museological Dimension of the Presentation The need to open the altarpiece completely and to be able to see the exterior of the wings in a new museum space with adequate climate control, safety, light and public access is incompatible with reinstallation in the Vijd Chapel. Of course, this argument is not recent: in 1894, Wilhelm von Bode’s decision to have the six wing-panels of the Ghent Altarpiece sawn through in order to display both sides of each panel on the museum walls was based on what was then a modern concept of museology: the public’s need to see things ‘properly’, at ease.22 Today, pressure from visual culture and the accessibility of bright colour reproductions might play an even greater role in discussions and decisions about the renewed display of the altarpiece after its restoration.

Fig. 1.7 Elisabeth Dhanens’s hypothetical reconstruction of the original installation of the Ghent Altarpiece (Gentse bijdragen tot de kunstgeschiedenis en oudheidkunde, vol. 22, 1969/72, p. 146)

The concept of an independent space to house the altarpiece was debated with architects and students of the faculty of architecture at Ghent University in 2011-2012 under the guidance of Professor Bart Verschaffel. As an academic exercise, the students freely designed fresh spaces for the altarpiece: they included a new building next to the cathedral, and using the episcopal palace behind the cathedral (this last concept involved politely

the ghent altarpiece revisited: 2012-2017

asking the bishop to move out!). The results of these exhilarating projects, la fantaisie au pouvoir, were shown in the Provinciaal cultuurcentrum Caermersklooster between September 2012 and June 2013. A future project, this time respecting the constraints of the law on monuments, will be shown in the same venue. Conservation Treatment, 2012-2017: a Fragile Balance between Respect for Authenticity, Material Ageing, History, Function, and Safety The conservation and restoration treatment proposed by the kik-irpa followed the recommendations stated in the report concluding the 2010 campaign23 and approved by the Advisory Committee.

9

Supports The overall condition of the supports is surprisingly good, considering that most of the panels are thinned and cradled and that all were kept in unstable climatic conditions. Some small splits and disjoints will have to be glued, cradles will be unblocked and older structural interventions reassessed. The stiff and heavy iron framework designed in 1951 with the intention of supporting the paintings safely (fig. 1.8) will have to be adapted in the short term as the panels are returned to the Villa Chapel. Indeed, the framework itself causes tensions in some panels and the mounting system involves screwing directly into the original frames, which has caused damage to the polyptych’s original structure and polychromy. Moreover, releasing

Fig. 1.8 Construction of the metal framework for the reinstallation of the Ghent Altarpiece in the Vijd Chapel in 1951

10

anne van grevenstein-kruse and hélène dubois

the painted panels from the framework is a cumbersome process, which makes it practically impossible to evacuate them from the glass case in an emergency. The technical aspects of a future display mode, whether in a new case or possibly using climate boxes, will be revaluated through a close collaboration between the team of conservators and various research groups. The issue of moisture barriers and the influence of past wax impregnations will also be investigated, in close collaboration with several research teams.24 Frames Frames embellished with gilding and polychromy, often imitating carved stone, are an intrinsic part of Early Netherlandish paintings. In particular, the frames of Van Eyck’s painting display dazzling illusionistic suggestions of carved letters and materials. The frames of the large centre panels were lost at some point in history but those of the wings are still conserved, although, except for the Adam and Eve panels, their structure and polychromy were drastically altered during treatments carried out in Berlin between 1823 and 1894. When the six double-sided wing-panels were acquired by Friedrich Wilhelm III of Prussia, the backs of the frames were covered with greenish overpaint that was partly removed in 1823, leading to the discovery of the quatrain at the bottom of the frames of the reverse side of the lower register but causing abrasion in the inscriptions (fig. 1.9).25 In 1894 the engaged frames were also sawn through, the hinges were removed and the mouldings were glued on a pine structure to adapt to the new presentation of the panels. The polychromy, severely damaged in places by the insertion of strengthening braces and locks, was then extensively restored (fig. 1.10) and now appears dark and atypically unrefined (fig. 1.11). This important and neglected part of the polyptych will be researched, the structure of the frames will be stabilized, and the sophisticated polychromy with its important inscriptions will be researched and conserved. Indeed, their authenticity, particularly the quatrain, is still the subject

Fig. 1.9 The Angel Annunciate panel in its original frame, photograph taken in Berlin before 1894

of debate.26 Undoubtedly this research will be one of the most rewarding opportunities presented by this conservation project. Noteably, the frames of the Adam and Eve panels – with their views of a cityscape and an interior on the reverse – were neither sawn through nor stripped of later repaints, and shadows of the original stone imitation with the lead-white traces of the joints on silver leaf on the reverse are clearly visible in raking light and on the X-radiographs (figs 1.12a-b).27 Removal of

the ghent altarpiece revisited: 2012-2017

11

Fig. 1.10 The Angel Annunciate panel in its original frame, after the major restoration that involved transforming the originally double-sided panels and frames into single-sided paintings. All metal fitting were removed and the polychromy was extensively overpainted

Fig. 1.11 The Angel Annunciate panel in its original frame, showing the darkening of the overpainted polychromy

the overpaint layers will be considered if at all feasible without alteration to the original surface with its fragile glazes. To regain the unity of the presentation while dealing with frames in very different conditions is a challenging issue for the treatment. A virtual reconstruction of the polychromy based on observations and laboratory anal-

ysis and a material reconstruction presented in a didactic format could also be considered. The examination and treatment of the original frames will be carried out with conservators specialized in polychrome sculpture, in a multidisciplinary collaboration that is one of the very strong assets of the kik-irpa.

12

anne van grevenstein-kruse and hélène dubois

Figs 1.12a-b A section of the frame of the Interior with City View panel: (a) in raking light; (b) X-radiograph showing traces of the lead-white joint in the stone imitation of the polychromy

Paint Layers The last overall conservation treatment of the painted surface dates from 1950-1951. The period allowed for this treatment was extremely short: following the recommendations of the international commission,28 the initial five months were extended to just over a year to carry out the consolidation of all the paint layers by impregnation with spike oil, beeswax and colophony (fig. 1.13), thin down the old varnish layers, locally remove overpaint, and retouch. The degraded varnish on the reverse of the wing-panels was barely thinned down and Philippot and Coremans were worried that the condition of the old varnishes would worsen, since cleaning could not be pushed any further.29 The 1947 ‘cleaning controversy’ about the treatment of the paintings at London’s National Gallery, the impact of influential art historians such as Ernst Gombrich, and the debates on the definition of the profession were high on the international agenda.30

Since 1951, the paintings have been varnished at least four times with modern synthetic varnishes that were thought to be more stable than natural resins31. Today, all the varnish layers, old and modern, are severely discoloured and have lost their transparency. Small paint flakes have become loose and some retouchings have become visible. The treatment can be summarized briefly in broad terms such as consolidation of the paint layers, removal of the synthetic varnish layers and thinning down of the older varnishes to achieve a homogenous and regular surface, removing discoloured old retouching step by step while assessing the presence and nature of overpaints and their possible removal. It is obvious that during this process a continuous flow of collateral research and analytical backup will support the decisionmaking process itself. With the support of the kikirpa research laboratories, with in particular a research project on the re-examination of the paint

the ghent altarpiece revisited: 2012-2017

13

Fig. 1.13 Albert Philippot applying a mixture of beeswax and colophony to impregnate the paint layers of the Hermits panel

samples taken in 1950-1951 and 1986, under the direction of Dr Jana Sanyova,32 and a large body of international experts, the treatment will be organized step by step, with regular meetings at crucial moments to facilitate the taking of fully informed decisions. The contribution of the diverse commissions will be integrated into the process and rationale of the treatment itself. Team and Organization of the Treatment The team in charge of the project is made up of conservators experienced in the research and treatment of Old Master paintings and Early Netherlandish paintings in particular.33 They are wellknown members of the international research community and are also familiar with the aesthetic and material evolution of Albert Philippot’s resto-

rations, the remaining varnish layers manipulated to modify and temper contrast. Time has moved on, loyalty remains but measured opinion has replaced solitary working. The conservators know how to recognize and conserve fragile shadows, the questions to ask and how to answer them. Furthermore, they teach in national conservation training programmes and lecture in Belgium and abroad, thus opening the project to knowledge transfer and student research. The treatment (which was initially planned to take five years and has subsequently been extended by an additional two years) consists of three phases. In each phase, some of the paintings are treated in a purpose-built studio in the Museum of Fine Arts in Ghent while the rest of the panels remain on view in the Villa Chapel. Photographs replace the

14

anne van grevenstein-kruse and hélène dubois

paintings in treatment so that visitors can still see the iconography of the altarpiece.34 In the museum the public can follow the ongoing work from a glazed visitors’ gallery and understand the need for time and financial means to support exhaustive research and meticulous execution. This open approach, which aims to give an informed insight into the work of the conservators and the reasons behind material changes and aesthetic transformations of the altarpiece, is essential for communicating the project: past controversies on restoration have been partly the consequence of the conservators’ isolation, as their work went on behind closed doors. The separation of the treated works of art from the public causes distrust of the ‘men of action who take risks’ as qualified by Gombrich, a perception that must be taken into account since lobbies and the media can express unmeasured opinion on conservation and restoration. Conservation Research, 2012-2017 The conservation research during the treatment will be carried out with three main paths in mind, with of course many overlapping areas: the identification of the artists’materials and working procedures, the condition of the paintings and their frames, and the material history of the altarpiece. This research will be carried out by the conservators, together with the laboratories as well as art historians specialized in Early Netherlandish painting, in particular through the Centre for the Study of the Flemish Primitives at the kik-irpa. The three paths of research will be constantly referenced in the methodology and decision-making process of the different phases of the treatment as well as the complex issue of the presentation and conservation of the altarpiece after treatment. The project will also be carefully documented by the conservators and kik-irpa photographers. The photographs of paintings and frames, taken at key moments during treatment, will be made accessible to experts and the public alike on the speciallydedicated website.35 Together with new infrared reflectograms, made after cleaning and especially

the removal of obscuring retouching and overpaint, as well as material analysis, they should enable both an understanding of the true material condition of the altarpiece and a review of the factors guiding the treatments as well as their final results. Identification of the Artists’ Materials and Working Procedures Together with the conservators, the laboratories will look again into the analysis of the materials used by the painter (or painters), and the techniques and working procedures applied in the altarpiece: pigments, fillers, binding media, underdrawing and paint build-up. Beyond the challenges of analysis, they will focus on linking objective analytical results to the painted forms and look at the implications of the similarities or differences in technical features for the division of labour within the chronology of the elaboration of the altarpiece. This research will rely also on the technical images produced during the 2010 campaign, which are accessible on the ‘closer to van eyck’ website (http://closertovaneyck.kikirpa.be). These images already enable new insights into Van Eyck’s practices such as the astonishingly freely applied underlayers blocking dark areas such as the trees and rocks that are clearly visible on the new infrared reflectograms (figs 1.14a-b). In the angels’ fair hair (figs 1.15a-c) these underlayers are only visible on the X-radiographs (fig. 1.15b), presumably because they contain lead white and no black. The extraordinarily complex and multiple paint build-up will have to be decoded. Samples analysed since 1951 show the presence of different preparatory layers and underdrawings and diverse paint constructions, to obtain specific variations in intensity and translucence.36 Since only very few paintings by Van Eyck have as yet been subjected to a detailed technical examination, it would be premature to conclude that this sophisticated build-up is specific to the Ghent Altarpiece. The completion of this monumental assemblage, made up of paintings of varying scales and involving the participation of different hands over several years,

the ghent altarpiece revisited: 2012-2017

could have required final adjustments such as the local or broad retouching of large areas. On top of this, extensive early restorations diagnosed by Coremans, principally in the Adoration of the Lamb, the Deity Enthroned, the Virgin Enthroned and the Singing and Musician Angels, complicate the interpretation of paint build-up (see below). The elucidation of the Van Eyck’s secrets is, of course, a pending challenge for all the specialists involved. Despite Giorgio Vasari’s oft-repeated association of Van Eyck with the invention of oil paint, the technical novelties are much more complex than the devising of a new binding medium. Coremans and Thissen had identified an aqueous binder in blue glazes and qualified the medium in most paint layers as drying oil with an undetermined additive – X.37 The use of different binding media in the superimposed paint layers, a technical feature identified in several Early Netherlandish masterpieces,38 associated with pigments of varying coarseness, would have been applied to obtain subtle modulations of different surface effects, colour saturation and hue. In a few samples, unpigmented intermediary layers are clearly part of the original technique, but their function is yet to be determined. Recent research has shown that this feature is not unique to the Ghent Altarpiece: the analysis of a number of paint samples carried out at the National Gallery scientific department during the restoration by Jill Dunkerton of the Portrait of Margaret van Eyck showed that in the red draperies Jan van Eyck used a very thin intermediary layer of varnish between the red glazes, a practice that was also documented in earlier medieval paintings.39 This new information implies that intermediary clear layers in a paint build-up cannot be systematically identified as varnish layers applied as a finish and therefore separating original from overpainting layers. The stylistic link of the altarpiece with polychrome sculpture, in particular for the trompe-l’oeil effects in its three-dimensional rendering and Claus Sluter as important source of inspiration, have been pointed out.40 This connection is further

15

strengthened by the use of applied brocades on the cloths of honour behind the monumental enthroned figures. This imitation of brocade, which was achieved by pasting onto the surface gilded tin leaves that had been pressed into a mould to obtain a pattern in relief then stiffened with a filler, is more often encountered on polychrome sculpture in the Low Countries and was yet unknown to Coremans and Philippot. Indeed, the first publications – by Mojmir Frinta and Thomas Brachert – only appeared in the 1960s, reflecting research carried out in Germanic areas.41 Johannes Taubert, another pioneer of polychromy study, had recognized the brocade imitations. He also suggested that gold ornaments, perhaps stars, could have decorated the once mattazurite-covered spandrels behind the enthroned figures, connecting the illusionistic paintings with an architectonic structure such as a canopy, with azurite-painted vaults decorated with stars, a feature often found in southern Germanic altarpieces.42 As noted by J.R.J. van Asperen de Boer, other technical features of the central enthroned figures are reminiscent of polychromy techniques of the period: the use of translucent glazes on silver and gold foils for the tiles, preparatory contour lines engraved in the ground in areas covered with metal foil, and the possible use of aqueous binding media for the blues, that could have produced a more matt, velvety finish contrasting with glossier glazes and shimmering gold.43 The description of the applied brocades was published only recently, following Ingrid Geelen’s examination during the urgent conservation campaign of 2010.44 The extremely degraded and overpainted condition of these areas makes it difficult to envision their original appearance and will warrant more detailed study to guide their conservation. Material Condition Both the natural evolution of the materials and the human interventions still influence the altarpiece’s state of conservation: the formation of craquelure,

16

anne van grevenstein-kruse and hélène dubois

Figs 1.14a-b Ghent Altarpiece, detail of the Hermits: (a) in normal light; (b) IRR. The trees in the background are freely blocked in with dark colours; the broad brushstrokes of the underlayer are clearly visible on the infrared reflectograms

the distortion of surfaces, the alteration of pigments and binding media, and the abrasion and detachment of paint layers are the main concerns. Whereas the discoloration and opacification of varnish layers is immediately noticeable, their chemical and mechanical interactions with the paint layers are also matters of concern that will be

investigated. Indeed, it is essential to understand which factors determine the apparent delamination of thin finishing paint layers, together with islands of varnish, that were already mentioned in 1936 (fig. 1.16).45 As will be seen from the paper by Spring and Morrison in this volume, laboratory analysis since

the ghent altarpiece revisited: 2012-2017

17

Fig. 1.14b

the 1990s has shed new light on degradation mechanisms caused by pigment-medium interaction, on how they affect the appearance of the paintings, but also how they can interfere with analytical results. There are therefore many facets to laboratory analysis that are relevant to the understanding of the degradation of materials and on the visual perception of the painted form.

The numerous and sometimes extensive changes introduced in the course of the painting of the Ghent Altarpiece are likely to be a significant contributing factor in the development of severe crack patterns in many areas. One particularly striking example of these changes is found on the drapery of the organ-playing Musician Angel: under the heavy brown and black brocade, marked by a

18

anne van grevenstein-kruse and hélène dubois

Figs 1.15a-c Ghent Altarpiece, detail of the Musician Angels: (a) in normal light; (b) X-radiograph; (c) IRR. The strands of hair appear clearly defined on the X-radiograph but not in infrared reflectography

dense network of deep cracks, is another, probably pale purple robe structured by different, stiff folds (fig. 1.17).46 The study of the impact of former treatments on the structure of the paint layers, the influence of heat and long term swelling solvents, and of impregnation with adhesives such as wax and

colophony, will be essential to the designing of some aspects of the treatment and the long-term conservation of the paint layers. Material History Research on the history of the altarpiece will have to be reviewed, particularly the historical events

the ghent altarpiece revisited: 2012-2017

19

Fig. 1.15b

and restorations that have affected its condition and which therefore significantly influence our perception of the paintings. Several aspects of the nature of the many treatments and transformations the altarpiece has undergone have been well researched already and Anton De Schryver and Roger Marijnissen’s exhaustive overview, published in l’Agneau Mystique au laboratoire, remains the reference for most of the archival documents rele-

vant to this subject.47 However, many questions are still controversial or unanswered, and the relevance of the historical sources will need to be ascertained by establishing clear connections between the written documents and the alterations identified by the technical examination and analysis of the paintings themselves. Modifications that have been carried out over the centuries to the altarpiece’s structure, and to the paintings and the poly-

20

anne van grevenstein-kruse and hélène dubois

Fig. 1.15c

chromy of the frames, must be taken into account when deciding on the nature of the treatment. In turn, material analysis will support the understanding of the early changes made to the altarpiece, which have been the subject of much controversy: were they carried out by the Van Eyck brothers, or did Jan make alterations to Hubert’s work? Do they reflect the collaboration of other painters within the workshop, or were they made by later artists?

Hitherto unnoticed compositional changes are bound to be discovered in the course of treatment by the conservators, who will spend countless hours of intense concentration in direct contact with the paintings, supported by advanced technology in technical analysis and imagery. The apparently extensive early overpainting of large areas of the paintings must be identified. This issue is particularly complex since most of the

the ghent altarpiece revisited: 2012-2017

materials used by the early restorers are similar to Van Eyck’s and to a great extent their paint layers would have aged in similar ways. These factors have led to contradictory conclusions from the researchers who studied this issue. During the restoration and examination of the polyptych under the direction of Paul Coremans in 1951 it was suggested that several areas of the inner side, in particular the Adoration of the Lamb, the Singing Angels and the enthroned figures, especially the Deity Enthroned, had been overpainted early on, and, presumably, in 1550 by Jan van Scorel and Lancelot Blondeel.48 This conclusion is based on the historian Marcus van Vaernewyck’s 1568 testimony that these renowned artists washed and cleaned the painting in many places49 and also on the fact that many features of that early overpainting are reproduced on the copy of the altarpiece that Michiel Coxcie produced in 1557. The representation of Utrecht Cathedral’s tower in an overpainting phase in the Adoration of the Lamb reinforced the assumption that Van Scorel, a canon of Utrecht, painted this area.50 The paint build-up of the red and green draperies, studied in microscopic cross sections, is very complex, comprising alternate applications of opaque and translucent paint layers. Intermediary unpigmented layers separating the paint applications were considered by Coremans as varnish layers separating the original from the overpainting. J.R.J. van Asperen de Boer, who conducted a technical study of the polyptych over several years, nuanced Coremans’s conclusions. On the basis of the available research he found it difficult to conclude that so many prominent areas were overpainted, maintaining that many changes, particularly in the Angels and Adoration panels, could have been made by Jan van Eyck following corrections directed by scholars and theologians.51 He also suggested that some parts might have been overpainted very early on, perhaps by others than Van Scorel and Blondeel but in any case before Coxcie made his copy. He also suggested that other changes were introduced later, such as the face of

21

the second Singing Angel, repainted between 1557 and 1625.52 The laboratory analysis carried out in the 1980s by Leopold Kockaert, Luc Maes and Jan Wouters (kik-irpa), mostly of existing samples taken in 1951, also disagreed with Paul Coremans’s and Jean Thissens’s conclusions regarding the presence of overpainting. In the view of Kockaert, Maes and Wouters the complex paint layer buildup is original and neither the Angels’ nor the Enthroned Deity’s draperies are overpainted.53 Archives The already extraordinary historical documentation will need to be reviewed and further completed, thanks to the collaboration with several archival depositories. For example, a first overview of archive documents in the archives of the Réunion des Musées Nationaux (Paris, Musée du Louvre) has revealed – in a bill submitted by the restorer Röser for the period 24 September 1798 to 30 December 1799 – a hitherto unknown reference to the restoration of the Deisis: 3 tableaux ceintrés en bois peints par Jean de Bruges ou van Eyck repant l’un un Pape, l’autre la Vierge & le 3 un Saint, nettoyé & restauré, hauteur 2 mètres largeur un mètre54 (fig. 1.18). This confirms the attribution by the administration of the restoration to Röser and the sawing of the upper parts of the boards to to facilitate framing.55 The excédent de planches were probably supports for ornaments that belonged to the original installation of the altarpiece. Bills sent in by this restorer reveal that he used esprit de vin (alcohol) in large quantities to clean paintings at great speed and systematically revarnished them. He occasionally mentions the removal of overpaint and of fillings overlapping onto the original paint, as well as carrying out retouching.56 Research in local archives in Belgium, such as the diocesan archive of St Bavo’s, has revealed further references to yet unknown treatments carried out in the nineteenth and twentieth centuries, such as washing the surface of the paintings with water and breadcrumbs to prepare them for photography.57

22

anne van grevenstein-kruse and hélène dubois

Figs 1.16a-b Ghent Altarpiece, detail of Joos Vijd: (a) in raking light; (b) macrophotograph from 1951: minute flakes of paint have detached from underlying layers in Joos Vijd’s face, a condition clearly visible in raking light (Photo atelier, 2010); this condition was already documented in 1951

This archival research is particularly important because it gives more insight into the lively history of the Ghent Altarpiece, which over time caused many alterations, particularly the dispersal of its different components. The material history and the condition of the different parts vary considerably, adding a complex dimension to the interpretation of degradations and possible changes of appearance over the years. Among the treatments known to us, the transformations carried out on the wing panels in Berlin between 1823 and 1920 are particularly remarkable. A careful study of photographs taken at different times before this major alteration

reveals essential information on the original polychromy and construction of the frames and on the evolution of their condition over several decades. Our quest for shadows of past interventions, of original traces of hinges, locks or hanging systems, relies mostly on these visual sources and on the technical examination of the construction of the frames in their present condition. Several original prints of the pictures taken before 1894 are kept in different archives, such as the Netherlands Institute for Art History (rkd) (Friedländer Archive) or the Deutsches Dokumentationszentrum für Kunstgeschichte – Bildarchiv Foto Marburg, and the

the ghent altarpiece revisited: 2012-2017

23

Fig. 1.16b

high quality digitization of these photographs offers great possibilities for research and documentation that are essential for the treatment (figs 1.9, 1.10). The kik-irpa also holds very rich documentation related to the more recent material history of the Ghent Altarpiece, particularly photographs taken after the return of the polyptych to Belgium following its retrieval from the Altaussee salt mines in 1945, and to the conservation treatment carried out in the Laboratoire central des Musées de Belgique between October 1950 and November 1951. The great majority of the technical photographs taken at the kik-irpa during the treatments to serve as research material and documentation were

never published. The large glass negatives (mostly 18 ≈ 24 cm) are now being systematically scanned to bring back to light a wealth of information that will give a more accurate insight into the work of our predecessors (fig. 1.19).58 Albert Philippot’s Techniques Albert Philippot, who restored the Ghent Altarpiece in 1951, worked on many Early Netherlandish masterpieces from the whole of Belgium (fig. 1.20). His expertise, together with the theoretical vision he and his son Paul constructed, have been transmitted through the next generations and are still very much at the core of kik-irpa’s conservation

24

anne van grevenstein-kruse and hélène dubois

Fig. 1.17 Ghent Altarpiece, detail of the Musician Angels: pronounced drying cracks mark dark areas in the organ-playing angel’s brocade mantle and the organ

practice. A study of his techniques has never been carried out and is called for in order to fully understand the work he did here. Albert Philippot was not a writer and his own notes are seldom to be found. Those who saw him

at work will confirm that the paintings he treated always looked well during treatment. The analysis of tonal values, their harmony and balance, essential in our perception of space, was central to his approach to cleaning. During cleaning he saturated

the ghent altarpiece revisited: 2012-2017

25

Fig. 1.18 Mémoire des tableaux nettoyés par le Cen Réser 9 Vendémiaire an 7 de la République jusqu’au 9 Nivôse 19 (Archives des musées nationaux MM an VII 3 Vendémiaire to 22 Nivôse)

the paint surface with linseed oil or varnish, tempered the highlights that stood out too harshly, and retouched zones that were optically disturbing for better legibility of the overall picture as the cleaning progressed. He often first ‘regenerated’ old varnish layers by swelling them with slowly evaporating spike oil and redistributing them over zones that lacked patina, a technique that his British

colleagues would have called ‘pushing the varnish around a bit’. In the 1950s, systematic conservation documentation was in its early years; detail photographs of the surface of the Ghent Altarpiece were taken, but a total image taken after cleaning and before retouching to document the overall condition does not exist. So we are looking at history, interpreted by a skilled, prudent and silent man.

26

anne van grevenstein-kruse and hélène dubois

Fig. 1.19 Demonstration of the consolidation technique with wax and resin to secure loose paint on a test painting. This document was made to explain the technique to the international commission that followed the treatment of the Ghent Altarpiece in 1951

Philippot introduced the objective documentation of his work later in his career, as this practice became more standard in the museum world. However, much precise information on his restorations can be gathered from the careful notes that were taken by interns and assistants and kept in the kik-irpa archives. One example of such documentation is the dossier written up on the treatment of the Trinity attributed at the time to Robert Campin, which Philippot restored in 1953. The assistant describes in detail the different phases of treatment and the materials used, illustrated with sketches. Surface cleaning, varnish removal, removal of overpaint, filling of the losses, the production and colouring of artificial craquelure in the fillings and the consolidation with wax and resin

are described, including comments on the effects of the products on the original paint (fig. 1.21). Interviewing conservators who worked with Philippot and applied and adapted his techniques themselves will also contribute to the reconstruction of his work.59 This information is crucial to understanding the physical condition of the paintings and bridging the gap between the theoretical discourse on aesthetics, history and ethics and its practical application through material interventions on unique works of art. Interdisciplinarity Beyond the survey of materials and techniques, the interpretation of the analytical results will be carried out by referring to the extended observation of

the ghent altarpiece revisited: 2012-2017

27

Fig. 1.20 View of the painting conservation workshop of the Central Laboratories of Belgian Museums, then housed in the Royal Museum of History in Brussels (1954). Visible among the paintings are Memling’s Triptych of St John the Baptist and St John the Evangelist, Dieric Bouts’s Justice Panels and Quinten Metsys’s Virgin and Child Enthroned, The Descent of the Cross by the Master of Frankfurt and St Martin Dividing his Cloak by Van Dyke

the paintings by the team of conservators during the treatment and through references to other paintings by Van Eyck, his contemporaries and predecessors.60 Crucially, the material results of analysis and their implications for the understanding and display of the altarpiece will have to be weighed up in a truly interdisciplinary collaboration between the conservators, the analysts and the specialist art historians in an ongoing dialogue, a structure that has been built over the years at the kik-irpa since the treatment of the Ghent Altarpiece in 1950-1951. As we all know, true interdisciplinarity is very difficult to achieve: highly specialized researchers

in different disciplines may not have the basic grounding in the other fields that they need in order to communicate. In tandem with the research an open exchange of ideas and cross-field interpretation has to go on, each specialist benefiting from the input of the other in a dynamic flow of ideas that prompt the questioning of results. Working Groups, Committees and the International Commission of Experts Interdisciplinarity within the project will be further strengthened by several working groups and advisory committees. Five different working groups and commissions, reflecting the different issues to be resolved, will follow the treatment. They will

28

anne van grevenstein-kruse and hélène dubois

Fig. 1.21 Extract from the treatment notes of The Trinity, dated 2 October 1953, at which time the painting was attributed to Robert Campin. The note reads: A. Impregnation à froid de la tête Christ 1.- Nettoyage térébenthine (tampon) pour enlever poussière, vernis à retoucher, autre crasse. 2.- Application mélange cire aspic (consistance de la crème fraîche) avec un pinceau doux. 3.- Tête recouverte d’un verre pour ralentir évaporation aspic

address preventive conservation and presentation, critical and historical issues, painting technique and analysis, scientific study of conservation issues and treatment of conservation-restoration.61 This last, international commission will be asked to meet more often and will be composed of expert conservators familiar with the treatment of Early Netherlandish paintings, and art-historians and conservation scientists specialized in this period.62 Of course we are not starting from scratch but are building on the expertise of our predecessors. With each round of laboratory analysis of the Ghent Altarpiece, from Jean Thissen and Paul Coremans, Louis Loose, Liliane Masschelein-Kleiner, Dolf van Asperen de Boer, Leopold Kockaert, Pim Brinkman and Jan Wouters until today, new information has been found and finer analysis interpreted in a context that is better known (fig. 1.22). Art historical studies of the Ghent Altarpiece are legion; authors have debated every aspect of the history, iconography, genesis, style and attribution issues. Like a Rubik’s cube, the Ghent Altarpiece and the artist’s (or artists’) process has been minutely tested and manipulated, resulting in proposals of different construction and elaboration phases from

Fig. 1.22 Jean Thissen observing a cross-section under a microscope, 1951

the angle of iconography, style, technique, history and linguistics, and the theories proposed have never failed to stimulate passionate discussions and occasional mordant controversies. After Hermann Beenken, Émile Renders, Max J. Friedländer, Erwin Panofsky, Otto Pächt, Maurice Brockwell, Elisabeth Dhanens, Albert Châtelet and, at this conference, Hugo van der Velden, Bernhard Ridderbos and Luc Dequeker, to name but a few of the many scholars who have worked on the altarpiece, it is hard to imagine the possibility of exploring new avenues of research in this field.63 Yet the treatment, the observations made, the amazing potential of technical photography illustrated by the ‘closer to van eyck’ website and the intensive collaboration with other specialists, will provide the possibility of linking material evidence to a broad context of different disciplines. Evidence on the execution techniques, original appearance and condition can be gained from different sources,

the ghent altarpiece revisited: 2012-2017

such as the rationale of the carpentry, the construction techniques and the proportions of the supports and the frames, the cadence of a hexameter in the quatrain, archaeological finds in the Vijd Chapel walls and floors, and microscopic examination of paint layers. Ideally, findings about the multiple aspects or facets of the altarpiece should converge to form a solid ground on which decisions in conservation and restoration can be made. Analysis In the coming years, as already mentioned, the conservators from the kik-irpa will naturally be working in collaboration with the kik-irpa laboratories, particularly with Jana Sanyova, who has started to reprocess and reanalyse the samples taken over sixty years ago by Coremans and his team.64 The research will involve the collaboration of several Belgian universities and international consortia and synchrotron analysis. It will concentrate on the identification of paint materials with special emphasis on the minor components and the trace elements and binding media. The analytical results will be interpreted in several contexts: the history of conservation techniques, the influence of treatments on the materials, pigment technology, and trade. Reconstitutions of painting techniques will be carried out by using fine analytical results, a very useful exercise for comprehending the technique, but also for estimating the paintings’ original appearance and the degradations that may have occurred, and developing conservation and restoration treatments. New research made possible by the spectacular finding of original paint in the rebate of the original frame of the stolen Just Judges during the conservation treatment of the copy in 2010 will bring us still closer to the original painting techniques and materials. Exploring the intricacy of the multilayered technique will certainly shed new light on Van Eyck’s techniques. The analysis of the binding medium, or rather, as we have seen, of the binding media, is certainly the most challenging issue of the laboratory analysis.

29

The samples from the 1950s, often containing all the layers of the complex paint stratigraphy, are an extraordinary source of information. The kik-irpa holds many samples from the polyptych which are in varying states of conservation and need to be inventoried and reprocessed before the new round of analysis. The documentation of the sampling sites has survived: Thissen, and Kockaert after him, sketched the samples in colour and noted primary results that hold useful information (fig. 1.23). The University of Ghent will contribute to the undertaking through an ambitious interdisciplinary project focusing on the use of non-invasive techniques to study the degradation processes and the influence of the material history on the appearance and the presentation of the altarpiece. This project, directed by Peter Vandenabeele (archeometry, chemistry), Luc Moens (analytical chemistry), and Maximiliaan Martens (art history) will involve four PhD research projects.65 Among the proposed examination techniques, high resolution digital microscopy will be used to document the condition of the paintings and the frames and provide a precise topography of the surface for further analysis. From Philippot’s microscope, which had to be positioned on the actual surface of the painting, and the photomicrographs that were taken (fig. 1.24), it also improves on the documentation conservators routinely gather today using a binocular microscope linked to a camera and computer to directly record and discuss their observations. The recent treatment and analysis of paintings by and attributed to Van Eyck in Berlin, Rotterdam and London and the study of panels in Dresden and Washington also offer the rich possibility of international collaboration.66 This dynamic integration of the practice of restoration into the interdisciplinary scientific field will provide the essential framework for a safer approach to the treatment.

30

anne van grevenstein-kruse and hélène dubois

Fig. 1.23 Card recording the stratigraphy of the golden rays and the sky around the dove in the Adoration of the Lamb panel

Conclusion Armed with skilled partnerships in several disciplines, we hope that we will meet the challenges awaiting us and, like Panofsky and Coremans, be carried on by the dynamics of a constantly evolving interdisciplinary dialogue to guarantee conservation and gain insight into Van Eyck’s vision. It is only in this context that the impact of the history of past interventions on the authentic vision of Van Eyck can be thoroughly assessed. It will enable us to ‘know what we are looking at’ and provide a crucial tool in the decision-making process for safe practice during treatment.

N OTES 1 See Van Belle in this volume. 2 Van Vaernewyck 1568, fols cxvij- cxix. 3 Coremans 1953. 4 Van der Veken, who with Albert Philippot had restored the altarpiece’s Adam and Eve panels in 1937 and had painted the copy of the stolen Just Judges panel, carried out an emergency conservation treatment of the paintings over a few days after the panels were returned from the salt mines of Altaussee, working under the supervision of Coremans. He attended the first meeting of the international commission to follow the treatment, on 10 November 1950, but was not involved in the restoration itself. Philippot was assisted by a number of craftsmen for the treatment of the supports and of the frames (KIK-IRPA conservation dossier). On Van der Veken’s forgeries, see Vanwijnsberghe 2008; specifically on his reputation among museum curators, Laemers 2008. 5 Pénéter l’art pour restaurer l’œuvr is the title of the volume edited by Catheline Périer D’Ieteren in homage to Paul Philippot: Périer-D’Ieteren 1990. 6 Coremans 1953, p.11.

the ghent altarpiece revisited: 2012-2017

31

Fig. 1.24 Albert Philippot controlling overpaint removal around the dove in the Adoration of the Lamb with a binocular microscope clamped to the edge of the panel

7 Paul Philippot also rightly insisted on the core importance of practice and craft in training the specific, critical skills of conservators: Philippot 1960. 8 Philippot 1967, p. 9. 9 The term ‘Technical Art History’ was first coined by David Bomford during the closing lecture of the 1995 IIC ‘Historical Painting Techniques, Materials and Studio Practice’ symposium in Leiden. He further discussed the evolution of this specialization in the introduction to Hermens 1998. 10 The composition of the advisory committee has varied throughout the project. It includes representatives of St Bavo’s Cathedral, the architect in charge of the cathedral’s restoration, the Flemish administration of monuments (Kunsten en Erfgoed) and fine arts (Topstukkenraad), the Province of East Flanders, the KIK-IRPA, the KMSKB-MRBAB, Monumentenwacht vzw and Lukas-art in Flanders vzw. Anne van Grevenstein-Kruse (University of Amsterdam), Ron Spronk (Queen’s University, Ontario/Radboud University Nijmegen) and Jørgen Wadum (National Gallery of Denmark/University of Amsterdam) were invited to join the council as international experts. For a full list, see http://closertovaneyck.kikirpa.be. 11 The severely tenting paint layers of the copy of the Just Judges were treated at the KIK-IRPA by Marie Postec between April and June 2010. 12 http://www.getty.edu/foundation/initiatives/current/panelpaintings (accessed October 2013).

13 Junior panel conservators, Jessica Roeders (Frans Hals Museum, Haarlem), Aline Genbrugge (KIK-IRPA) and Renzo Meurs (freelance conservator, Amsterdam) took part in the programme. Meetings in Ghent, coordinated by Antoine Wilmering (Getty Foundation) were attended by Su Ann Chui (J. Paul Getty Museum, Los Angeles), Jose de la Fuente Martinez (Museo del Prado, Madrid), Ray Marchand (Hamilton Kerr Institute, Cambridge), George Bisacca and Alan Miller (Metropolitan Museum, New York), Britta New (National Gallery, London), Jørgen Wadum (National Museum of Denmark, Copenhagen), Paul van Duin (Rijksmuseum, Amsterdam), Ingrid Hopfner (Kunsthistorisches Museum, Vienna), Béla Nagy (Budapest), Andrea Santacesaria (Opificio delle Pietre Dure, Florence), Hélène Verougstraete (Université Catholique de Louvain, Louvain-la-neuve), Al Brewer (Royal Collection, Windsor), Livia Depuydt and Hélène Dubois (KIK-IRPA) and Anne van Grevenstein (University of Amsterdam). 14 The dendrochronological examination of the Adam and Eve panels could not be carried out in 2010 because their original, engaged frames would have had to be dismantled, an invasive operation that could not be justified within the scope of the examination campaign. Peter Klein attempted to date the panels on the basis of the X-radiographs (see reports on the website). The examination of the wing panels, with the exception of Adam and Eve, was carried out by Josef Vinckier in 1986 (Vinckier 2002). 15 CHARISMA (Cultural Heritage Advanced Research Infrastructures. Synergy for a Multidisciplinary Approach to Conservation-

32

anne van grevenstein-kruse and hélène dubois

Restoration) is an integrating activity project carried out in the FP7 Capacities Specific Programme co-funded by the European Commission (http://www.charismaproject.eu/home-page.aspx, accessed October 2013). MOLAB is a group of the Transnational Access programme of CHARISMA offering access to a portable set of advanced analytical equipment for in situ non-invasive measurements on artworks. 16 Prof. Arnold Janssens and Michel Depaepe, PhD candidate Lien Debacker (University of Ghent), Anne-Cathérine Olbrechts (Monumentenwacht Oost-Vlaanderen), Bart Ankersmit (Rijksdienst Cultureel Erfgoed Amsterdam), Prof. Henk Schellen (Eindhoven University of Technology). 17 The Klimaatnetwerk Vlaanderen and Klimaatnetwerk Nederland are associations of professionals and institutions from the heritage and academic sectors in Flanders and the Netherlands who develop and implement methods and strategies to regulate the climate in historic buildings and museums. Shin Maekawa, expert in preventive conservation from the Getty Conservation Institute, has been advising on this issue. 18 Goedleven 1986; Cnops 1986. 19 Borchert 2008, p. 32. 20 Van Asperen de Boer 2004; Châtelet 2011. 21 Dhanens 1969-1972. 22 De Schryver, Marijnissen 1953, pp. 59-60. 23 Van Grevenstein-Kruse 2011. 24 Jean Albert Glatigny, Prof. Dr Henk Schellen (Eindhoven University of Technology), Paul Van Duin (Rijksmuseum, Amsterdam), Lucca Uzielli (University of Florence), Josef Grill (University of Montpellier). The NWO-funded ‘Climate4Wood’ project 2012-2017 has been involved so far in developing research on the effect of the exhibition conditions (metal structure, climate in the showcase, short and long term modifications to the presentation). Since the presentation of this lecture, a study for the improvement of the climate and of presentation of the altarpiece in the Villa Chapel has been carried out by the firm HELICON. 25 Waagen 1824, p. 99. On the discovery of the quatrain, see Van der Velden 2011, p. 11. 26 The silver leaf was identified in 1951, as well as the presence of glazes but the authenticity of this decoration, as well as the gilding on the front of the frames, was doubted. The stone imitation, heavily restored in 1894, was thought to be modern. Coremans 1953, pp. 121122. 27 The presence of silver leaf was confirmed by hand-held XRF during the 2010 examination campaign (Jana Sanyova (KIK-IRPA) and Bart Vekemans (University of Ghent): see Van Grevenstein 2010, pp. 36-54). 28 On the interaction of the commission with the treatment, see Dubois, Van Grevenstein 2011. 29 Coremans 1953, pp. 10 and 91. 30 Coremans was a member of the Scientific Advisory Committee led by J. Weaver that addressed the cleaning issues and produced the ‘Weaver Report’: Weaver 1950. A number of articles relating to different approaches to cleaning paintings, first published in Museum in 1950 and 1951, were compiled in a publication produced by UNESCO: Rousseau 1950. The post-war public controversy continued in the 1960s through heated scholarly exchanges in articles in the Burlington Magazine: see Wechsler 1987. 31 Ketone-based retouching varnish produced by Talens. KIKIRPA conservation dossier. 32 The four-year project (2012-2015) entitled ‘The Mystic Lamb in the laboratory 60 years after Paul Coremans. The contribution of new analytical techniques’ is supported by BELSPO (Belgian Federal Science Policy), Project MO/39/011. 33 Livia Depuydt, who leads the team of conservators, restored Van Eyck’s Virgin by the Fountain from the KMSKA. The team is composed of Bart Devolder (on-site coordinator), Hélène Dubois (research

coordinator), Nathalie Laquière, Claire Mehagnoul, Marie Postec, Françoise Rosier and Griet Steyaert. 34 Although in black and white prints rather than colour reproductions so as to avoid any confusion by the visitors. 35 Conceived as an extension of the initial ‘closer to van eyck’ website (http://closertovaneyck.kikirpa.be). 36 Coremans 1953, pp. 69-76; Brinkman et al. 1984-1985; Brinkman et al. 1990. 37 In her paper for this conference, Melanie Gifford revoked her earlier identification of an aqueous medium in a blue glaze of the Washington Annunciation. 38 See for example Campbell et al. 1994, pp. 30-31, and particularly the keynote paper by Spring and Morrisonin this volume. 39 These results were presented at the Van Eyck Studies Colloquium by Jill Dunkerton, Rachel Morrison and Ashok Roy. The report on the restoration and technical data were published online (http:// www.nationalgallery.org.uk/paintings/research/the-restoration-ofmargaret-the-artists-wife/introduction). At the same colloquium a new study of the Washington Annunciation was also presented by Melanie Gifford, Cathy Metzger, Suzanne Lomax, John Delaney, Rachel Morrison and Marika Spring. 40 Borchert 2008, p.14. The comparison of the Diesis with the sculptures of Romanesque reliquaries was made by De Tolnay 1938, p. 5 ff. The conception of the paintings, in particular the upper register, as illusionistic and ambiguous three-dimensional representations inserted in an intricate monumental structure is developed by BrandtPhilip 1971. 41 Frinta 1963; Brachert 1964. Recent archival research by Delphine Steyaert on neo-Gothic polychromy techniques has shown that the sculptor Leopold Blanchaert, who stripped and reworked medieval sculptures before a new polychromy was applied by Adrien Hubert Bressers, had recognized that applied decorations in relief were made with wax (Steyaert 2014). 42 Dr Johannes Taubert shared this information with Anne van Grevenstein in 1971, during the conservation of the Annunciation by Veit Stosz in Nuremberg, Bayerishes Landesamt für Denkmalfpflege. Taubert had been a researcher at the KIK-IRPA in the early fifties and remained in close contact with the team for the study of polychrome sculpture under the guidance of Agnes Gräfin Ballestrem. Taubert also shared this visual analysis with J.R.J. van Asperen de Boer (Van Asperen de Boer 1976, pp. 145-146 n. 10). The presence of a monumental canopy crowning the altarpiece is discussed at length by Dhanens 1969/72 and by Brandt-Philip 1971. On the spandrels, see Genbrugge and Roeders in this volume. 43 As suggested by Van Asperen de Boer 2004, pp. 175-176. See also n. 37, above. 44 Geelen, Steyaert 2011, pp. 382-393. This outstanding book gives an overview of the applied brocade techniques used on different supports in the Low Countries. 45 Report by Georges Hulin de Loo, Leo Van Puyvelde and Josef Van der Veken, 1936, published in Coremans 1953, pp. 64-65 n. 92. 46 A paint sample from the robe was shown to contain lead white, red lake and azurite: Brinkman et al. 1990, p. 35. 47 Before these scholars, Canon de Kesel of St Bavo’s Cathedral devoted several publications to the subject. 48 Coremans 1953, pp. 98-99, 101-117, the red and green draperies of the Singing Angels (pp. 98-99), the brocades and the tiles in the panels of the enthroned figures (pp. 101-105), the faces of the Virgin Enthroned and John the Baptist Enthroned (p. 102) and large areas in the Adoration (pp. 106-117). 49 Van Vaernewyck 1568, fol. 117 v. 50 See Dubois, Sanyova, Vanwijnsberghe in this volume. 51 Van Asperen de Boer 1979, pp. 155-163, 172-178. 52 Ibid., p. 157. 53 Brinkman et al. 1990, pp. 35-37.

the ghent altarpiece revisited: 2012-2017

54 Mémoire des tableaux nettoyés par le Cen Réser 9 Vendémiaire an 7 de la République jusqu’au 9 Nivôse (Archives des musées nationaux MM an VII 3 Vendémiaire jusqu’au 22 Nivôse). 55 Archives des musées nationaux 1 BB4 séance n°169 du 13 fructidor an 6 (30 août 1798), transcription in Emile-Mâle 1994, p. 66. 56 On Röser, see Jamois 2005-2006. 57 First overview carried out by Dominique Deneffe and Hélène Dubois. 58 The KIK-IRPA holds over 3000 photos of the Ghent Altarpiece, documenting its condition and appearance and the way it has been displayed over the years. These photographs were digitized and made accessible through a database in a project conducted by Luc Stockart, Julie Mauro, Marie-Christine Claes and Damien Yernault. 59 Several conservators were interviewed about their practice through projects conducted by Professor Joyce Hill-Stoner (University of Delaware) and the Secco Suardo Foundation. 60 An overview of existing technical data on Van Eyck’s paintings at the National Gallery in London, the Prado and the Louvre, was carried-out by Marie Postec and Hélène Dubois in 2012 through the ARCHLAB programme of the CHARISMA project. All the scholars and scientists supporting this programme are warmly thanked for their help and expertise. 61 For these working groups, see http://closertovaneyck.kikirpa.be. 62 The international commission of experts assembled at the beginning of the project included Maryan Ainsworth, Liesbeth De

33

Belie, Till-Holger Borchert, Veronique Bücken, Lorne Campbell, Leslie Carlyle, Jill Dunkerton, Susan Farnell, Nicole Goetghebeur, Melanie Gifford, Régine Guislain-Wittermann, Babette Hartwieg, Lizet Klaassen, Cathy Metzger, Uta Neidhardt, Elke Oberthaler, Catheline PérierD’Ieteren, Marika Spring, Ron Spronk and Jørgen Wadum. They were later joined by Michael Gallagher and Maximilian Martens. 63 On the historiography of the altarpiece, see for example Ridderbos in this volume. 64 See n. 29, above. 65 Archeometrical study of the Ghent altarpiece, Special Research Fund, Concerted Research Action (2013-2017), University of Ghent. Supervisor: Peter Vandenabeele. 66 The Three Marys at the Tomb from the Museum Boijmans van Beuningen and the Crucifixion from the Gemäldegalerie were both restored in 2012 for the exhibition on pre-Eyckian painting, The Road to Van Eyck (curated by Friso Lammertse and Stephan Kemperdick). The materials of the Rotterdam painting are studied in the collaboration with the NWO Science4Arts Programme (2012-2017), in particular the multidisciplinary international PAinT research project ‘Paint Alterations in Time. Implications of pigment-binder interactions for conservation, presentation and storage of oil paintings from Van Eyck to Mondrian’ supervised by Prof. Piet Iedema (University of Amsterdam). The restoration and study of the Portrait of Margaret Van Eyck (Groeningemuseum, Bruges) and the examination of the Dresden Triptych both appear in this volume.

Fig. 2.1 Ghent Altarpiece, Adoration of the Lamb, detail, gemstones in the canal surrounding the Fountain of Life

Fig. 2.2 Ghent Altarpiece, Hermits, detail, coral and gemstones

2

Gems in the Water of Paradise. The Iconography and Reception of Heavenly Stones in the Ghent Altarpiece Marjolijn Bol

ABSTRACT: In 1566-1568, the Flemish chronicler Marcus van Vaernewyck wrote a comprehensive account of all the things depicted in the Ghent Altarpiece. In his description he points towards two details that at first glance are rather inconspicuous – the precious stones that lie scattered in the canal surrounding the Fountain of Life and the coral that ‘seems to grow from the mossy stones’ depicted near the stream of water in the Hermits panel. These are indeed odd details which, interestingly, have never been studied iconographically. The presence of these precious stones in and around water gains further significance when we discover that they can be found scattered on pebbled beaches in countless other fifteenth- and sixteenth-century paintings. In this paper I hope to explain the genesis and meaning of the painterly motif of gemstones in the water of Paradise in both the Christian tradition and pagan and medieval lore about the nature and origin of gemstones. Following rivers through vast mountainous landscapes, the discovery of precious stones was considered a sign that one was close to finding the long-lost Garden of Eden…

—o— In his diary, Van die beroerlicke tijden in die Nederlanden en voornamelick in Ghendt, 1566-1568, the Flemish chronicler Marcus van Vaernewyck (15181569) wrote a comprehensive account of all the things depicted in the Ghent Altarpiece. In this description he did not fail to notice two details that at first glance are rather inconspicuous: the precious stones that can be seen through the clear

water of the canal surrounding the Fountain of Life (fig. 2.1), and the coral that, according to Van Vaernewyck, ‘seems to grow from the mossy stones’ depicted near the stream of water in the panel depicting the Hermits (fig. 2.2).1 In light of the lengthy and ongoing debate regarding the iconographical implications of Van Eyck’s addition of the Fountain of Life to the Ghent Altarpiece, it is remarkable that the costly gemstones in its water have received little attention.2 This becomes all the more surprising when we consider the fact that similar precious stones can be found scattered in and around small streams in numerous other Early Netherlandish paintings.3 In this paper I argue that both the presence and meaning of the gemstones in the water surrounding the Fountain of Life and the coral4 and precious stones at the feet of the Hermits can be explained by studying a wide range of medieval lore about the nature and genesis of precious stones. Late-medieval theological literature, natural philosophical texts, travel chronicles, and maps show that people believed gemstones ‘grew’ in Adam and Eve’s earthly Paradise and, carried by the four paradisal rivers, were spread over earth. As Paradise was thought still to exist somewhere as a real place, people started travelling the four rivers in order to find it. The discovery of an abundance

36

marjolijn bol

of precious stones in or near a river thus became an indication of the proximity of the long-lost Garden of Eden. In the lower register of the Ghent Altarpiece the well-established literary motif of the ‘gems of the water of Paradise’ became visual for the first time and, subsequently, was readily adopted by many of Van Eyck’s followers. The Gems and the Fountain of Life in Biblical Scripture? In order to articulate how the various types of medieval lore could have contributed to Van Eyck’s addition of a fountain with precious stones to the Ghent Altarpiece, I will first discuss how the painter and his patrons could have been informed by biblical scripture. With the addition of the fountain and its identifying inscription, hic est fons aque vite procedens de sede dei + agni,5 Van Eyck refers to a well-known passage from the Revelation to John. St John writes that in the New Jerusalem of his vision he saw ‘a pure river (fluvius) of water of life, clear as crystal, proceeding out of the throne of God and of the Lamb.’6 Yet, Revelation does not refer to precious stones in the river of the water of life, nor does it mention a fountain. For this reason, Van Eyck’s inclusion of the fountain is usually explained by the fact that fountains were traditionally associated with salvation and eternal life and thus are a likely motif for an altarpiece with the mystery of Redemption as its main subject.7 In the rare instances that the gems in the canal are addressed, they are considered a reflection of the fact that St John famously describes how the New Jerusalem is known for containing an abundance of jewels and other precious materials.8 Interestingly, the prophecy of Zechariah (the eleventh of the twelve Minor Prophets) provides additional evidence for Van Eyck’s depiction of a fountain, as opposed to the river of living water that is found in Revelation.9 Zechariah, who is himself depicted in the far left lunette of the Ghent Altarpiece with a banderole inscribed with a verse from his prediction, prophesies the Redemption of

mankind.10 He writes that on the day of the Last Judgement ‘there shall be a fountain (fons) opened’ and ‘it shall be in that day, that living waters shall go out from Jerusalem; half of them toward the former sea, and half of them toward the hinder sea: in summer and in winter shall it be.’11 Zechariah’s prophecy thus provides us with the first evidence for Van Eyck introducing a fountain of living water to the Ghent Altarpiece. But in order to understand the presence of the precious stones in the canal surrounding it, we need to look further. Theologians, Gems, Fountains and the Paradise on Earth The centre panel of the Ghent Altarpiece depicts the Adoration of the Lamb, a scene based on a passage from Revelation that is traditionally read on All Saints Day.12 With the inclusion of the fountain of living water, Van Eyck ingeniously transformed this conventional image of the Adoration of the Lamb into a representation of the Heavenly Jerusalem. He does not depict it as the great city in the sky with the jewelled walls known from numerous earlier works, but rather as an earthly garden, with trees, flowers and herbs all rendered with such precision that, to this day, botanists are able to identify them.13 In this way, and without visual precedent, Van Eyck associates the heavenly New Jerusalem with the first Paradise on earth: Adam and Eve’s Garden of Eden. In theological writings of the fifteenth century the typological analogies of the Old and New Testament, and accordingly the many parallels between the two versions of Paradise, were fervently debated. In Genesis, in the verses describing how the ‘Lord God planted eastward in Eden’ a garden of earthly delights, we read for instance that parallel to Revelation’s New Jerusalem ‘the tree of life was in the midst of the garden’ and, again similarly, it includes a river that ‘went out of Eden to water the garden […]’14 These parallels aside, it will become clear that in order to understand the presence and significance of the precious stones in Van Eyck’s fountain of living water, the way in which medieval lore

gems in the water of paradise

develops the subsequent verses from Genesis is instrumental. In those verses we are told how the paradisal river flows out of the garden and is thence divided into four large rivers called Pishon, Gihon, Hiddekel and Euphrates: […] and from thence it was parted, and became into four heads. The name of the first is Pison: that is it which compasseth the whole land of Havilah, where there is gold; And the gold of that land is good: there is bdellium and the onyx stone. And the name of the second river is Gihon: the same is it that compasseth the whole land of Ethiopia. And the name of the third river is Hiddekel: that is it which goeth toward the east of Assyria. And the fourth river is Euphrates.15 In his two influential commentaries on the book of Genesis, St Augustine of Hippo (354-430) expands on and explains the previous passage. In A Refutation of the Manichees he adds that besides gold, aromatic resin and onyx, the Pishon carries carbuncles (a translucent red stone) and leek-green stone (probably emerald).16 Here, for the first time, Augustine provides us with an unequivocal account of a variety of precious stones in paradisal streams. He explains the meaning of the presence of these gemstones when he connects the four paradisal rivers to the four cardinal virtues. According to St Augustine, the Pishon is associated with prudence because: […] this prudence [Pishon] then goes round the land [Havilah] which has gold and carbuncle and leek-green stone, that is, a discipline of life that glistens brightly as if refined from all earthly dross, like the best gold; and truth which no falsehood can overcome, like the brilliance of the carbuncle which is not overcome by night; and eternal life, which is signified by the greenness of the leek-green stone, because its vigour never withers.17

37

All aspects of St Augustine’s explanation of the meaning of Pishon closely correlate with the pure state of mankind before the Fall that was hoped to be restored with Christ’s Second Coming in the New Jerusalem. In fact, St Augustine’s text can be considered pivotal in the present argument, because not only does he account for the presence of a variety of gemstones in paradisal waters by drawing close parallels between earthly Eden and Heavenly Jerusalem, he was also the first to write about human salvation in terms of historical progression.18 This theme, as was pointed out earlier, is very important for our understanding of the iconography of the centre panel of Ghent Altarpiece.19 Augustine’s historical reading of the Bible also resulted in the idea that the four rivers of Paradise could still be found on earth. In his second commentary, The Literal Meaning of Genesis, St Augustine explains that this is true because the four paradisal rivers can still be identified in the regions they flow through: About these rivers, why should I bother any more to establish that they are real rivers, not just metaphors, as though they did not exist but only names signified something, seeing that they are obviously known in the regions they flow through, and indeed their fame has spread abroad to practically all peoples?20 About two centuries after St Augustine, a similar and equally influential idea about Paradise and the four rivers issuing forth from it can be found in the Etymologies of Isidore of Seville (c.560-636). Isidore is of particular relevance here, because he is the first to introduce the idea that the four rivers spring forth from a single source. In Book 14, describing ‘The Earth and its Parts’, Isidore writes in the entry to ‘Asia’ that here, in the ‘Eastern part of the world’, earthly Paradise is located.21 He explains that its name is ‘Garden’ in Greek and ‘Eden’ in Hebrew, meaning ‘Delight’ and that in this garden of delights there are many kinds of wood and fruittrees, the tree of life and that it is always spring.22

38

marjolijn bol

Isidore furthermore asserts that ‘a spring bursts forth in the centre irrigating the whole grove and is divided into the headwaters of four rivers.’23 Importantly, this last remark, as will later become clear, inspired subsequent authors to include a ‘fountain’ in their descriptions of Paradise on earth. In his entry about ‘rivers’ (De fluminibus), writing about the names of famous rivers, Isidore moreover claims that ‘the Euphrates is a river in Mesopotamia rising from Paradise, well supplied with gemstones; it flows through the middle of Babylonia.’24 As with St Augustine, for Isidore of Seville too, the Garden of Eden is a real location somewhere in the East and he likewise believes that the four paradisal streams are real rivers still extant, irrigating earth and supplying gemstones. Despite this ‘reality’ of Paradise, however, Isidore does not neglect to point out that access to it was blocked off after the fall of mankind, and that it is now well-protected by a garrison of angels and ‘fenced in on all sides by a flaming sword, that is, encircled by a wall of fire, so that the flames almost reach the sky.’25 While Original Sin caused Adam and Eve’s (and thus mankind’s) expulsion from Paradise, people believed until well into the sixteenth century that their paradisal garden could be found somewhere on earth and, by way of analogy to the New Jerusalem, held the promise of Redemption. By travelling the known rivers flowing out of the long-lost Garden of Eden one might come close to finding it again. In what follows, it becomes clear how the gems of the water of Paradise have a special place in all of this. Natural Philosophers on the Gems in the Water of Paradise Influenced by the ideas of contemporary theology, early-medieval natural philosophy similarly theorized that gems are carried from the Garden of Eden with the help of the paradisal rivers. Additionally, being interested in the genesis of materials in the natural world, natural philosophers attempted to give a more physical explanation of the origin of gems in streams.

Ingeniously combining the four elements, Hildegard of Bingen (c.1098-1179), Benedictine abbess of the Rhineland, explains in her Physica (1151-1158) how nature produces precious stones.26 Hildegard claims that gems are made in the East, in areas where the sun’s warmth causes the mountains to have heat as powerful as fire. Accordingly, when foam from the boiling rivers comes into contact with the hot mountains, gemstones are formed in the space of about three to four days.27 When the gems are dry they fall from the mountains into the sand ‘like flaking fish scales.’28 Finally, when the rivers flood again, the gems are carried ‘to other countries where they are later discovered by human beings.’29 In this way, Hildegard has given a resourceful physical explanation of how gemstones are formed from water and fire, helped by earth and air, while at the same time not neglecting their divine origin in the paradisal rivers springing from the Garden of Eden in the East. Thomas of Cantimpré (1201-1272), ‘pupil’ of the great philosopher Albertus Magnus (c.12001280), makes the connection between contemporary theology and natural philosophy even more explicit. Calling on St Augustine, Pliny, Solinus, Isidore of Seville and Jacques de Vitry as his authorities, Cantimpré writes in his Liber de natura rerum (1228-1244) that at the centre of earthly paradise there is a fountain (fonte) of the clearest water.30 This fountain has so much water that it divides into four streams and, starting in India, these rivers irrigate the earth. According to Cantimpré, one of the four rivers, the Pishon, falls in a miraculous manner from a mountain in India.31 In the introduction to his book on stones, he writes furthermore that terrestrial Paradise produces gemstones that are carried to other regions by the four rivers.32 Bartholomeus Anglicus (c.1203-1272), the other great medieval encyclopedist, corroborates most of Cantimpré’s ideas about the paradisal fountain and the four rivers, but in a special entry ‘on rivers’ he additionally claims that precious stones are carried from Paradise by the River Pishon specifically.33 As both the works of Cantimpré and

gems in the water of paradise

Anglicus were highly influential, their ideas about a large paradisal fountain and the Pishon being the carrier of the precious paradisal stones would resonate for centuries to come. Well into the fifteenth century Cantimpré was used as a source for many encyclopedias and Anglicus’s encyclopedia was translated into many languages.34 In the midfourteenth century the German scholar Konrad of Megenberg (1309-1374) writes one such translation and adaption of Cantimpré’s work, using it as the basis for his Buch der Natur (c.1349). I mention it here because a fifteenth-century illuminated codex of Megenberg’s work (c.1442-1448) contains an interesting frontispiece miniature to book six, ‘Von den Edel Stainen’ (fig.  2.3). The painted page shows two men pointing towards a stream of precious stones that appears to pour from a mountain. In the background, directly behind the stream of gemstones, a church connects this ‘river’ of gems to their divine genesis. The miniature appears to combine the idea that gems flood from paradise in the four rivers with Cantimpré’s claim that the Pishon falls from a huge mountain in India. In Search of Paradise on Earth As the Garden of Eden was thought to have been a real place on earth, medieval explorers tried everything in their power to find it. Although it was believed that one could only enter the Garden by the mercy of God, beautiful vegetation, a warm climate, and especially the rivers and an abundance of precious stones, were considered indications of the proximity of terrestrial Paradise.35 These journeys in search of Paradise were often recorded in written form and, judging by their widespread circulation in manuscripts and numerous translations into the vernacular, they were a very popular genre in medieval lore. Although most travel chronicles are characterized by the rather grand imagination of their authors and were sometimes even completely fictitious, they do give yet another interesting insight into medieval interpretations of the gems of the water of Paradise.

39

Two of the most famous chronicles are the socalled letter of Prester John and the Travels of Sir John Mandeville. Testifying to its enduring popularity the letter of Prester John survives in about a hundred copies in manuscripts dating from the twelfth to the seventeenth century.36 Prester John was believed to rule a kingdom bordering the Garden of Eden and this explained the many marvels travellers discovered in his empire. In one of the oldest versions of the letter, dating back to the twelfth century, its writer argues that the Pishon, coming from a fountain (fons) in Paradise, can be seen emerging in St John’s land and that it carries gold and precious stones.37 In three Middle High German, Middle English and Middle French versions of the letter, contemporary with the Ghent Altarpiece, we are told that in the land of Prester John a desert can be found that no one is allowed to enter and that includes a mountain from which floods a stream of gemstones without water that comes from Paradise.38 The travel account of Sir John Mandeville, a supposed English knight, first started circulating between 1357 and 1371.39 Although now believed to be completely fictitious it gained tremendous popularity and, like the letter of Prester John, was translated into many languages. Mandeville relates that on his travels he did not see Paradise itself but, having visited the lands of Prester John, came very close to it.40 He states, like many writers before him, that it is the Pishon that carries precious stones from Paradise: And in the most high place of Paradise, even in the middle place, is a well that casteth out the four floods that run by divers lands. Of the which, the first is clept Pison, or Ganges, that is all one; and it runneth throughout Ind or Emlak, in the which river be many precious stones, and much of lignum aloes and much gravel of gold. […] And men there beyond say, that all the sweet waters of the world, above and beneath, take their beginning of the well of Paradise, and out of that well all waters come and go.41

40

marjolijn bol

Fig. 2.3 Attributed to the workshop of Diebold Lauber, Konrad von Megenberg, Das Buch der Natur, frontispiece miniature to Buch VI: Von den edeln stainen, c.1442-1448, paper, 39.8 x 28 cm, Heidelberg, Universitätsbibliothek, (Cod. Pal. germ. 300, 320v)

gems in the water of paradise

None other than Christopher Columbus (14511506) himself is known to have used the above chronicles when writing his own travel diary. Indeed, when setting foot in South America he famously proclaimed that because of the beauty and riches of this land it must be close to Paradise: I believe that this water may originate from [paradise], though it be far away and may come to collect there were I came and may form this lake. These are great indications of the earthly paradise, for the situation agrees with the opinion of those holy and wise theologians, and also the signs are very much in accord with this idea, for I have read or heard of so great a quantity of fresh water coming into and near the salt. And the very mild climate supports this view.42 Mapping the Garden of Eden Maps of the world also disseminated ideas about the geography of the Garden of Eden and its four rivers. Early medieval maps often borrowed from the writings of the authors mentioned earlier. In some cases, maps were even made to illustrate such texts. Based on these writings, cartographers depicted the Garden of Eden as a place located somewhere in the Orient with the four paradisal rivers flowing out of the long lost Garden of Delights. Precisely in the time of Van Eyck, at the start of the great age of exploration, the time of numerous pilgrimages and crusades, maps were less frequently based on textual sources and instead were created with a growing concern for geographical accuracy and scale.43 The Garden of Eden, however, is still depicted in most of them.44 For instance, a Mappa Mundi (1448) made in Konstanz by Andreas Walsperger (b. c.1415) depicts Paradise as a large fortified castle built on a massive mountain out of which the four rivers flow in different directions.45 In another fifteenth-century example, the so-called Borgia World Map, not only Eden but also the Land of Prester John is marked as a geographical location.46

41

Rulers of the late Middle Ages generally had maps of the world in their possession, and Jan van Eyck’s patron, Philip the Good (1396-1467), was no exception. By sending people on pilgrimages and crusades to the East and the Holy Land, Jerusalem, they attempted to gain knowledge of the world, which was subsequently recorded on lavishly illustrated maps.47 A few documents suggest that Van Eyck also undertook such travels for Philip the Good and, if we are to believe Bartholomaeus Facius (c.1400-1457), Van Eyck himself made a ‘very accurate’ circular representation of the world for the Duke of Burgundy, in which ‘you may distinguish not only places and the lie of continents but also, by measurement, the distances between places.’48 From Van Eyck’s Paradise into Early Netherlandish Painting As Eden was believed to be located somewhere in the East it has been argued that, from the fifteenth century onwards, new knowledge of the plant life from faraway lands prompted painters to render the Garden of Eden with increasing botanical accuracy as an eastern garden stretching as far as the eye could see. The centre panel of Van Eyck’s Ghent Altarpiece is the first representation of such an open terrestrial Paradise.49 Using well-established motifs from medieval lore Van Eyck shows us the large paradisal fountain with the Pishon issuing forth, thus transforming the Pishon into St John’s river of living water: It brings not only precious stones but also life, as is indicated by the luscious vegetation on its banks (fig. 2.1). The precious stones painted near the stream of water in the right panel depicting the Hermits (fig. 2.2) can now be considered an indication of the idea that, by following the earthly rivers, the Hermits, followed by the Pilgrims, are on their way to adore the Lamb in the Garden of Delights. Travelling towards this newly opened Eden, they are about to reach long-desired Redemption.50 That the implications of the presence of gems in paradisal waters were not as enigmatic in the

42

marjolijn bol

Fig. 2.4 Dieric Bouts the Elder, Way to Paradise, after 1468, oil on panel, 115 x 69.5 cm, Lille, Palais des Beaux Arts (inv. no. P. 820)

gems in the water of paradise

fifteenth century as they appear to us today becomes clear when studying riverbanks in other Early Netherlandish paintings. After noticing an abundance of small jewels in a painting depicting St John on Patmos in the Museum Boijmans Van Beuningen, I have been on a search of my own and discovered the gems of Paradise in numerous other panel paintings from the mid-fifteenth to the sixteenth century. Significantly, the paradisal gems are always found in works by painters who are, in one way or another, indebted to the work of Jan van Eyck; what is more, all of these works incorporate the motif in a similar and meaningful way. The first painter to follow Van Eyck’s example and include the paradisal gems in his works appears to have been Dieric Bouts the Elder (c.1410-1475). The gems can be found near streams in several of his works and those of his followers. The most telling example is the left wing of a (former) triptych known as the Way to Paradise (fig. 2.4).51 In this work an angel points the elect on the day of the Last Judgement towards a large fountain in the background of a luscious paradisal landscape. Out of this fountain flow the four rivers, one of which seemingly flows out of the painting towards the beholder, much as in the Ghent Altarpiece. Significantly, the elect can be seen walking towards the fountain along these four rivers, again congruent with both the Ghent Altarpiece and the ideas postulated by medieval lore. In the foreground of the painting, Bouts depicted a variety of precious stones, this way identifying the river flowing towards the viewer as the Pishon. Bouts and his patron show that not only did they closely observe Van Eyck’s altarpiece but that they also understood the meaning behind the inclusion of the fountain with paradisal gems in such detail that they could inventively elaborate on the theme.52 That this was indeed the case can also be seen in several of Bouts’s other paintings where he included the gems of the water of Paradise. In one example, a scene from the Gospel of St John is represented. The painting shows us the moment where St John, after seeing Christ, exclaims: ‘Behold the Lamb of God’ (Ecce

43

Fig. 2.5a After Dieric Bouts the Elder, Ecce Agnus Dei, oil on panel, 73 x 56 cm, Berlin, Staatliche Museen zu Berlin, Preussischer Kulturbesitz, Gemäldegalerie (inv. no. 533c)

Fig. 2.5b Detail of fig. 2.5a, gemstones

Agnus Dei) (fig. 2.5a).53 The gems can be found near the feet of Christ on the bank of a great river winding through a vast landscape (fig. 2.5b). By means of the precious stones, here used as signs promising Salvation, Bouts again shows his ability to ingeniously combine Eyckian motifs in a meaningful way.54

44

marjolijn bol

Figs 2.6a-c (a) Hieronymus Bosch, Garden of Earthly Delights, 1500-1505, oil on panel, 220 x 389 cm, Madrid, Museo del Prado (inv. no. P02823);

(b) detail of Fig. 2.6a, left panel, fountain, gemstones;

gems in the water of paradise

45

(c) detail of Fig. 2.6a, centre panel, gemstones

Another important painting that was certainly informed by Van Eyck’s gems of the water of Paradise is Hieronymus Bosch’s (c.1450-1516) Garden of Earthly Delights (fig. 2.6a). In the left panel of this triptych, depicting Adam and Eve in the Garden of Eden, Bosch painted a large pink fountain resting on a base of precious stones (fig. 2.6b). A river flowing from this fountain can be seen winding through a mountainous landscape. In the centre panel, Bosch again painted a fountain large enough to provide water to the four paradisal rivers. Significantly, in the creek in the lower half of the panel, precious stones can be found scattered on a small sandy beach (fig. 2.6c).55 In most of the other paintings I found the gems of Paradise on the banks of rivers that wind through vast landscapes illustrating a specific set of biblical narratives. Too numerous for all of them to be discussed in detail, the gems of the water of Paradise can be found in river landscapes portraying the

following subjects:56 St John the Evangelist on the Island of Patmos,57 Adam and Eve in the Garden of Eden,58 depictions of Paradise as a terrestrial garden,59 St John the Baptist,60 St Christopher,61 Charon’s crossing of the River Styx,62 and the Apostles on their way to Emmaus.63 A painting by Joachim Patinir (c.1480-1524), Charon Crossing the River Styx, beautifully illustrates how sixteenthcentury ‘landscape’ painters use the gemstones to indicate the vicinity of Paradise on earth (figs 2.7a-b). In his painting Patinir seamlessly combines a story from ancient mythology with Christian lore. He depicts how Charon, the ferryman of the underworld, transports the souls of the dead to the gates of Hades. While on the right, Patinir depicted his vision of Hell, to the left of the river he painted a terrestrial Paradise (clearly inspired by Bosch) that includes a fountain and a large river flowing out of it. Gemstones are scattered on a small sandy beach in the foreground of

46

marjolijn bol

Figs 2.7a-b Joachim Patinir, Landscape with Charon Crossing the Styx, 1520-1524, oil on panel, 64 x 103 cm, Madrid, Museo del Prado (inv. no. P01616);

earth is abandoned, the stones likewise disappear from paintings altogether. For Van Eyck’s contemporaries, however, the gemstones scattered along the banks of rivers were signs that Paradise on earth, and hence possibly Redemption, was close. Although the real Garden of Eden was fenced off by impenetrable mountains, walls of fire, winged lions and garrisons of angels, Van Eyck and later painters afford us a visit, if only with our eyes. (b) detail of Fig. 2.7a, riverbank, gemstones

the painting. The path to this Paradise is pointed out by an angel and the gemstones on the riverbank can thus be considered signs that one is well on the way to finding it. It is significant that while the gems of the water of Paradise can be found in works produced well into the sixteenth century I have yet to find them in a single painting made in the seventeenth century. It seems that at the very moment when the idea that the Garden of Eden was a real place on

NOTES * The idea for this paper first took shape during my internship in the Museum Boijmans Van Beuningen, and I would like to express my sincerest gratitude to Dr Peter van der Coelen for this thought-provoking experience. I wish to express special thanks to Prof. Dr Jeroen Stumpel for our stimulating discussions. The present paper finally materialized at the Max Planck Institute for the History of Science in Berlin and for this inspiring fellowship my deepest appreciation goes to Prof. Dr Sven Dupré and to Dr Nadia Baadj who kindly read and commented my paper. Needless to say, remaining errors and shortcomings are my own. 1 Vanderhaegen 1872-1881, pp. 144-145: […] up een fonteijne waer af dwater schijnt te vloeijen, ende zo clear dat men die cleene steenkins in den gront ziet. […] In dander deure ziet men de heremijten ende

gems in the water of paradise

maechdekins commen uut die wonderlicke eenelicke rootsen ende heremijtaigen, daer onder uut sommighe bemoste steenen corael schijnt groeijende […]. 2 There is such a massive body of scholarship about the Fountain of Life as a literary and iconographic motif that a complete overview cannot be provided here. The most significant general studies of the motif are Underhill 1910, Underwood 1950, Miller 1986, and more recently Blacksberg 2005 and several essays in Baert 2005. The implications of the presence of the Fountain of Life on the Ghent Altarpiece are usually discussed in relation to an anonymous painting with similar subject matter and related iconography now in the Prado (School of Jan van Eyck (?), The Fountain of Grace and the Triumph of the Church over the Synagogue, Madrid, Museo del Prado, inv. no. P01511), see Panofsky 1964, pp. 216, 446-447 nn. 1-4, 439-440 n. 4; Bruyn 1957; Pemán y Pemartín 1967; Dhanens 1980; Pächt 1994, pp. 131-134 ff; Herzner 1995, p. 51 ff. (also gives a survey of literature) and more recently Borchert 2002, pp. 22, 237. For a survey of older literature, Fransen 2009, pp. 105-125 and Herzner 2011 (with an updated survey of recent literature). 3 Although the presence of these gems in other Early Netherlandish paintings has been remarked upon in a few instances, the origin of the motif in the Ghent Altarpiece and its connection to the sources discussed in the present article has been overlooked thus far. Cf. nn. 52, 53, 55. 4 Until the eighteenth century coral was believed to be a precious stone and the best corals were considered to come from India, see for instance Lexikon des Mittelalters and Theophrastus On Stones for the first occurrence of this idea. 5 ‘This is the fountain of the water of life, proceeding from the throne of God and the lamb.’ 6 Rev 22:1 and Dhanens 1973, p. 56. All English Bible citations come from the Authorized Version (King James). 7 Moreover, Erwin Panofsky points out that fountains frequently occur in other biblical passages, Panofsky 1964, vol. 1, p. 216 (and nn. 2 and 3 from that page). For the association of the octagonal font with the ceremony of baptism, see Cardon 2005, pp. 151-153. And for the centre panel of the Ghent Altarpiece referring to the Last Judgement, see Brand-Philip 1971, pp. 98-107. Also compare Van der Velden 2011b, pp. 140-141, who argues that the fountain was added with Josse of Burgundy’s baptism in mind and Van Asperen de Boer 1979, p. 194, who asserts that the fountain appears not to have been planned from the outset, as it is not underdrawn and is painted on top of much of the foliage (the latter is easy to see with the naked eye; I was able to confirm it myself during the documentation of the altarpiece in 2010). 8 Rev 21: 18-21; Panofsky 1963, p. 216 and Herzner 1995, p. 39. 9 To my knowledge this passage has not yet been brought up in studies of the Ghent Altarpiece. 10 EXULTA SATIS FILIA SYON JUBILA ECCE REX TUUS VENIT (‘Rejoice greatly, O daughter of Sion; shout [O daughter of Jerusalem:] behold, thy King cometh unto thee’); Dhanens 1973, p. 52. 11 Zech 13:1, 14:8. 12 Rev 7: 9. 13 Pächt 1994, pp. 147-148 and Sellink 2002, pp. 213-215. 14 Gen 2: 9-10. Regarding precious stones, there is another interesting parallel between the ‘old’ and ‘new’ paradise that again has not been mentioned in studies of the Ghent Altarpiece. Ezekiel, one of the four major prophets of the Old Testament and painted by van Eyck to the left of the Lamb, writes that the Garden of Eden, like the New Jerusalem, blazed with precious stones: ‘Thou hast been in Eden the garden of God; every precious stone was thy covering, the sardius, topaz, and the diamond, the beryl, the onyx, and the jasper, the sapphire, the emerald, and the carbuncle, and gold: the workmanship of thy tabrets and of thy pipes was prepared in thee in the day that thou wast created.’ (Ezek 28:13). The ‘Tree of Life’ is mentioned in Rev 22: 2. 15 Gen 2: 10-14. 16 Rotelle 2002, p. 69: ‘Now a river was coming forth out of Eden, and was watering Paradise; […] The name of one, Pishon; this is

47

the one which goes round the whole land of Havilah; there is gold there, indeed the gold of that land is the best; there is carbuncle there and leek-green stone. […]’ 17 Rotelle 2002, p. 81. 18 Scafi 2006, p. 62. 19 Although the present argument might have implications for ideas about the genesis of the Ghent Altarpiece, they will not be elaborated here. For a recent discussion concerning this topic see Van der Velden 2011a, pp. 5-39; Herzner 2011c, pp. 127-130; Van der Velden 2011b, pp. 131-141. 20 Rotelle 2002, p. 356. 21 Barney 2006, p. 285. 22 Barney 2006, p. 285. 23 Ibid. 24 Ibid., p. 281. 25 Ibid., p. 285. 26 Throop 1998, pp. 137-138. 27 Ibid. 28 Ibid., p. 137. 29 Ibid., p. 138. 30 Boese 1973, p. 351: De fonte qui est in medio paradisi […]. 31 Boese 1973, p. 351: […] Nam Phison qui et Ganges ex quodam monte Indie scaturiens, iterum incipit super terram cursum ostendere aperte defluendo, ut dicit Orientalis historia. 32 Boese 1973, p. 355: Proinde lapides, qui de terrestri paradiso ad nos per quatuor flumina deferentur, magis pretiosi sunt atque rarrisimi. Sed et alii, qui in diversis regionibus inveniuntur, pretiosi sunt ac virtute potentes, et maxime hii qui in orientis partibus sunt. Sunt autem claritate fulgoris ceteris cariores, et hoc, quia ibi sunt vapores elementorum a sordibus puriores. 33 I have used the 1485 Middle Dutch translation by J. Bellaert for the relevant passages cited here, Bellaert 1485, ll. 42059-42214: […] die fonteyne clam op ende netted dat paradijs dat gedeilt is in vier ryuiren genesi ij’ and ‘Ende van dier meer daer dat water van bouen neder valt spruten die vier ander ryuieren als physon die men ganges heit […]. And see also his book vanden ryvieren, lines 35895-35908: ‘[…] dese riuier [physon] oueruloyet mit gulden sande ende mit duerbaer ghesteynte ende in sinen oeuer wassen welrukende bomen ende oec cruden die goet ter medecinen sijn. […].’ The same data can be found in Jacob van Maerlant’s (c.12301300) Den naturen Blume (c.1266): Gysseling 1998, ll. 14855-14880. 34 Also compare Jan van Boendale’s Der leken spieghel (1330) in which poem Boendale also speaks about the paradisal fountain and the precious stones, see Vries 1844-1848, p. 78: In des paradijs pleyne, Springhet eene scone fonteine, Die soe groot es, heb ic verstaen, Datter vier rivieren af gaen. […] En daet die planteyt van desen. Men vinter oic in saphiere, Ende alderhande ghesteente diere. 35 Scafi 2006, p. 51. 36 The authoritative study on the letter and the documentation about Prester John is Zarncke 1876. 37 Ibid., p. 123. 38 Ibid., pp. 152-153 (Middle High German), p. 182 (Middle French) and p. 135 (Middle English): ‘And a 3 iourneys long fro that see, ben gret mountaynes; out of the whiehe go the out a gret flood, that comet he out of paradys: and it is futlc of precious stones, with outen ony drope of water: and it rennethe thorghe the desert, on that o sydc; so that it makethe the see gravely: […].’ Compare also to the Megenberg miniature (fig. 2.3), here the gemstones also seem to flood from the mountain without water. 39 Higgins 1997, p. 6. 40 Pollard 1900, pp. 200-201: ‘[…] Of Paradise ne can I not speak properly. For I was not there. It is far beyond. And that forthinketh me. And also I was not worthy. But as I have heard say of wise men beyond, I shall tell you with good will. […].’ 41 Ibid. 42 Ibid., pp. 240-241. 43 For an excellent study into the history of the cartography of Paradise, see Scafi 2006.

48

marjolijn bol

44 Ibid. 45 Karl-Heinz Meine 1983, pp. 17-30. 46 Scafi 2006, p. 212. 47 Van Eyck’s missions undertaken in the duke’s service have been discussed by, among others, Dhanens 1980. 48 Baxandall 1964, pp. 102-103 and n. 26. 49 Depictions of paradisal gardens first become popular in images of the Virgin in the so-called Hortus Conclusus, one of the earliest examples of which can be found in Frankfurt (Master of the Upper Rhine, Das Paradiesgärtlein, c.1410, Frankfurt, Städel Museum, inv. no. HM54). For a discussion of this topic, see Delumeau 1995, pp. 121-134. 50 Notice moreover that the water also appears to proceed to the left wing of the Ghent Altarpiece depicting the Knights of Christ as demonstrated by the wet hoofprints in the ground. 51 This painting has been thoroughly analysed in the monograph on Bouts by Catheline Périer-D’Ieteren 2006, cat. 16. 52 In an essay about Bouts’s iconography Didier Martens suggests how Bouts’s painting can perhaps be understood from the explanation of the road to Redemption by Dionysius the Carthusian (1402-1471), because Dionysius writes how the souls of the blessed await the Day of Judgement in the Garden of Eden. Martens also briefly discusses Bouts’s fascination with the Pishon because of his inclusion of its precious stones in several of his paintings, but does not connect the origin of the motif to the Ghent Altarpiece or the medieval lore discussed here, see Périer-D’Ieteren 2006, pp. 68-70. 53 Jn 1: 35-37. 54 In another famous triptych known as The Pearl of Brabant, Bouts depicted St John the Baptist on the left wing (c.1454-1462, Munich, Bayerische Staatsgemäldesammlungen, Alte Pinakothek, inv. no. WAF 77). The gems from the water of Paradise are scattered across the bank of a small perennial stream trickling from some rocks. In the background, and apparently the source of the stream at the feet of John the Baptist, we see a river flowing from a large mountain. It is possible that Bouts here refers to the idea that Paradise was fenced off by a large mountain out of which the river Pishon issues forth carrying precious gems. Compare also Hans Memling’s Allegory with the Virgin (c.1479-1480, Paris, Musée Jacquemart-André, inv. no. 1035) where the Virgin is seen guarding a mountain out of which water and precious stones flow. For the iconographical interpretation of this painting, see Baert 1997. 55 See Vandenbroeck 1989, pp. 52-53 and 1990, pp. 69-71. For a brief discussion of the connection between Bosch’s Garden of Earthly Delights and the Ghent Altarpiece, see Bax 1956, p. 68. The possible implications of the gemstone lore discussed in the present paper for the interpretation of the Bosch panels could prove a fertile avenue for further research. 56 I am still finding more examples so this list is by no means complete.

57 Follower of Dieric Bouts, Rotterdam, Museum Boijmans Van Beuningen, inv. no. 1083; Simon Bening, New York, Brooklyn Museum, inv. no. 11.505; Anonymous, Southern Netherlands, sixteenth century, Antwerp, KMSKA, inv. no. 5042; Tobias Verhaecht, St Petersburg, State Hermitage Museum, inv. no. 8694; Pieter Pourbus, Madrid, Museo del Prado, inv. no. 7757. For the iconography of gems and St John the Evangelist, also compare Voragine’s Golden Legend, see Caxton 1973, p. 176: ‘Isidore […] saith this: St. John the Evangelist transformed and turned rods of trees into fine gold, the stones and gravel of the sea into precious gems and ouches, the small broken pieces of gems he reformed into their first nature […].’ 58 Hugo van der Goes, Vienna, Kunsthistorisches Museum, inv. no. GG 5822; Jan Gossaert, Malvagna Triptych (on both side panels when closed), Palermo, Galleria Regionale de Sicilia, inv. no. 75; Jan Gossaert, Madrid, Museo del Prado, inv. no. 1930.26. 59 Dieric Bouts the Elder, Lille, Musée des Beaux-Arts, inv. no. P. 820; Hieronymus Bosch, Garden of Earthly Delights, left and centre panel, Museo del Prado, Madrid, inv. no. P 02823. 60 Dieric Bouts the Elder, Pearl of Brabant, left panel, Munich Bayerische Staatsgemäldesammlungen, Alte Pinakothek, inv. no. WAF 77; Dieric Bouts the Elder, Ecce Agnus Dei, Munich, Bayerische Staatsgemäldesammlungen, Alte Pinakothek,, inv. no. 533C; Follower of Dieric Bouts, Staatliche Museen zu Berlin, Preussischer Kulturbesitz, Gemäldegalerie, Berlin, inv. no. 15192; Juan de Flandes, Miraflores Altarpiece, centre panel, Madrid, Juan Abelló collection. Juan de Flandes also includes the gems in a painting depicting the Crucifixion, recently acquired by the Museo del Prado, Madrid. It is moreover interesting to note that although gems occur in many panels following Rogier van der Weyden’s Baptism of Christ, there are none in either of the copies of that triptych (Berlin and Frankfurt). In fact, they are not present in any of Van der Weyden’s works and in this respect the gems of the water of Paradise can be considered a typical Eyckian motif. 61 Follower of Van Eyck, Philadelphia, Philadelphia Museum of Art, inv. no. 342; Albrecht Bouts, Modena, Galleria Estense, inv. no. 320; Lieven van Lathem, Prayer Book of Charles the Bold, M 37, fol. 18, Los Angeles, The J. Paul Getty Museum; Joachim Patinir, El Escorial, San Lorenzo de El Escorial; Joachim Patinir, Antwerp, Museum Mayer van den Bergh, inv. no. 129; Quinten Matsys and Joachim Patinir, Cassel, Musée Départemental de Flandre, inv. no. 2004.4.1; Herri met de Bles, Rotterdam, Museum Boijmans Van Beuningen, inv. no. 2437. 62 Joachim Patinir, Madrid, Museo del Prado, inv. no. P01616. 63 Herri met de Bles, 1535-1540, Vienna, Kunsthistorisches Museum, inv. no. 1006; Herri met de Bles, Antwerp, Museum Mayer van den Bergh, inv. no. 40, Follower of Herri met de Bles, Rotterdam, Museum Boijmans Van Beuningen, inv. no. 2475.

Fig. 3.1 Ghent Altarpiece, Interior with City View, detail, the townscape of Sluis seen through a window

3

The Adoration of the Lamb. Philip the Good and Van Eyck’s Just Judges Luc Dequeker

ABSTRACT: Van Eyck’s Ghent Altarpiece illustrates Burgundian court ideology. It deals with millennial expectations of unity in the church after schism, restoration of the Latin Kingdom in Jerusalem, ultimate Jewish conversion, instituting God’s kingdom on earth. The painting was probably intended as a gift from Castilian merchants in Bruges to the Prinsenhof in Ghent, to celebrate Philip the Good’s marriage to Isabella of Portugal in January 1430 and the baptism in Ghent of the ardently desired successor to the ducal title, Josse, on 6 May 1432. The project had a Jewish-converso background. Evidence of this is found in significant traces of Jewish symbolism. Probable sources of Jewish influence are former rabbi and later bishop, Paul of Burgos, and his son Alonso de Cartagena. Dionysius the Carthusian is a candidate source for the mystical aspects. Theological objections and disillusionment at Josse’s early death resulted in the corrected painting being moved to the Vijd Chapel in St John’s Church, the later St Bavo’s Cathedral.

—o— To those familiar with Christian doctrine, the theology of Van Eyck’s Ghent Altarpiece might seem obvious at first sight.1 God the Father, flanked by the Virgin Mary, mother of Christ, and John the Baptist, and angels singing and playing musical instruments, presides over a celestial scene. The heavenly scene depicts the Adoration of the Lamb. Patriarchs, prophets and church members surround the Fountain of Life. The octagonal fountain and the Lamb stand for baptism and the Eucharist. Adam and Eve represent the Fall. The cross next to the Lamb underlines redemption through Christ’s sacrifice.

When closed, the altarpiece displays God’s promise of redemption. The pagan sibyls of Eritrea and Cumae corroborate the proclamations that prophets Zechariah and Micah issue. John the Evangelist, supposed author of the Revelation to John, and John the Baptist, patron of St John’s Church (later St Bavo’s Cathedral), are depicted in the lower register, next to the kneeling figures of patrician Joos Vijd and his wife Elisabeth Borluut, endowers of the Vijd Chapel in St John’s. A closer view reveals that the altarpiece’s theology is less common than generally thought. The painting has a political substratum that refers to court ideology and Philip the Good’s marriage to Isabella of Portugal in 1430.2 Despite the work’s Christian character, it has a Jewish substratum as well. Jews adoring the Lamb do not represent the Old Testament; they represent the hoped-for conversion of the Jews at the end of time. To reach both substrata, one must study the texts on the altarpiece’s panels before interpreting the iconography. Word and image are inextricably linked.3 Brief biblical citations should be read as references to their broader scriptural context and to the meaning of their themes in Van Eyck’s time and milieu. Texts painted on the panels proper bear greater authenticity than less reliable captions on the framework.

52

luc dequeker

‘Out of You Shall Come Forth a Governor for Israel’ Let us look first at the sibylline texts. The sibyl in white robe is the Sibyl of Cumae. To the right, beautifully dressed in green, is the Sibyl of Erythrae. Awkward captions on the framework inverting their names may be old, but not original. The banderole text identifies the Sibyl in green. This text is attributed to the Erythrean Sibyl in Augustine’s De Civitate Dei (XVIII, 23) and reads Rex altissimus adveniet per secula futurus scilicet in carne (‘The supreme king will come at the end of time in human form’). It refers to Christ’s incarnation and to his return at the end of time. The Sibyl of Cumae’s text, borrowed from Virgil’s Aeneid (Song VI, vv. 50-51), illustrates her visionary skills, Nil mortale sonans afflata es numine celso (‘You do not speak human language, [your language is] motivated from on high’). She gazes upward. Could there be some relation to the words De Visione Dei (‘On the Vision of God’), legible in the Virgin’s book in the depiction of the Annunciation? It may well refer to the then ongoing dogmatic discussion on the Visio beatifica. Unlike Thomas Aquinas, Dionysius the Carthusian (doctor extaticus), a compatriot of Hubert and Jan van Eyck, thought that the Virgin saw God face to face during her lifetime, as did Moses and the Apostle Paul.4 The Erythrean Sibyl might be modelled on a painting that Jan van Eyck made of Isabella of Portugal in 1429. Her right hand rests significantly on her pregnant womb.5 The depiction of the Annunciation would have fit in well with court ideology and Philip the Good’s dynastic hopes. The painting seems to apply to the duke the angel’s words addressed to the Virgin. A royal saviour (of the dynasty) will be born, adveniet rex. The townscape of Sluis, seen through the window of the room in which the Annunciation takes places, depicts the intended reference clearly.6 The duke waited anxiously for his bride in Sluis, a harbour north of Bruges (fig. 3.1). As luck would have it, she arrived there on Christmas Day, 1429.

The prophets’ banderoles are similarly constructed. Scriptural texts on the first coming of Christ are applied to the duke’s expectations for the Burgundian dynasty. The prophets declaim their texts from huge Bible codices. Zechariah, to our left, proclaims: Exulta satis filia Syon (filia Ierusalem). Jubila, ecce Rex tuus venit (‘Shout aloud daughter of Zion, daughter of Jerusalem; for see your king is coming to you’) (Zech 9: 9).7 An attentive listener hears the whole passage, a message of peace: ‘His cause won, his victory gained […] The warrior’s bow shall be banished. He shall speak peaceably to every nation’ (Zech 9: 9-10). The prophet’s message was addressed to Jerusalem and in the Gospels it applies to the Messiah’s entry into the city (Mt 21: 5; Jn 12: 14). On the altarpiece, the prophet addresses Mary (note his index finger). Micah, on the other side, also addresses Mary: Ex te egredietur qui sit dominator in Israel (‘Out of you shall come forth a governor for Israel’) (Mic 5: 2). Applying biblical texts to court ideology was not uncommon in the fifteenth and sixteenth centuries. Philip the Good’s solemn entries into Bruges (December 1440), and into Ghent eighteen years later (23 April 1458), were richly illustrated with biblical texts and mystery tableaux (called toog / togen in Dutch). The mystery tableau representing the Adoration of the Lamb, known as the Toog op de Poel, was the most important scene, honouring the Duke after the humiliating rebellion of the Ghent citizens in 1453. The iconography owed much to Van Eyck’s altarpiece.8 Like Zechariah and Micah, the open altarpiece references Isaiah. The large initial C in the open Bible on John the Baptist’s lap shows that the book is open to the first word of Isaiah, chapter 40: Consolamini. This chapter starts the second part of Isaiah’s prophecies (Isa 40-55). The whole opening paragraph, verses 1 to 11, is pertinent here: ‘Comfort, comfort my people, it is the voice of your God […]. Her [Jerusalem’s] penalty is paid […]. Cry to the cities of Judah, Your God is here’.

the adoration of the lamb

Without being cited explicitly, Isaiah may also underlie the representation of the golden halo between the imperial crown and the Lamb, right at the altarpiece’s geometric centre.9 It represents divinity and is designed as a golden disc with rays representing an aniconic manifestation or emanation of God. It constitutes the main, if not the only, source of light on the Adoration of the Lamb panel. ‘The [heavenly] city [of Jerusalem descended to earth] has no need of sun or moon to shine upon it; for the glory of God gives it light, and its lamp is the Lamb’ (Rev 21: 23). This refers back to Isaiah: ‘Arise, Jerusalem, rise clothed in light; your light has come and the glory of the Lord (kabod Adonai) shines over you […]. The sun shall no longer be your light by day, nor the moon shine on you when evening falls; the Lord shall be your everlasting light, your God shall be your glory’ (‫)לתפארתך יהוה כבוד‬ (Isa 60: 1, 19). Isaiah’s key word, ‫ – תפארת‬Tiferet, can be read on the red hat with Hebrew characters worn by a worshipper on the far side of the group to the left of the altar (fig. 3.2). It seems as if the golden halo projects the Hebrew letters onto the hat. Only the letters on the front are lit. Tiferet characterizes the whole group of Jews adoring the Lamb (see Lk 2: 29-32).10 ‘Praise and honour to him who sits on the throne and to the Lamb’ The theology of the Revelation to John (or Apocalypse) which is displayed along the altarpiece’s vertical axis, substantiates the altarpiece’s fundamental unity. Upper and lower registers, including the wing-panels, belong together (fig. 3.3). The main theological points of interest – crown, golden glory, Lamb on the altar, Fountain of Life – are situated at intentionally geometric points in the work. The honorary title sabaot on the Deity’s chest strap corresponds to the name of God – adonai (yhwh) sebaot – in the Hebrew Bible, Pantocrator in Greek. The letters A and O, the latter written as a Greek omega, refer to Christ’s solemn claim to divinity at the beginning and end of

53

Fig. 3.2 Ghent Altarpiece, Adoration of the Lamb, detail, Hebrew mark of honour for Jewish converts

Revelation: ‘I am the Alpha and the Omega […] who is, and who was, and who is to come, the Beginning and the End, the Almighty’ (Rev 1: 8, 21: 6, 22: 13). Van Eyck most certainly referred here not to God the Father but to the glorified Messiah at the end of time. The combination of the monogram ihesus xps on the throne’s cloth of honour and the emblem of the pelican is decisive, duly confirmed by the overpainted original ihc + xps.11 As in Revelation there is no reference to the Trinity. The dove in the golden glory, presumed to represent the Holy Spirit, is an additional, later interpretation.12 The title sabaot further evokes Isaiah’s vision in which seraphs proclaim, ‘Holy, holy, holy is the Lord of Hosts (Adonai / YHWH Sebaot) (Isa 6: 3), quoted in Revelation as an identifier of the Pantocrator, Jesus Christ, ‘Holy, holy, holy Lord Almighty (παντοκράτωρ Pantocrator), who was and is and is to come’ (Rev 4: 8). The text on the Pantocrator’s mantle, Rex Regum et Dominus Dominantium, pays tribute to the Lamb after its epic victory over Babylon: ‘He is Lord of lords and King of kings, and his victory will be shared by his followers, called and chosen and

54

luc dequeker

Fig. 3.3 The earthly City of God. Three-dimensional presentation of the Ghent Altarpiece (author’s photo arrangement)

faithful’ (Rev 17: 14). Again after its final victory over the scarlet beast with seven heads and ten horns: ‘He is robed in a garment drenched in blood […]. On his robe and on his thigh there is written the name: King of kings and Lord of lords’. (Rev 19: 13, 16).13 The crown of the last world emperor lies at the feet of the victorious and glorified Messiah to confirm this tribute. Van Eyck portrayed the Pantocrator as the Judge at the end of time. He holds a crystal sceptre, not the sharp two-edged sword of Rev 1: 16, 19: 15. The message is not threatening, as apocalyptic images normally are. The Ghent Altarpiece, faithful to the authentic meaning of Revelation, does not

deal with the world’s ruin, but with the full realization of the Kingdom of God on earth at the end of time. The altarpiece looks at salvation history from an eschatological perspective. Notwithstanding iconographical differences, the presentation of the Pantocrator is consistent with traditional Maiestas Domini images of God seated on a throne, the right hand raised in blessing, the left holding the Book of Life, inscribed with Alfa and Omega. Van Eyck’s Pantocrator wears the papal tiara of the summus pontifex who stands against the Antichrist. The iconography is also related to Eastern Orthodox Deisis (or ‘humble petition’) imagery.

the adoration of the lamb

The crowned Virgin and John the Baptist, both sitting on thrones with cloths of honour, figure as intercessors before the Judge at the end of time. Their gestures are not just imploring. They have completed their intercessory tasks. Mary now offers psalms of thanksgiving, while John uses Isaiah to proclaim redemption. The Virgin’s crown has a threefold symbolism. Gemstones at the bottom refer to her royal title. Lilies and roses in the centre adorn the Bride of Christ. As ‘daughter Zion’ she represents New Jerusalem, coming down from heaven adorned as a bride (Rev 21: 2). Twelve stars at the top represent the beleaguered woman of the Apocalypse: ‘A great sign appeared in heaven: a woman clothed with the sun, the moon under her feet and on her head a crown of twelve stars’ (Rev 12: 1). The Baptist wears a green mantle above his rough, desert camel’s hair garment. It recalls the prophet’s mantle, transmitted by Elijah to his disciple Elisha, as token of his mission (2 Kings 2). Jesus himself identified the Baptist as the prophet Elijah who at the end of time would come to intercede and convert the faithful, as prophesied by Malachi (2 Kings 2: 11; Mal 3: 23-24, Mt 11: 13-14). Inscriptions in the decorative arches above the Deisis group confirm the Maiestas Domini as Van Eyck understood it. The inscription above the Pantocrator’s throne reads: Hic est Deus potentissimus propter divinam maiestatem. Summus omnium optimus propter dulcedinis bonitatem. Remunerator liberalissimus propter immensam largitatem (‘This is God Almighty, Divine Majesty. Most high. Excellent in lovely goodness. Supreme above all. Generous Retributor, with immense forgiveness’). The arches above the Virgin’s throne display a text from the Wisdom of Solomon (Sapientia): Hec est speciosior sole et super omnem stellarum disposicionem luci comparata invenitur prior – Candor est enim lucis eterne et speculum sine macula Dei (‘She is more radiant than the sun and surpasses every constellation […]. She is the brightness that streams from everlasting light, the flawless mirror of the active

55

power of God’) (Wis 7: 29). The text honours the Virgin as Sedes Sapientiae, seat of divine wisdom. It may correspond to the Visio Dei in the book on which Mary has been meditating in the Annunciation panel. The arch above the Baptist displays a non-biblical text that defines John’s mission. It marks the transition from ‘Law and Prophets’ to the Gospels and the Epistles, Hic est Baptista Joannes, Major homine, Par angelis, Legis summa, Evangelii satio, Apostolorum vox, Silentium prophetarum, Lucerna mundi, Domini testis (‘This is John the Baptist, greater than man, equal to the angels, summarizing the Law, seed of the Gospels, Apostle’s voice, silence of the Prophets, lamp of the world, witness of the Lord’). Reading down the vertical axis of the altarpiece to the lower register we first meet the golden glory, whose apocalyptic background has already been mentioned. Does the Lamb on the altar also have an apocalyptic background? In the Christian-Catholic liturgy the reading for All Saints Day, 1 November, is on the adoration of the Lamb by the tribes of Israel (Rev 7: 2-12). But which Lamb is celebrated on Van Eyck’s altarpiece: the Paschal Lamb (Ex. 12: 1, 1 Cor 5: 7) or the Ram of Revelation? The Baptist’s witness on the altar is evident: Ecce Agnus Dei qui tollit peccata mundi (Jn 1: 29). This is the Passover Lamb, icon of Christ’s sacrifice. But if we were to apply a hermeneutic interpretation could we not discern the Ram of Revelation (άρνιον) behind the Passover Lamb (άμνός)? The Latin translations, which guide modern authors, use the word agnus for both, which obliterates the distinction. The very fact that Van Eyck situates the Lamb’s altar on a mound in a panorama of Jerusalem, with the Temple on Mount Zion, shows that he intended the Revelatory rendering: ‘Then I looked, and on Mount Zion stood the Lamb, and with him were a hundred and fortyfour thousand’ (Rev 14: 1). In fact the Lamb on the altarpiece has been redrawn. Below the ears of the present Lamb we can see the smaller ears of an earlier image. Is some

56

luc dequeker

older image hidden underneath? If so, it would mean that Van Eyck’s Adoration of the Lamb refers, perhaps primarily, to the investiture of the Lamb in Revelation 5: 1-14, not just to the liturgy in Revelation 7: 10 (‘Victory to our God who sits on the Throne, and to the Lamb’). The main biblical references would be Revelation 5: 5 (‘Do not weep; for the Lion from the tribe of Judah, the Scion of David, has won the right to open the scroll and break his seven seals’) and Revelation 5: 12-13 (‘Worthy is the Lamb, the Lamb that was slain, to receive all power and wealth, wisdom and might, honour and glory and praise […]. Praise and honour, glory and might, to him who sits on the throne and to the Lamb for ever and ever’). The best parallel to the imagery would be the miniature in a Carolingian bible, illustrating the seven churches of the Apocalypse, with the Lamb on the altar breaking the seals, as the four Evangelists unveil Moses’ face.14 Whichever it may be, it is clear that the Lamb on Van Eyck’s altarpiece represents the suffering and victorious Christ, whereas the Pantocrator represents Christ, the Messiah, glorified, Judge at the end of time. The third main point of interest on the lower part of the vertical axis is the Fountain of Life. The inscription on the stone basin’s edge reads: hic est fons aque vite procedens de sede dei + agni (‘This is the source of life-giving water flowing from the throne of God and the Lamb’). The text refers to various passages in Revelation about the ‘water of life’ in New Jerusalem. Referring to the elect standing before the throne of God, we read ‘The Lamb who is at the heart of the throne will be their shepherd and will guide them to the springs of the water of life’ (Rev 7: 17). Then, in Christ’s self-revelation, is ‘I am the Alpha and the Omega, the beginning and the end. A draught from the water-springs will be my free gift to the thirsty’ (Rev 21: 6). Finally, the description of Heavenly Jerusalem notes ‘Then he [the angel] showed me the river of water of life, sparkling like crystal, flowing from the throne of God and of the Lamb’

(Rev 22: 1). These texts allude to the prophet Ezekiel’s vision of a stream of life-giving water flowing from Jerusalem’s restored temple (Ezek 47: 1-12) as well as to Zechariah 14, which refers to the Jewish Feast of Tabernacles (Sukkot), a passage pertinent to Van Eyck’s painting: ‘On that day [of yhwh] living water shall issue from Jerusalem […]. Then the Lord shall become king over all the earth […]. The nations shall come up year by year to worship the King, yhwh sabaot, and to keep the pilgrimfeast of Tabernacles’ (Zach 14: 8, 9, 16). On the Jewish Feast of Tabernacles, Jesus proclaimed in the temple that ‘If anyone is thirsty let him come to me; whoever believes in me, let him drink.’ As Scripture says, ‘Streams of living water shall flow out from within him’ (Jn 7: 37-38). Next to Ezekiel’s vision, Van Eyck’s Fountain of Life might also allude to the rivers of Paradise (Gen 2: 10-14).15 Precious stones in the waters of Paradise surrounding the basin may refer to Thomas of Cantimpré’s Liber de natura rerum.16 Moreover, the pile of gemstones to the left of the basin may refer to the idea that the Jewish Talmudic tradition holds precious values hidden like gems among worthless customs.17 There are those who deny the authenticity of the Fountain of Life as an original part of the painting for there is no underdrawing. My belief, however, is that the Fountain of Life is essential to the altarpiece’s theological programme. The fountain’s architectural form is irregular. In its present form, the basin is octagonal, whereas the brass fountain suggests a hexagon. Six splashes on the water’s surface reveal six lower spouts with another four above them. Add to that the two flasks that the angel at the top is pouring and we have twelve jets of water. The geometrical analysis of the altarpiece allows us to restore the basin’s correct perspective. Let me give a few examples of the numerous hexagonal fountains and pedestals – seats of dignity – in fifteenth-century painting and sculpture. Regarding the Virgin’s coronation and Song of Songs, the tapestry entitled Vierge Glorieuse (200 ≈

the adoration of the lamb

280 cm, Brussels, 1485. Paris, Musée du Louvre), shows the Madonna as Hortus Conclusus and fount of life.18 A miniature at the Feast of Tabernacles in the Duke of Alba Bible (Toledo; Madrid, 14221433, Duke of Alba priv. coll.), translated from the Hebrew into Castilian by Moses Aragel, includes a hexagonal fountain of life located in the Jerusalem temple and illustrating Sukkot.19 Among the many hexagonal pedestals or bases are Claus Sluter’s early-fifteenth-century Well of Moses or Great Cross at Champmol and the recently restored Descent from the Cross triptych attributed to Bernardo Martorell (Barcelona, first half of the fifteenth century; Lisbon, Museu Nacional de Arte Antiga), which shows the rock of Golgotha in the form of a roughly hexagonal pedestal. Judaism sees the millennium as a thousand-year period that follows upon the world’s six-thousandyear existence. This corresponds to the six days of creation. Hence the term ‘chiliasm’ or ‘millenarianism’, the doctrine of or belief in a future (and typically imminent) thousand-year age of blessedness, beginning with or culminating in the Second Coming. On the seventh day, God Almighty completes creation forever. Van Eyck evokes six thousand years of the world’s existence in the hexagonal Fountain of Life. It represents the created cosmos in contrast to chaos. The hexagon is a foundation for the heavenly kingdom on earth. The altarpiece shows the seventh millennium as the one in which Jews, with the seal of God on their foreheads, and gentiles of every nation and tribe (Rev 7: 3, 9) adore the Lamb. Pilgrims and soldiers on their way to Jerusalem complete the image. With Jerusalem in the background, just below and parallel to the upper register’s heavenly figures, the altarpiece illustrates the ever-present, ultimate kingdom of God. Starting off as a hexagonal pedestal alluding to the millennium, Van Eyck’s Fountain of Life was reinterpreted into an octagonal baptismal font.20 In the Revelation to John, millennium expectations coincide with the celebration of the Lamb’s holy wedding: ‘Alleluia! The Lord our God, sover-

57

eign over all, has entered on his reign! Exult and shout for joy and do him homage, for the weddingday of the Lamb has come! His bride has made herself ready, and for her dress she has been given fine linen, clean and shining […]. Happy are those who are invited to the wedding-supper of the Lamb!’ (Rev 19: 6-9). The first human couple, Adam and Eve evoke a holy wedding. The Singing and Musician Angels situated between man’s ancestors and the Blessed Virgin represent the Song of Solomon (Canticles). Their melodies evoke the Fountain of Life, the bridegroom addressing his beloved: ‘My sister, my bride is a garden close-locked, a fountain sealed’, the bride responding: ‘The fountain in my garden is a spring of running water’ (Song 4: 12, 15). The Wisdom of Solomon (Sapientia) lies open on the singer’s lectern. The text on the arch above the Virgin quotes this book. Carved figures adorn the lectern. They are obviously sages and may even be Moses ben Maimon (Maimonides) and Averroes (Ibn Rushd). The City of God on Earth A four line inscription in front of the Pantocrator’s throne lists the characteristics of the millennial City of God shown in the designated altarpiece panels (fig. 3.4).21 vita sine morte in capite + iuventus sine senectute in fronte gaudium sine merore a dextris + securitas sine timore a sinistris Life without death on top + Youth without aging in front Joy without sorrow to the right + Security without fear to the left. Directions a dextris and a sinistris refer to the Pantocrator’s proper right and left. The left side is the side of condemnation, the right the place of election. What is depicted at the top means life, not death. The front deals with eternal youth. Left

58

luc dequeker

Fig. 3.4 Ghent Altarpiece, Deity Enthroned, detail, lettered step, the altarpiece’s iconographic programme

proclaims security and the end of fear. The right shows happiness that overcomes sadness. The first motto, Vita sine morte in capite, applies to Adam and Eve, as well as to Cain and Abel in the lunettes. Normally they signify death and murder. That is why their framework captions read Adam nos in mortem praecipitat (‘Adam pushed us to death’) – Eva occidendo obfuit (‘Eve caused damage through death’). In medieval iconography, Cain and Abel usually designate antagonism between Church (Abel) and Synagogue (Cain). Van Eyck goes beyond enmity and original sin. The captions on the framework modify significantly the representation’s meaning. They cannot be original. Indeed, when opening the small upper doors of the Annunciation panel, humanity’s ancestors appear immediately to the left and right of the Pantocrator as icons of redeemed humanity at the end of time (fig. 3.5). Moreover, Cain and Abel are in the group of redeemed Jews on the right side of the altar. Both are bareheaded, Cain (with black hair) carries God’s sign of protection on his forehead (Gen 4: 15).22 Adam’s gesture is also significant. It has nothing to do with temptation, sin or shame. Adam does not reach for the forbidden fruit, nor to his throat where, according to legend, the forbidden fruit became stuck. Rather, he brings his left hand deli-

cately to the place that the rib from which Eve was sculpted once occupied. According to medieval mystical theology, this is a symbolic reference to the blood and the water flowing from the wound of the crucified Christ (Jn 19: 34). See the pierced side of the Lamb, foreshadowing church’s ‘birth’. Eve is depicted as pregnant. According to mystical theology, she is the mother of the early Christian church, i.e. the Jews who first recognized Jesus as the Messiah. In a delicate gesture of love she holds the sign of salvation in her hand (fig. 3.6). It is a special yellow citrus fruit with rough skin, known in Judaism as the etrog for the Feast of Tabernacles (Lev 23: 39-40). Christian sources, at least from thirteenth century onwards, allude to the fruit as pomus adami (Adam’s apple / Paradise apple), the oldest attestation being found in Thomas of Cantimpré’s Liber de natura rerum (c.1245). Under the authority of Jacques de Vitry, Bishop of Acre in Palestine (d. 1240), Thomas describes the pomus adami as a yellow ‘apple’ with rough skin, in which one could find the bite-marks left by Adam after eating the forbidden fruit.23 As I show below, the same yellow fruit is in the group of Jewish worshippers, where it is one of the four species used in celebrating Sukkot or Tabernacles. No Christian sources relate

the adoration of the lamb

Fig. 3.5 Ghent Altarpiece, redeemed humanity. Opening the altarpiece, first step (author’s photo arrangement)

59

60

luc dequeker

Fig. 3.6 Ghent Altarpiece, detail, Eve with the etrog, the fruit of redemption

the pomus adami to the etrog used in celebrating Tabernacles. Van Eyck’s theological programme obviously knew the connection. The Jewish Aramaic Bible (the Targum) connects the etrog to the ‘apple of love’ in the Song of Solomon (Song 2: 3).24 Dionysius the Carthusian, who may have been involved in the altarpiece’s theological programme, knew of the tradition from Alexander of Hales.25 Van Eyck replaced the traditional forbidden fruit of the Fall with the Sukkot etrog: Vita sine morte in capite. The next motto, Iuventus sine senectute in fronte, refers to the hexagonal Fountain of Life discussed above. After six thousand years, the universe will not be old, human history will not end. Creation

will be renewed and confirmed in the next millennium, which will be marked by ‘A new heaven and a new earth […]. The holy city, new Jerusalem, coming down out of heaven from God’ (Rev 21: 1-2). The third motto, Securitas sine timore a sinistris, is intentionally written on the right side of the panel.26 This is Judge’s left side, the ‘sinister’ side of condemnation. The motto inverts the traditional meaning. Here it refers to the group of Jews on the panel’s left side, i.e. on the Pantocrator’s proper right (fig. 3.7). It runs contrary to the traditional position of the Jews on the sinister (left) side at the Last Judgement, as depicted, for instance, in the fifteenth-century panel now in the Museo del Prado entitled The Fountain of Grace and the Triumph of the Church over the Synagogue (originally Segovia).27 The Spanish panel is a flagrant repudiation of Van Eyck’s presentation of redemption and the Adoration of the Lamb. Three leading figures, king, priest and prophet, mark the group as Jewish. They stand in a straight diagonal line, mirroring the three popes on the opposite side. They represent the Jewish nation’s three pillars, rather than particular individuals. The first is King David, author of the Psalms. A golden ray, sign of divine inspiration, touches his mouth. The priest in the middle is dressed in a white robe with a laurel wreath on his head, as is fitting for the Jewish Day of Atonement. Because of the wreath, this figure is often erroneously associated with the Roman poet Virgil. However, he is the archisynagogos, the high priest representing the Jewish nation. La Fuente de Vida shows the archisynagogos as blind and humiliated. Van Eyck depicts him with Jewish sidelocks and traces of blood under the laurel wreath. Like Eve he holds the yellow citrus fruit, the etrog (here with twig and leaf), sign of redemption. The prophet in blue next to him holds a myrtle branch. The man in green behind has a sprig of river willow. These four species refer to the eschatological Feast of Tabernacles, more precisely to Zechariah 14: 16, cited above in relation to the Fountain of Life. In its description

the adoration of the lamb

61

Fig. 3.7 Ghent Altarpiece, Adoration of the Lamb, detail, Ecclesia ex Circumcisione

of the New Jerusalem, Revelation reprises Zechariah’s prophecy: ‘There will be no more curse’ (Zech 14: 11). The throne of God and the Lamb will be there and his servants shall serve Him’ (Rev 22: 3). Kneeling figures next to the Fountain, presenting four open books, represent the Jewish ‘tradition’, the transmission of the Law (the oral Torah) in four steps.28 Above the kneeling group stands St Jerome, explaining the Torah to Jews wearing blue caps. The whole scene on the left side of the Adoration of the Lamb panel shows Jewish conversion as a necessary preparation for Christ’s return to earth at the end of time. This was a major issue in Spain during the first half of the fifteenth century. Van Eyck goes beyond the centuries-old hostility between Church and Synagogue. He returns to the early Christian theme of close relationship within the

universal church of the ecclesia ex circumcisione (circumcised [i.e. originally Jewish] Christians) and the ecclesia ex gentibus (Christians from [other] nations’). The following motto, Gaudium sine merore a dextris, applies to the church on the right (fig. 3.8). The most distinctive figures in the group are the three popes in the foreground. Like the Jewish pillars, they, too are arranged in a straight line. They stand for the restoration of unity in the church after the Western Schism (1378-1417). Two popes pray together from a liturgical book: Martin V, the pope of reconciliation, elected at the Council of Constance in 1417, and John  XXIII (sic), the antipope of Pisa, who was ultimately reconciled with Martin V. The pope at the back, with a sharp face, is probably Benedict XIII, the last antipope of Avignon. He is known for his consistent and radical effort during the years 1413-1415

62

luc dequeker

Fig. 3.8 Ghent Altarpiece, Adoration of the Lamb, detail, Ecclesia ex Gentibus

for Jewish conversion in preparation for the end of time. His presence in Van Eyck’s altarpiece underlines the need to convert the Jews before the millennial kingdom can begin. The kneeling apostles mirror the kneeling Jewish representatives of the oral Torah. The apostles are entrusted with transmitting the Gospel of Christ to future generations. Unexpectedly, the group includes Judas, as well as his substitute Matthias (Acts 1: 26).The fifteenth-century theology of universal redemption accepted Judas’s ultimate repentance.29 Thomas Aquinas stands at the top, across from St Jerome. Both are touched by a golden ray, a sign of divine inspiration. Thomas Aquinas and St Jerome were authoritative masters to Spanish converts. In the middle distance, more servants of the Lamb approach the altar in two groups. They rep-

resent the immense multitudes, from all races and tribes, peoples and tongues that joined the tribes of Israel as members of the universal church in the City of God (Rev 7: 9). Next to themes of universality, restored unity in the church, and Jewish conversion, the liberation of Jerusalem was a prerequisite condition for the arrival of the millennial messianic kingdom on earth. Crusaders and pilgrims shown heading for Jerusalem on the altarpiece’s wing-panels aim at the ultimate liberation of the Holy City and the restoration of the Latin kingdom in the Holy Land. The heavenly heralds with cross-bearing banners, milites Christi (Knights of Christ), evoke the crusade’s device: Dieu le veut! Contrary to common opinion, Hubert and Jan van Eyck are not shown among the dynastic figures on the Iusti Iudices (Just Judges) panel. Rather, three Burgundian dukes are

the adoration of the lamb

depicted, Philip the Bold, John the Fearless and, dressed in black, Philip the Good. 30 History is well aware of their interest in a new crusade. Apparently, Burgundian court ideology thought the millennial empire would be realized in the duchy of Philip the Good.31 Two additional cityscapes in dark colours on the horizon apply the apocalyptic vision of the New Jerusalem to the Burgundian ideology. At the right we see the city of Ghent, with the gothic St John’s Church still under construction and the silhouette of Prinsenhof, the duke’s newly restored residence (fig. 3.9). To the left, we see Utrecht’s fourteenthcentury Dom tower, representing Philip the Good’s political interests in the diocese of Utrecht. From Ducal Court to Joos Vijd’s Memorial Chapel – A New Hypothesis Van Eyck’s millennium altarpiece centres around the ultimate redemption of the Jews and on Burgundian court ideology. The idea came apparently not from Philip the Good’s entourage, but from Castilian merchants in Bruges, to whom, in Octo-

63

ber 1428, the duke had granted the privilege of a local consulate.32 While Dionysius the Carthusian may have been involved in the painting’s mystical aspects, the real candidates for sources of the fifteenth-century Jewish and millennium programme are two Castilian converts. They are Rabbi Salomon Ha Levi (later Bishop Paul of Burgos [d. 1435]), and his son Alonso de Cartagena (d. 1456), who succeeded his father as bishop.33 Alonso’s Defensorium Unitatis Christianae, published with the Council of Basel (1434-1439) in mind, emphatically claims that Jewish converts were equal to any Christians from (other) nations. Alonso may have had contact with Jan van Eyck in Burgos in 1428, at the court of John II of Castile. I believe that the altarpiece may have been intended as a gift to Philip the Good on the occasion of his marriage to Isabella of Portugal (1429-1430), and the solemn baptism (in Ghent) of their second son Josse on 6 May 1432. This date is mentioned in the altarpiece’s dedicatory inscription (quatrain) in memory of the child who would die prematurely on 21 August 1432.

Fig. 3.9 Ghent Altarpiece, Adoration of the Lamb, detail, Ghent cityscape

64

luc dequeker

Had it been accepted, the altarpiece would have been placed in the restored chapel of the Prinsenhof.34 Disappointment brought on by the death of the eagerly hoped for Burgundian successor, theological objections to the project’s presumed chiliastic overtones and paradisiacal enthusiasm, perhaps even some degree of anti-Judaism at court, may all have contributed to the duke and duchess’s failure to accept the work.35 Ghent patrician Joos Vijd, who was familiar with the duke and his court, acquired the altarpiece for the memorial chapel he endowed in 1435 in St John’s Church – the later St Bavo’s Cathedral. Van Eyck’s millennium painting underwent significant modification to adapt it to traditional medieval salvation history. N OTES 1 Schmidt 2005. 2 Jolly 1987. 3 De Baets 1961. 4 Swenden 1960. 5 The cryptic title on the Erythrean Sibyl’s body, Meiaparos, written at the end with a Greek Σ, may refer to Isa. 7: 14 (Dequeker 2011, pp. 15-16). 6 Dequeker 2011, Appendix IV. 7 Bible texts from New English Bible (Oxford/Cambridge, 1970). 8 De Baets 1961, Annex pp. 612-614. 9 Geometric analysis, Dequeker 2011, pp. 2-6; Appendix I. 10 Hebrew-like characters on the cloths of honour on the thrones of the Virgin and the Baptist are illegible. Further examination of the applied brocade might solve the problem. Geelen, Steyaert 2011. Geelen 2012. 11 Geelen, Steyaert 2011, p. 388. 12 Van Asperen de Boer 1979. Reinterpretation already seen in the Toog op de Poel (1458). 13 On the structure of the Revelation to John, Dequeker 2011, Appendix III. 14 Paris, St Denis. Now in the Abbazia di San Paolo fuori le Mura, Rome. Van der Meer 1978, fig. 46. Dequeker 2011, fig. 14.

15 Van der Meer 1938, pp. 64, 173. 16 Thomas Cantimprensis 1973. Translated by Jacob Van Maerlant, Der Naturen Bloeme (Stenenboek): Ghemenelike sijn best die stene / Die ons comen groet ende clene / Bi den rivieren neder te warnen / Uten paradyse ghevaren; / Stene die comen van Orient / Die sijn verren wel bekent (Verwijs 1980). 17 Thus Raymundus Martinus (c.1280), Prooemium, V and IX. Paulus Burgensis 1435, Prologus. 18 Dequeker 2011, pp. 57-58. 19 Dequeker 2011, pp. 107-108; Nordström 1967, fig. 127; Sterling 1967, pp. 36-37. 20 Dequeker 2011, fig. 16. The wrong reading HONI instead of AGNI at the end of the inscription (De Baets 1961, p. 600) may be due to redrawing. 21 Dhanens, 1965, p. 59 and Van Asperen de Boer 1979, p. 178 justify the authenticity of the inscription, notwithstanding the repentir. 22 The sign on Cain’s forehead is original. Christian tradition interpreted Cain’s sign as a condemnation. Biblical and Jewish traditions understand the sign as a mark of God’s protection. Ita Raimundus Martinus (c.1280) ), fol. 646-647. 23 Thomas Cantimpratensis 1973. Translated by Jacob van Maerlant (c.1300): Arbor Ade, dats bekent, es een boem in Orient / Die Adaems boem heet bedi, als ons seghet Jacob van Vetri / Omdat si appele draghen vele, schone ende van vaerwen ghele /Ende elc appel met ere beten, so dat men sienmach ende weten / Dat God toghet in aertrike, Adaems sonden wel dorperlike (Verwijs 1980, pp. 197-206). James Snyder 1976. 24 Dequeker 2011, p. 67. 25 Dequeker 2011, p. 67. Dionysius Carthusianus, Opera 7, 345 C. 26 The underdrawing shows the mistaken position, which the painter corrected. 27 Fransen 2009. 28 Mishnah, Proverbs of the Fathers I: 1: ‘Moses received the Torah from Sinai and committed it to Joshua, and Joshua to the Elders, and the Elders to the Prophets; and the Prophets committed it to the men of the Great Synagogue.’ 29 According to famous repentance preacher Vicente Ferrer O.P. (Valencia c.1350 – Vannes 1519), a confident of Pope Benedict XIII, Christ on the Cross would have forgiven Judas’s repentant soul, liberated from his body after hanging. Brettle 1924, pp. 41, 44. Dequeker 2011, pp. 128-131. 30 Post 1922, pp. 120-125. 31 Jolly 1987. 32 Vandewalle 1994. 33 Serrano 1942. 34 Lievois 2000. 35 Chiliasm was never condemned formally by the church, only considered ‘dangerous’ under specific circumstances.

Fig. 4.1 Paul Coremans

4

‘Revenons à notre Mouton’. Paul Coremans, Erwin Panofsky, Martin Davies and the Mystic Lamb Hélène Dubois, Jana Sanyova and Dominique Vanwijnsberghe

ABSTRACT: In 1953 two major books came out that were to contribute greatly to the understanding of the Ghent Altarpiece – Paul Coremans’s L’Agneau Mystique au laboratoire and Erwin Panofsky’s Early Netherlandish Painting. Before becoming the best of friends, the scientist and the scholar had learned to know and appreciate each other. They soon realized that their different approaches – technical and art historical – were complementary. To bring about a close collaboration, they set out to organize in Brussels a seminar entirely devoted to the polyptych. It gave a team of leading experts ample opportunity to discuss the new findings and interpretations using laboratory documents and scientific imagery, before examining the altarpiece in situ. The results of these interdisciplinary moutonnements were carefully recorded but never published. They are nevertheless an inspiration and food for thought for all those who are currently involved in the research and the treatment of the Ghent Altarpiece.

—o— The year 1432, mentioned in the quatrain on the frames of the Ghent Altarpiece, is of paramount importance for the history of the polyptych, as scholars have repeatedly shown. In the more recent past another date can be associated with a turning point in its critical appraisal. In 1953, within the space of a few months, two major books appeared that were to contribute greatly to the understanding of the polyptych. The first, L’Agneau Mystique au laboratoire,1 was produced by an interdisciplinary team led by Paul Coremans (fig. 4.1), the director

of what would later become the Royal Institute for Cultural Heritage, more familiarly known from its Dutch and French acronyms as the kik-irpa.2 It describes the examination and conservation treatment of the panels carried out in 1951 and reflects the collaboration with the international advisory commission. It was one of the very first publications devoted entirely to the technical analysis of a masterpiece and, as such, it served as a model. The second publication, Early Netherlandish Painting,3 was written by one of the most influential art historians of his time, Erwin Panofsky (fig. 4.2). In a long chapter entirely devoted to the Ghent Altarpiece,4 Panofsky reveals his revolutionary views on the polyptych with a particular focus on the complex evolution of its iconographic concept. Coremans and Panofsky actually knew each other and even shared their findings before the publication of their respective works. Their views were fairly compatible and over the years they developed a close friendship. They were both very conscious of the limits of their investigative methods and acknowledged that they were far from solving all the questions raised by the altarpiece. They were aware of the provisionality of their conclusions and were convinced that they would make considerable progress if they united forces and continued the discussion with other experts. This is

68

hélène dubois, jana sanyova and dominique vanwijnsberghe

Fig. 4.2 Erwin Panofsky (M.A. Holly, Panofsky and the Foundations of Art History, 1984)

what inspired them to organize what they called a ‘symposium’, a series of scholarly meetings entirely devoted to the Ghent Altarpiece. This paper is devoted to those sessions, the records of which were never published.5 It is also meant as homage to the scholars who were involved in this fascinating chapter of research on the Ghent Altarpiece. A large number of records that are used here are kept in the archives of kik-irpa, currently being catalogued by Dominique Deneffe. These include many of Panofsky’s letters to members of the Institute and, above all, to his ‘dear friend and colleague’ Paul Coremans.6 It all starts with a letter from Panofsky to Coremans, dated June 13, 1949.7 Panofsky has consulted Coremans’ recent monograph on the Ghent Altarpiece.8 It is in the first place a ‘picture book’ with dozens of photographic details of the altarpiece taken after its return from the salt mines of Altaussee in 1945. Panofsky is puzzled by the Mediterranean vegetation in the background of the Adoration of the Lamb panel and wants to know if it is ‘consistently superimposed (on the paint layer)

throughout, or only party superimposed, or partly not’. Coremans’ reply is brief: it would take weeks of continuous work to answer this apparently simple question.9 Panofsky also enquires about the possibility that the upper side of the Adoration panel might have been cut, as suggested by Beenken.10 One feels a certain unease on the part of Coremans. Obviously he was not acquainted with his celebrated American colleague’s bibliography and he asks for a complete list of publications, ‘to be sure of knowing completely your scientific work on this subject.’11 The ice is broken between the scientist and the scholar after Panofsky’s visit to Brussels in 1951. In a letter to Carl Nordenfalk,12 Panofsky reports that Coremans presented him with the latest results of his investigations. From then on, the two men get to know and appreciate each other. Panofsky – Pan to his friends – returns to Brussels in 1952 and meets the team that has been working on the restoration of the Ghent Altarpiece. In a lyrical letter addressed from Sweden to René Sneyers, Coremans’ right-hand man, Panofsky uses for the first time what would become the code name of the altarpiece: the ‘Mouton’ (fig. 4.3).13 In 1953, both Coremans’ and Panofsky’s manuscripts are about to be sent to press and Panofsky invites his Belgian colleague to Princeton to lecture on his ‘startling revelations’ as he calls them, probably to anticipate any nasty surprises since his book is due to come out after Coremans’: I pray to God that you have not made any discoveries in the meantime which may explode my new theories just as your previous findings did my other ones.14 Coremans spends two months in the United States, giving no less than twenty-one lectures in eight different locations! On 12 March he gives a paper at the University Museum of Art in Princeton and is invited to the Panofsky’s home, where he meets other scholars. From now on, the ‘Dear Friend and Colleague’ simply becomes ‘Dear Paul’.

revenons à notre mouton

69

Fig. 4.3 First mention of the ‘Mouton’ in a letter from Erwin Panofsky to René Sneyers, dated 14 August 1952

L’Agneau Mystique au laboratoire rolls off the press a few weeks later and is sent right away to Princeton, with a letter inviting Pan to come to Belgium to re-examine the Ghent Altarpiece with the support of the Belgian American Educational Foundation (baef).15 These meetings will allow them to share their views on the ‘Mouton’ – to moutonner as they call it. Later on, as their partnership deepens, the two men will display an amazing linguistic creativity, labelling their encounters as séances Mouton,16 Commission de moutonnements internationaux,17 recherches moutonnières,18 palabres moutonnants19 or, very modestly, gazouillements Mouton.20 Up until the end, they will dream of being able to ‘finish up the Mouton’.21 And as Coremans confesses in 1960: Ce sacré animal me donne des démangeaisons […] Plus je pourrais moutonner, plus je serais heureux (‘This bloody animal gives me an itch. The more I could ‘lamb’, the happier I would be’).22 On 1 June, Panofsky acknowledges the safe arrival of Coremans’ book and readily accepts his invitation to take part in the moutonnements:

I cannot imagine a more useful and, at the same time, more pleasurable vacation than a sojourn in Brussels enlivened by moutonnements with you and your associates.23 The idea takes shape in the summer of 1953 and materializes a year later. The first challenge is to set up a small group of international experts, advocatus diaboli as Panosky calls them. One name emerges: Martin Davies (fig. 4.4), the future director of the National Gallery in London. According to Panofsky, he will be a perfect antidote to his vivid imagination: ‘he seems to be a man who does not believe anything [...] he will reduce me to order’.24 Early Netherlandish Painting appears at the beginning of 1954. The very first copy is sent to Coremans, who, right away, reads the chapter on the ‘Mouton’. He is absolutely convinced by the whole theory.25 But Panofsky sends him new protestations of modesty: I am only too conscious of the fact that much more might be done and that a great many details stand in need of correction. To discuss all these points with you and to make all the necessary corrections is the main purpose of our forthcoming visit.26

70

hélène dubois, jana sanyova and dominique vanwijnsberghe

Je suggère tout simplement que chacun garde son opinion et que l’on délimite deux phases dans la création et l’aménagement du polyptyque – phases que nous appellerions A et B.29

Fig. 4.4 Martin Davies

Now that the two books are out, now that they are in the hands of reviewers and fellow scholars, the urgent need for new moutonnements becomes obvious. Coremans takes the initiative to define the theme of these study days. They will tackle the ‘Eyckian phase of the Ghent polyptych’27 i.e. Panofsky’s theory according to which the altarpiece is composed of originally unrelated elements left behind by Hubert, transformed or finished by Jan. They will also examine the alterations that each panel or group of panels underwent during the Eyckian phase. And he goes on by proposing to go further and explore other important issues such as the different types of tile in the upper part, Beenken’s theory of an original altarpiece in inverted T-shape, the tower of Utrecht Cathedral, and the quatrain.28 A last thorny issue (perhaps even a moot point), is the respective share of Hubert and Jan van Eyck in the altarpiece. Coremans proposes an odd compromise:

It takes another few weeks to agree on the members of the scientific committee. The core members, Coremans, Panofsky and Davies, will be joined punctually by Hélène Adhémar, curator at the Louvre, and Karel G. Boon from the Rijksmuseum in Amsterdam.30 The Salon des refusés can boast such prominent personalities as Friedrich Winkler31 and Otto Pächt. Before heading off to Europe, Panofsky lists the specific points he wants to discuss in order of priority: 1. What he calls the ‘transitional zone’ in the lower register of the painting in which changes are visible to the naked eye; 2. The original shape of the panels, more specifically whether the Angel Musicians might have been cut round on top; 3. The changes in the lower part of the ‘upper triptych’.32 The working sessions with Davies and Coremans start on 6 July in the premises of the Laboratoire, where all the technical documentation – X-radiographs, ultraviolet and infrared photographs – has been gathered. They can be compared to large photographs and close-ups of the altarpiece. Cross sections of the paint layer can also be reexamined by the laboratory. During the second week, Adhémar, Boon and Lavalleye join them. Every word is carefully transcribed in circumstantial reports made by the secretaries of the Centre for the Study of Flemish Primitives – the ‘three recording angels’ as Panofsky kindly calls them,33 Nicole Veronee-Verhaegen, Jacqueline Folie and Anne Carton de Wiart. The reports are then handed over to each participant and further discussed in subsequent meetings. All these notes and documents are kept in the archives of the Institute.34 During those discussions Panofsky’s ideas are submitted to a critical examination. According to him, Hubert’s contribution included the lower tier, the enthroned figures above, as an autonomous Deisis, and the Singing and Musician Angels,

revenons à notre mouton

initially designed as organ shutters. Jan would have adapted and completed these panels, while painting on his own Adam and Eve and all the reverse sides of the wings. A particularly critical area in this theory is the famous ‘transitional zone’ with the landscape and the sky in the lower register. It was of vital importance in Panofsky’s concept, as according to him it would have been remodelled when the polyptych was assembled. The observation was not new: several changes in this area had been noted before the laboratory examination of 1951.35 It was already obvious at that time that many elements of the landscape had not been planned initially. Indeed, they had not been prepared by an underdrawing and they were not blocked out from the underlayers of the background and of the sky, but rather painted on top of them. The whole, still unresolved question was of course to establish whether these revisions were Eyckian – presumably Jan transforming Hubert’s work – or rather subsequent changes. And it is worth noting that, independently of Panofsky, Coremans and his team had associated the weaker pictorial quality of several zones, unworthy of Van Eyck’s typical accuracy, with very old restoration repaints.36 Therefore, these problèmes particuliers as Coremans called them37 – and especially the ‘transitional zone’ – dominated the 1954 summer sessions. It was then essential to decode the important compositional changes in this critical area, in particular the dove and the landscape, while situating the material data in a historical and stylistic context. Indeed, the first six sessions, organized en petit comité (Coremans, Panofsky and Davies), were devoted to the study the Adoration of the Lamb and its wings. But even during the study of other parts the following week, the ‘transitional zone’ regularly crops up in the discussions. Most questions are directed to Coremans and his laboratory as the art historians seek to understand the interpretation of the technical documents Coremans had presented in L’Agneau Mystique au laboratoire, and to check its validity.

71

When, on 16 July, they all go to Ghent to devote one day – just one! – to the study of the polyptych in situ, the guests are forced to acknowledge, not without frustration, that the surface appearance of the paint does not match the deep structure observed in the X-radiographs.38 Faced with this evidence, they qualify some of their findings and request additional laboratory documents for the future moutonnements. Panofsky returns to Princeton visibly satisfied with his Belgian adventure, disapproving only of the very Belgian tendency ‘to kill one by kindness’.39 He welcomes the fact that his theory fared rather well in the discussions. To continue: […] our 1001 desiderata have still to be dealt with, and God knows what your next exploration in October will bring to light. Yet the evidence already on hand seems to converge, more or less, in one direction, and when we meet again, like witches of Macbeth, we may be ready for some kind of public announcement.40 There will be no second time. From 1955, Coremans is caught up in the case of the fake Vermeer and the lawsuit filed by the Dutch art collector D.G. Van Beuningen,41 a difficult situation that will only be resolved in 1957. Panofsky invites his friend Paul to spend several months at Princeton to complete the work.42 But Coremans, faced with mounting responsibilities, is unable to take a long sabbatical. In Belgium his presence is needed to co-ordinate the construction of the new building of the Institute, which will open in 1962. So that when he sombrely turns down a trip to Princeton in October 1959, Coremans signs de facto the death warrant of the moutonnements.43 The very promising seminar of 1954 has never led to a publication. And the implications of the findings by Coremans and his team on the technical execution and the genesis of the Ghent Altarpiece were rarely revisited. One of the few attempts in this direction is the fundamental article by J.R.J. van Asperen de Boer,

72

hélène dubois, jana sanyova and dominique vanwijnsberghe

published in 1979.44 The Dutch scholar conducted several campaigns of infrared reflectography of the polyptych, re-examined paint cross sections and inspected the panels with a binocular microscope. He emphasized the likely intervention of assistants, at least in the first stages of the painting process,45 a particularly persuasive hypothesis given the size of the work. In nearly sixty years, progress in the technological study of paintings has revealed unsuspected aspects of Van Eyck’s approach and technique,46 such as the subtle transformations introduced in the course of painting by this unparalleled master of illusionism – as several papers given at the 2012 Van Eyck Symposium have illustrated.47 Yet the issue of workshop participation has not been reviewed. The ongoing conservation and restoration treatment of the altarpiece (2012-2017)48 as well as the re-examination of Coremans’ samples49 will at last provide the extraordinary opportunity to reconsider the theories passionately discussed over the years. Coremans’ legacy to his Institute, including notes on and photographs of the Flemish Primitives assembled since 1943 and the numerous paint samples from the 1951 treatment, are a rich source for further moutonnements. They will allow us to reassess still unresolved issues such as the genesis of the transitional zone and the attribution of early overpaint. Let us now dwell on these issues to show that they have lost none of their relevance. As we have seen, the ‘transitional zone’, the dove in particular, puzzled Panofsky who had proposed that the Adoration panel was initially about 11 centimetres taller and presented at the top a golden glory from which rays emanated toward the different groups surrounding the altar. He took the view that the heavenly court originally represented in the lower register did not require the presence of the dove of the Holy Spirit.50 During the moutonnements Panofsky insisted several times that the laboratory check whether the presence of the barbe and unpainted edge around the panel provided indisputable evidence that the size of the painting had not been reduced.51

Fig. 4.5 Coremans’s sketch describing the barbe along the edge of a panel painting

Coremans patiently explained these features with several sketches (fig. 4.5). On 12 July, Panofsky, agreed resignedly that, as reported in L’Agneau Mystique au laboratoire, the format of the central panel was original.52 Coremans showed that the somewhat heavily executed dove had been added over the paint layers of the sky, and that no trace of gold leaf could be detected in paint cross sections.53 These observations seriously undermined Panofsky’s theory of the ‘golden glory’. The area of the dove had been repeatedly sampled in 1951 because the original character of this layer had been questioned from the very start of treatment. With the support of a national and an international commission,54 Albert Philippot had removed the old, probably sixteenth-century grey clouds to reveal the rainbow visible today.55 It was, therefore, an unexpected development – and one that Panofsky would have gladly welcomed – that when during the examination of the painted surface with a binocular microscope in 1978, Van Asperen de Boer indeed observed traces of gold under the dove, in areas that had not yet been sampled.56 Since neither the X-radiographs of the whole surface, completed in 1986, nor the recent infrared images have revealed any traces of a golden glory so far, this part of the panel, a major issue of the moutonnements, will understandably be ‘under close surveillance’ in the coming years. Hopefully the observations carried out during treatment and further analysis will provide irrefutable evidence that will solve this particularly vexing issue.

revenons à notre mouton

In any case, it is highly probable that the dove already appeared in the painting in 1458, since in that year the Adoration of the Lamb, complete with dove, was reproduced in the form of a tableau vivant as part of the pageant organized in Ghent for the Joyous Entry of Philip the Good57 (fig. 4.6). But is it part of an Eyckian revision or, as Van Asperen de Boer suggested, of a non-Eyckian intervention prior to 1458?58 In 1951, the detailed description of the pageant, although published in 1839-1840, had not attracted the attention of the scholars who followed the treatment.59 Coremans, his colleagues and the commission agreed that the heavily executed dove, the crown of the Enthroned Deity and the inscription on the steps below that figure were the result of a very old restoration, carried out before Coxcie’s copy of 1557, which displays these details.60 The 1550 intervention by the painters Jan van Scorel and Lancelot Blondeel, although only mentioned by the Ghent historian Marcus van Vaernewyck in 1568,61 appeared as a likely candidate. The addition of the tower of Utrecht Cathedral and old revisions in the landscape in the Adoration of the Lamb were naturally associated with Van Scorel’s position as a canon of Utrecht, even though this very detail also appears in the landscape of the Virgin and Child with Chancellor Rolin in the Louvre.62 In 1951, the commission concluded that the Adoration of the Lamb, the Enthroned Deity and the Singing Angels had been largely overpainted in

73

the sixteenth century. Coremans and his colleagues also based their conclusions on the fact that the stratigraphy of the areas containing elements unworthy of Van Eyck included opaque paint layers on top of very thin, presumably worn finishing glazes and the presence of thin unpigmented ‘transition varnish layers’ in between the top glazes in the red mantel of the Enthroned Deity.63 According to Coremans, these varnish-like layers separate the original from two later overpaints. This stratigraphy, recently re-examined in the laboratory, turns out to be even more intricate: a microscopic thin section prepared from a paint sample from this area shows at least three levels of red glazes, some of them separated by thin, varnish-like layers (fig. 4.7). This paint build-up that, in 1951, appeared atypical of Van Eyck, requires careful interpretation, as recent technical examinations of Van Eyck’s paintings have shown that the artist indeed applied intermediate, transparent, varnish-like layers between glazes64 which may have been used to saturate a matte surface in order to finalize the modelling or modify the form.65 Van Eyck finished his paintings to perfection and paint samples of the Ghent Altarpiece often show a complex stratigraphy, alternating opaque, semi-opaque and transparent layers, which, as already concluded by Brinkman, Kockaert and their co-authors, belong to the same paint build-up.66 Glazes of different compositions overlap, as in other Flemish and German paintings of the same period.67

Fig. 4.6 Reference to the dove in the description of the pageant in Ghent in 1458

74

hélène dubois, jana sanyova and dominique vanwijnsberghe

1 2

3 4

5 6 7 8 9 10 11 12 13 14 15 16 17 18

Fig. 4.7 Paint thin-section from the red mantel of the Deity Enthroned. Microphotographs in transmitted light with bright field illumination showing the layer structure composed of three opaque and semi-transparent red layers (nos 3, 4, 6), several red glazes (5, 7, 8, 9, 12, 14, 15) and varnishes (11, 13, 16, 17, 18, 19)

The scholars who met in the summer of 1954 knew the description of the 1458 pageant and were faced with a dilemma since the apparently irrefutable material evidence contradicted the historical description. Their conclusion is a puzzling compromise: if these elements actually existed before 1458, they must in any case have been repainted in the sixteenth century!68 Obviously, the distinction between original features and very old overpaint needs to be addressed cautiously. Renewed examination of cross sections using highly sophisticated techniques should help to characterize these many individual layers. This interpretation must be carried out, as noted by the participants of the moutonnements in 1954, by comparing the analytical results with observations on the panels. The interpretation of changes in style and iconography should also be considered by highly specialized art historians. Although the results of extensive discussions in 1954 never led to a publication, the still unanswered questions are an inspiration and food for thought for all those who are involved in the research and conservation treatment of the Ghent Altarpiece. The moutonnements document all stages of the gradual construction of knowledge, from the questioning of material evidence to the development of explanatory models and theories. The participants pro-

ceeded by trial and error, driven by the dynamics of constantly evolving questioning. Obviously, the interdisciplinary dialogue that started sixty years ago must continue and be part of the advisory process accompanying the treatment. It will benefit from new data on the artist, the art work itself, and the historical records. Above all, it is the direct confrontation of the various theories with the ‘naked’ altarpiece during conservation treatment that will provide new evidence for fresh moutonnements. NOTES * The authors wish to express their gratitude to Alistair Watson and Christina Currie for their careful editing of this paper. Thanks are also due to Dominique Deneffe for facilitating access to the Archives of the Royal Institute for Cultural Heritage (KIK-IRPA) and to Cécile Glaude for her help with the analysis of paint samples. 1 Coremans 1953. 2 Masschelein-Kleiner 2000. 3 Panofsky 1953. 4 Panofsky 1953, pp. 205-232. 5 In 1955, Coremans briefly reported on this meeting and expressed the hope that its results would be published in full to advance understanding of this complex painting. See Coremans 1955. 6 Some of the letters are published in Panofsky 2006, Panofsky 2008 and Panofsky 2011. 7 KIK-IRPA, Archives, 1949-1965, II. General Correspondence, C. Foreign Correspondents, Erwin Panofsky (abbreviated 1949-1965 EP). 8 Coremans and Janssens de Bisthoven 1948. 9 KIK-IRPA, Archives, 1949-1965 EP, ref. 5/0/20529/PC/RB and L. 3/2/21570/PC/RB. 10 Beenken 1933; Beenken 1933-1934; Beenken 1936. 11 KIK-IRPA, Archives, 1949-1965 EP, ref. L. 3/2/21570/PC/RB. 12 Panofsky 2006, pp. 210-212, no. 1486. 13 Même en Gotland, île enchantée et jadis Carrefour des courants artistiques où l’on trouve, dans la même église, des vitraux Rhénans, des sculptures Parisiennes et des peintures murales Russo-Byzantines, le problème du ‘Mouton’ […] ne me laisse pas dormir (KIK-IRPA, Archives, 1949-1965 EP). 14 KIK-IRPA, 1949-1965 EP. Published in Panofsky 2006, pp. 379-380, no. 1580. Panofsky probably refers to the observation made by Coremans, Philippot and Sneyers that the format of the Adoration of the Lamb was never altered (Coremans, Philippot and Sneyers 1951, p. 5), thereby eliminating the key argument in his hypothesis on the genesis of the altarpiece and its relationship to the Washington Annunciation. Cf. Panofsky 1935, pp. 460-461 and Panofsky 1938, pp. 419-429. 15 Letter from Coremans to Panofsky, 13 May 1953 (KIK-IRPA, Archives, 1949-1965 EP, ref. L3/2/51611/PC/HD), published in Panofsky 2006, pp. 446-448, no. 1619. 16 Letter from Coremans to Panofsky, 30 September 1954 (KIKIRPA, Archives, 1949-1965 EP, ref. 5/0/63489/PC/JF). 17 Letter from Panofsky to Coremans, 13-17August 1954 (KIKIRPA, Archives, 1949-1965 EP, published in Panofsky 2006, pp. 596-598, no. 1708a – on p. 598, read ‘jeunes filles en fleur’ instead of ‘jeunes tilles’). 18 Letter from Panofsky to Coremans, 13 June 1960 (KIK-IRPA, Archives, 1949-1965 EP). 19 Letter from Coremans to Panofsky, 30 September 1954 (KIKIRPA, Archives, 1949-1965 EP, ref. 5/0/63489/PC/JF).  20 Letter from Coremans to Panofsky, 15 October 1954 (KIKIRPA, Archives, 1949-1965 EP, ref. 5/0/63759/PC/JF).

revenons à notre mouton

21 Letter from Panofsky to Coremans, 5 January 1959 (KIKIRPA, Archives, 1949-1965 EP); Letter from Coremans to Panofsky, 28 February 1964 (KIK-IRPA, Archives, 1949-1965 EP, ref. 5/0/157145/ PC/LP): Moi aussi, je voudrais terminer le ‘mouton’, mais pas sans vous. Sinon, ma foi, le ‘mouton’ restera dans sa cave jusqu’au moment où, après nous, un autre reprendra les recherches. 22 Letter from Coremans to Panofsky, 21 April 1960 (KIK-IRPA, Archives, 1949-1965 EP, ref. 5/0/111830/PC/MV). 23 Letter from Panofsky to Coremans, 1 June 1953 (KIK-IRPA, Archives, 1949-1965 EP, published in Panofsky 2006, pp. 446-447, no. 1619). 24 Letter from Panofsky to Coremans, 25 January 1954 (KIKIRPA, Archives, 1949-1965 EP). 25 Letter from Coremans to Panofsky, 5 February 1954 (KIKIRPA, Archives, 1949-1965 EP, ref. L5/0/58437/PC/HD). 26 Letter from Panofsky to Coremans, 1 March 1954 (KIKIRPA, Archives, 1949-1965 EP). 27 Letter from Coremans to Panofsky, 5 March 1954 (KIK-IRPA, Archives, 1949-1965 EP, ref. L3/2/58685/PC/HD). 28 La théorie de l’agencement d’éléments disparates qui forment maintenant le polyptyque de Gand. Elle comprendrait ensuite l’examen de chacun des panneaux ou des groupes de panneaux et des modifications d’état (changements de composition) que ces panneaux auraient pu subir durant cette phase eyckienne. En supplément, nous pourrions aussi essayer de nous mettre d’accord sur quelques points très importants, mais apparemment posteyckiens: les trois dallages, le couronnement rectangulaire du panneau central, la tour d’Utrecht et l’ordre de succession ‘Ermites-Pèlerins’; enfin, le quatrain. Letter from Coremans to Panofsky, 5 March 1954 (KIKIRPA, Archives, 1949-1965 EP, ref. L3/2/58685/PC/HD). 29 Ibid. 30 Letter from Coremans to Panofsky, 15 April and 12 May 1954 (KIK-IRPA, Archives, 1949-1965 EP, ref. 5/0/59766/PC/JF and 5/0/60400/PC/JF). 31 Letter from Panofsky to Coremans, 6 April 1954 (KIK-IRPA, Archives, 1949-1965 EP, published in Panofsky 2006, pp. 557-559, no. 1688a): ‘The presence of Winkler would not make me particularly happy, not only on account of the political past but also because, in contrast to Heydenreich, he has done too much work in the Flemish field. His concept of Hubert, like Hulin de Loo’s, has crystallized around the “Hand G” miniatures, and I am afraid that he will find it very difficult to make a fresh start. If I had my own way, I should rather eliminate the question of those miniatures altogether and stick to the Monument exclusively’. 32 Letter from Panofsky to Coremans, 21 May 1954 (KIK-IRPA, Archives, 1949-1965 EP, published in Panofsky 2006, pp. 579-581, no. 1701).  33 Letter from Panofsky to Coremans, 13-17 August 1954 (KIKIRPA, Archives, 1949-1965 EP, published in Panofsky 2006, pp. 596598, no. 1708a). See also Folie 1996-1998, p. 225. 34 KIK-IRPA, Archives, Dossier Ghent Altarpiece, Comptes rendus du Symposium ‘La phase eyckienne du polyptyque de Gand’ (abbreviated DGA, Comptes rendus). 35 Panofsky mentions studies by de Tolnay and Hulin de Loo (KIK-IRPA, Archives, 1949-1965 EP, Letter from Panofsky to Coremans, 13 June 1949). He probably refers to de Tolnay 1939, p. 21, 48 n. 42 and Hulin de Loo 1931, p. 126. 36 Coremans 1953, pp. 98, 108, 112, 114. 37 Coremans 1953, pp. 98-125. 38 KIK-IRPA, Archives, DGA, Comptes rendus, July 16, 1954: En étudiant l’œuvre même, on a pu se rendre compte que, à de très nombreux endroits, l’aspect de surface ne correspondait pas aux structures telles que celles-ci apparaissaient, notamment en radiographie. 39 Letter from Panofsky to Udo von Alvensleben, 19 July 1954 (Panofsky 2006, pp. 594-595, no. 1707). 40 Letter from Panofsky to Coremans, 13-17 August 1954 (KIKIRPA, Archives, Archives, 1949-1965 EP, published in Panofsky 2006, pp. 596-598, no. 1708a).

75

41 ICOM 1956. 42 From 11 January 1960 to 8 April 1960. Letter from Panofsky to Coremans, 9 December 1958 (KIK-IRPA, Archives, 1949-1965 EP). 43 Letter from Panofsky to Coremans, 28 October 1959 (Panofsky 2008, pp. 543-544, no. 2399). 44 Van Asperen de Boer 1979. 45 Van Asperen de Boer 1979, p. 212. 46 See for example Gifford 2000, Brinkman et al. 1984-1985 and Brinkman et al. 1990. 47 See for example, the papers by Périer D’Ieteren and Dunkerton, Morrison and Roy in this volume. 48 See the keynote paper by Van Grevenstein-Kruse and Dubois in this volume. 49 The re-examination of samples taken during the 1950-1951 treatments and in 1986 is the focus of a four-year project (2012-2015) led by Jana Sanyova at KIK-IRPA, supported by BELSPO (Belgian Federal Science Policy) and entitled ‘The Mystic Lamb in the laboratory 60 years after Paul Coremans. The contribution of new analytical techniques’ (MO/39/011)’. 50 Panofsky 1953, pp. 218-219. 51 KIK-IRPA, Archives, DGA, Comptes rendus, 9-10 July 1954. 52 KIK-IRPA, Archives, DGA, Comptes rendus, 12 July 1954. 53 KIK-IRPA, Archives, DGA, Comptes rendus, 10 July 1954. 54 The international commission was composed of R. Huyghe (Musée du Louvre), H. Plenderleith (representing the National Gallery, London), A. van Schendel (Rijksmuseum) and representatives of St Bavo’s Cathedral, civic authorities and museums. A national commission prepared the international meetings and comprised the museum curators P. Fierens and W. Vanbeselaere as well as P. Coremans, R. Sneyers and A. Philippot. The introductory meeting on 10 November 1950 involved representatives from UNESCO, ICOM, curators and scholars, each representing a different approach to painting restoration: C. Brandi, P. Hendy, N. Maclaren (both London, National Gallery) and G.L. Stout (Worcester, Massachusetts): see Coremans 1953, p. 8 n. 2 and Dubois, Van Grevenstein 2011. 55 Coremans 1953, pp. 92-93, pls XXXII-XXXIII. 56 Van Asperen de Boer 1979, pp. 193-194. 57 Serrure, Blommaert 1839-1840. The description of the pageant was analysed in depth in Bergmans 1907. For a more recent survey of the Joyous Entry and its relevance to the arts in Ghent, see Dhanens 1987. 58 Van Asperen de Boer 1979, pp. 70-71. 59 Although the description of the tableau vivant is mentioned in L’Agneau Mystique au laboratoire (in the chapter dealing with the material history of the polyptych, see De Schryver, Marijnissen 1953, pp. 22 and 34), its full implications are not explored in the discussion on the different phases of the polyptych (Coremans 1953). 60 These elements are also mentioned in the description of the 1458 pageant. 61 Van Vaernewyck 1568, fol. 117 v, as referred to in De Schryver, Marijnissen 1953, no. 11, p. 36. 62 Pächt 1956, p. 268. 63 Coremans 1953, p. 105, pls XIX.3, XXIVbis.4. 64 A thin transparent layer sandwiched between two layers of red glaze was observed in a cross section taken from the red dress of the Portrait of Margaret van Eyck (Bruges, Groeningemuseum). This painting was restored and studied at the National Gallery, London: see the contribution by Dunkerton et al. in this publication. 65 They may also be an exudate of binding medium or a substance formed over the years through the interaction of pigments and binding media. 66 Brinkman et al. 1990, pp. 32, 37. 67 For example Dunkerton 2008 and Sauerberg et al. 2009. 68 KIK-IRPA, Archives, DGA, Comptes rendus, 12 and 14 July 1954.

1397? 1406

1391

1377

1404

1406

1408

1400

1389

1368

1368

1412

1403

1400

1406

1391

1404

1392

1397

1399

1389

1389

1379 1407 1407 1405

1405 1395 1406

1395

1395 1404 1402 1407 1406

Fig. 5.1 Diagram showing the planks that make up the panels: planks from the same tree are the same colour; the dates represent the year of the last ring measured for each tree

5

Results of Three Campaigns of Dendrochronological Analysis on the Ghent Altarpiece (1986-2013) Pascale Fraiture

ABSTRACT: Between 1986 and 2013, the Royal Institute for Cultural Heritage has undertaken three campaigns of dendrochronological analysis of the Ghent Altarpiece. This paper presents information collected by this study, which involves the twelve panels that compose the altarpiece. In addition to the resulting chronological data the specific characteristics of the wood, commonly termed ‘Baltic oak’, will be described; the different trees used to make the panels will be detailed; and their mode of procurement and their provenance(s) discussed. The data acquired, of different kinds and degrees of precision, deepens our understanding of this masterpiece of Western art and the context of its production in the Middle Ages.

—o— The Ghent Altarpiece by Hubert and Jan van Eyck has been at the heart of art historical research for decades and its material history has intrigued specialists for generations.1 Dendrochronological study of the altarpiece began in 1986, thereby confirming early scientific interest from a discipline which was itself in its early stages and particularly so in relation to art history.2 From the initial opportunity to study six of the wings in June 1986, when the altarpiece was moved to its new protected chapel in St Bavo’s Cathedral, to the completion of the latest dendrochronological analysis in May 2013, when the Adam and Eve panels were removed from their frames for the first time, two

generations of dendrochronologists have studied it.3 This paper presents the results and conclusions of twenty-seven years of dendrochronological research undertaken on the twelve panels that make up the Ghent Altarpiece. Description of the Wood Used for the Altarpiece Each of the five panels from the upper part of the altarpiece and the four wings of the lower part are composed of three oak planks (Quercus robur or Q. petraea) which are between 20 and 32 centimetres wide; the Adam and Eve panels are made of two planks measuring 15-17 centimetres wide (fig. 5.1). Maximum lengths reach 169  centimetres for the Virgin Enthroned and John the Baptist Enthroned and 212-214 centimetres for Deity Enthroned and Adam and Eve. The Adoration of the Lamb panel is comprised of four planks of between 32 and 37 centimetres in breadth and more than 243 centimetres in length. As the wings have been sawn through their thickness,4 their original depth is difficult to estimate. However, the depth of the centre panels is 3 centimetres. Most of the planks studied by dendrochronology are of unusually large dimensions.5 The wings are painted on both front and back. Therefore, their surfaces were perfectly planed and they do not display any tool marks. By contrast, the

78

pascale fraiture

centre panels still show traces of plank splitting, which was the most common method for obtaining oak planks in the fifteenth century.6 Indeed, oak is particularly well suited to splitting because of its characteristic medullary rays (fig. 5.2a).7 Such rays help to guide the split, especially in logs with rectilinear wood grain, as is the case with the altarpiece panels. As splitting follows the oak rays, it produces quarter-cut planks, in theory at least (fig. 5.3).8 Indeed, all the planks of the altarpiece that have been studied were found to be radial cuts (fig. 5.2a). A radial cut by splitting, unlike the semi-radial cut (fig. 5.2b) or the worst tangential cut (fig. 5.2c), gives much waste but allows for the extraction of the best quality planks, thereby minimizing the risk of wood shrinkage during changing environmental conditions. Next to its mechanical advantages, a radial cut has positive implications for dendrochronology. Quarter-cut planks present tree rings that are perpendicular to the edges of the planks, making dendrochronological measurement more reliable,

unlike semi-radial planks, which necessitate incremental measurement which in turn increases the likelihood of error, particularly when rings are narrow and rays are thick (figs 5.2a-b). Moreover, radial planks give the maximum number of visible rings, in contrast to tangential planks that contain only a few rings and generally not enough for dating (fig. 5.2c).9 The core and the sapwood10 are unsuitable for making quality supports since they are susceptible to deformation and/or rot (fig. 5.3). It was thus recommended that craftsmen remove these parts. On the Ghent Altarpiece, while core is entirely absent, sapwood rings have been found on four planks, as in the John the Baptist Enthroned panel. Here, the worm-eaten sapwood weakened the support and was replaced by healthy wood in 1978 following an accident.11 Despite the presence of these sapwood fragments, the first and last growth rings of the trees were removed from all of the planks. But these sapwood remnants indicate an economical use of

Figs 5.2a-c Different orientation of plank cutting: (a) Radial or quarter-cut plank: medullary rays are parallel to the flat edges of the plank and rings are perpendicular. On this example, rays are very straight, which is typical of Baltic oak; these help to guide the split. The radial cut produces planks with maximum number of rings, easy to measure in an orthogonal way (Adoration of the Lamb panel, second plank from the bottom); (b) Semi-radial plank: rays and rings are oblique, at 45° to the flat edges of the plank; there are fewer rings on the planks and necessitate incremental measurement (Dutch panel painting, 17th century); (c) Tangential plank: rays are perpendicular to the flat edges of the plank and rings are parallel (tangential); thus there are very few of these on the plank (wall panel, Liège, 17th century)

results of three campaigns of dendrochronological analysis

Most recent tree ring Oldest tree ring

3

2

I

1

D II C

B

A

Bark

Core Heartwood

Sapwood

Cambium

Fig. 5.3 Diagram of the transversal section of an oak trunk showing: – the various parts of the wood: core, heartwood, sapwood, cambium (and the bark around the wood); – the succession of annual growth rings, the earliest immediately next to the core and the most recentlyformed at the periphery of the trunk, under the cambium; – diagram showing the way waynscots (in grey) are cut, as quarters or eighths of a log (I); these are further split or sawn to produce several planks per waynscot (II); – different positions of radial planks (in orange), which influence interpretation of the tree’s felling date: plank with cambium and complete sapwood (A), plank without cambium but with some sapwood rings (B), planks without sapwood (only the sapwood is absent in case C; sapwood and some of the heartwood rings are missing in case D); – different orientations of planks in the original log, according to the cutting technique used (in green): radial or quarter (1), semi-radial (2), tangential (3)

the logs since craftsmen removed just the minimum of wood necessary (or in these cases just below the minimum). The growth rate corresponds to the average amount of wood produced per year. Slow growth is characterized by narrow rings, rapid growth by wide rings.12 Oaks with slow growth produce softer wood than rapid growth trees13 and slow growing trees are relatively impervious to deformation due to shrinkage or dilation. They are thus more suitable for

79

making stable supports for paintings. Identification of the growth rate for the planks of the altarpiece reveals homogeneity within the panels. Most of the planks (twenty-two out of thirty-five) show slow to very slow growth, the mean ring width being less than 1.2 millimetres (fig. 5.4a). The growth rate of seven other planks can be considered medium (fig. 5.4b), relatively rapid at the beginning of the sequence and then slowing. Finally, six elements show a more rapid growth, with the mean ring width reaching 2 millimetres per year (fig. 5.4c). Planks with such characteristics were obtained from remarkable oaks. To produce such wide radial elements (more than 30 centimetres for the most part) from slow-growing trees, oaks several centuries old were used: indeed some of the dendrochronological sequences number more than 350 rings, even without the first and last growth rings. Moreover, such long and regular planks indicate that these trees would have grown with exceptional regularity, both in growth rhythm and growth shape. The oaks used to produce the supports of the Ghent Altarpiece were exceptional, even for oaks from the primeval Baltic forests. Methodology Measurement Recording During the dendrochronological analysis of a work of art, as opposed to the studies of buildings or archaeological material, the dendrochronologist does not physically remove a sample, because of the precious nature of the object. Therefore direct access to the rings is necessary. For softwood such as pine or spruce, measurements can be made on the longitudinal surface of the tree, which corresponds to the flat side of a radial board. However, for oak, we consider that only the transversal section of the tree provides conditions good enough for precise measurements.14 For the centre Adoration of the Lamb panel the lateral edges of the planks are suitable, the grain of the wood being horizontal; for all the other panels, the end grain is present on the horizontal edges with the wood grain being vertical (fig. 5.1).

80

pascale fraiture

Figs 5.4a-c Macro-photographs of the edges of oak planks (transversal section, semi-millimetric scale) for dendrochronological measurements (campaigns of 2010 and 2013), with growth rhythm distinguishable one from the next: (a) Very slow growth rate for the second plank from the bottom of the Adoration of the Lamb panel: very narrow rings; (b) Medium growth rate for the central plank of the of Virgin Enthroned panel: rings neither too narrow nor too wide; (c) Rapid growth rate for the left plank of the John the Baptist Enthroned panel: rings thicker than 2 mm

Several techniques for acquiring dendrochronological data are available. The most commonly applied is the recording of ring widths directly from the object, using a magnifying glass or microscope. This method was used in 1986 by Jozef Vynckier

(fig. 5.5a). However, conscious of the fact that later checking is impossible once the work is reframed, he had the visionary idea of also recording the ring series on an intermediate support. He worked with kik-irpa photographer Jean-Luc Elias, who took high quality silver-film images calibrated with a scale (fig. 5.5b). These images could be analysed in the laboratory long after the reframing of the wings. They played – and still do play – the role of samples from which the tree rings can be remeasured as needed, as with cores taken from beams in a roof frame building. In 2010, this concept of indirect measurement technique has been entirely developed and is systematically applied to works of art. Ring series are recorded on digital macrophotographs calibrated on a millimetric scale (figs 5.4a-b-c) and ring widths are measured on the computer screen using special software.15 Precision of measurement is to 1/100 millimetre, instead of the 1/10 millimetre previously obtained with a manual lens. The advantage of this indirect method is that it provides the dendrochronologist with recordings of the rings which can be referred to at will, for verification during measurement, or for future re-evaluations, as could be done with the 1986 campaign. To record precise measurements, wood cleaning is often necessary in order to make the separation between the rings perfectly clear. This step has also been improved, in keeping with the ethics of conservation-restoration. In 1986, rigid blades were used to clean the wood quite drastically (fig. 5.5b). In subsequent campaigns, various techniques were used in order to remove the least material possible, depending on the degree of cleaning required. The most conservative technique consists of simply brushing the plank edges; this process was sufficient for the cleanest edges or when growth was not too slow (fig. 5.4 c). In other cases, superficial refreshing using sharp, flexible ‘Stanley cutter’ blades, was also required, in particular for less regular edges of very slow growing planks; this was generally partial and done along with brushing. Finally, a prototype for the laser apparatus developed at the University

results of three campaigns of dendrochronological analysis

81

This uses the kik-irpa tree-ring database series, the vast majority of which is predisposed toward fifteenth-century Flemish paintings19 and is the result of a well-established international collaboration between European laboratories. This ensures a constant exchange of data, metadata and additional expertise.20

Figs 5.5a-b Dendrochronological analysis of certain wings of the altarpiece, carried out by Jozef Vynckier in 1986 when the altarpiece was moved to the Villa Chapel: (a) direct measurement of the ring widths with a manual lens; (b) Example of a silver-film picture of ring series calibrated with a scale, to complement the direct recording of the dendrochronological sequences

of Liège (ULg/CEA) within the framework of the author’s university research at the end of the 1990s was used on the Adoration of the Lamb panel, again in tandem with brushing. This system enabled surface cleaning without abrasive chemical or mechanical action.16 Data Processing by Computer The next stages of the dendrochronological analysis – correlation of the ring series and dating itself – will not be discussed here as the complete methodological protocol employed for analysing or reanalysing the ring series recorded during the three campaigns has been published in the scientific reports of the 2010 and 2013 campaigns.17 In brief, the dendrochronological study of the Ghent Altarpiece has benefited from recent advanced software.18

Results Planks Produced from the Same Tree Methodologically speaking, there is no established approach that allows for the identification of planks from the same tree. The dendrochronologist reaches a conclusion on the basis of different criteria.21 In practice, the strong similarity between dendrochronological graphs of ring series (fig. 5.6a) and high correlation rates in dendrochronological tests provide convincing arguments. Added to these are the comparison of mean ring width22 and the contemporaneity of the sequences compared, which reflect the tree’s growth (the growth rate and the age of the tree).23 Finally, observation of wood anatomy may contribute to the answer in part, such as when a growth irregularity is seen on two different planks. In some cases, we can easily deduce that two wood elements come from the same log. The central and the right planks from the Deity Enthroned panel came without doubt from the same tree: the similarity between the graphs is striking (fig. 5.6a); the mean ring widths are similar;24 the macroscopic anatomic patterns of the wood are perfectly comparable and the dendrochronological sequences have a contemporaneous ring series.25

1250

1250

1250

1300

1300

1300

1350

1350

1350

Deity Enthroned, left plank Deity Enthroned, right plank

Adoration of the Lamb, 2nd plank from the bottom Deity Enthroned, central plank

Adoration of the Lamb, 2nd plank from the bottom

Deity Enthroned, left plank

Deity Enthroned, right plank

Figs 5.6a-c Drawing of ring series originating from the same tree or from neighbouring trees: (a) Ring series from two planks of the Deity Enthroned panel which are almost completely identical: there is no doubt that they come from the same log; (b) Ring series from the third plank of the Deity Enthroned panel and a plank of the Adoration of the Lamb panel, which again are virtually identical, proving that they originated from the same log; (c) Cross-comparison between the ring series of both pairs of planks showing excellent correlations but not enough to conclude that all these planks originate from the same oak rather than from neighbouring trees

1200

1200

Fig. 5.6b

Fig. 5.6c

1200

Fig. 5.6a

Deity Enthroned, central plank

pascale fraiture

results of three campaigns of dendrochronological analysis

Likewise, consideration of these criteria for all the planks of the altarpiece demonstrates that several groups of wood elements come from a single tree. In total, the thirty-five planks of the altarpiece are from twenty-three different oaks at most, since two groups of three planks and eight pairs of planks have been identified as being respectively produced from the same log (fig. 5.1). In other cases, however, it is more problematic to come to a decision as the aforementioned criteria become less meaningful. In such situations, there is no way to conclude definitively whether or not planks come from the same log or from neighbouring trunks that grew in similar conditions. For example, let us consider four planks made up of two pairs; the first includes two planks from the Deity Enthroned and the other pair the third board of the Deity Enthroned as well as a plank from the Adoration of the Lamb. The first pair resemble each other and clearly come from the same tree, while the second pair also resemble each other and they too appear to come from the same tree (figs 5.6a-b). The dendrochronological dilemma is whether or not both distinct pairs of planks originally came from the same tree. Indeed, the different criteria to be considered between the pairs do not reach the quality level required for planks within the pairs (fig. 5.6c): they could, therefore, have come from one log or from neighbouring trees in the same forest. This implies that other pairs or trios of planks may exist within the altarpiece, but which dendrochronology cannot identify with any certainty. Provenance Beginning in the fourteenth century, many historical sources attest to the trade in the exportation of ‘Baltic’ oaks to the Southern Netherlands by Hanseatic merchants via the Baltic ports.26 Dendrochronology contributes to the understanding of this activity since most of the panels used by Flemish and other northern European painters from the fourteenth to the middle of the seventeenth century were made with this exported

83

wood.27 As well as being available in great quantity, it had mechanical properties that enabled the extraction of quality planks.28 Unsurprisingly the Ghent Altarpiece represents one more example of the recourse to this imported wood: dendrochronology demonstrates that all its planks are made of oak originating from this vast ‘Baltic’ export zone. Dendrochronology and archival documents reveal that Baltic material exported to the west in the thirteenth and fourteenth centuries came mainly from the Gdansk-Pomerania region (Poland). For timbers exported later, historical sources tend to indicate a shift in this activity to north-eastern regions, such as Königsberg (Kaliningrad) and Courland, and to southern Poland. The port of Gdansk still played an important role in this trade since it is located at the mouth of the Vistula River, which represented a vital route for wood supply (fig. 5.7a). Some studies of fifteenth century objects appear to indicate the surroundings of the Białowieża forest as one of the sources exploited for western exportation. However, it is obvious that different regions and woodlands were used concurrently.29 That said, it is difficult to specify the provenance zone of the wood used in the Ghent Altarpiece because the reference chronologies used for dating are themselves of undetermined Baltic provenance. The reference chronologies are composed of ring series from archaeological and artistic material produced in western Europe using imported Baltic oaks.30 However, to precisely locate a provenance within the Baltic area, it is necessary to dispose of local chronologies from the potential provenance sources. By this we mean the chronologies constructed on the basis of oak that was locally grown and used, so that these chronologies reflect the growth of the oaks in this precise region.31 However, as most of the native oaks from these regions were exported, almost no oak was used by local populations (at least none of equal quality to those exported). As a result, the construction of local oak chronologies in Baltic countries is a

84

pascale fraiture

Riga

Kalmar

COURLAND (polish in 1561)

Tver

Livonian Brothers of the Sword Dau

SAMOGITIA

BALTIC SEA

GRAND DUCHY OF MOSCOW

gav

Miedniki

Moscow

a Polock

Königsberg

POMERANIA

OV IA AZ

PODLASIE Pinsk

I

BrzeĞü (Brest)

M

ul

L

a

Nowogród Siewierski (Novgorod-Severski)

Turów

t

Sandomierz

Horodlo

KINGDOM OF POLAND

KINGDOM OF BOHEMIA

VOLHYNIA

ia

A

Chelm

ip Pr

I

rta Wa

S

Lublin

Luck

WiĞlica Kraków

PrzemyĞl

Kijów (Kiev)

UKRAINE

Lwów (Lemberg)

RED RUTHENIA

Brno (Brünn)

PODOLIA

LITHUANIAN PODOLIA

Kamieniec Podolski (Khmelnytskyï)

Bou

g

Siret

Pest t

tr

ies

e

Laúi

KINGDOM OF HONGRIA

Dn

ou Pr

Falticeni

Danub

ts

epr

Braclaw (Brastlav)

BUKOVINA Esztergom (Gran)

Done

Dni

Halicz

Koszyce (Kassa)

Buda

Kursk (Koursk)

Czernihów (Tchernihiv)

E

LESSER POLAND

a

Vi st

Sieradz

BraĔsk (Briansk)

GRAND DUCHY OF LITHUANIA

g

Ok

S

GREATER POLAND

Nowogródek (Novogroudok)

Bu

Warsaw

na Des

Wloclawek

Toula

Minsk

Grunwald (StĊbark)

Gniezno

ProznaĔ

Krewo

Malbork (Marienburg)

ToruĔ (Thorn)

SmoleĔsk

Wilno (Vilnius)

pr

PRUSSIA

Chelmno (Kulm) Nieszawa (Nessau)

Kowno (Kaunas)

Królewiec (Kaliningrad)

WiaĨma

Witebsk

Nièmen

Dnie

GdaĔsk

Wroclaw (Breslau)

a

NOVGOROD LAND

LIVONIA (polish in 1561)

PiltyĔ

Vol g

KINGDOM OF SWEDEN

Kalocsa

PRINCIPALITY OF MOLDAVIA

Cluj (Kolozsvár)

TRANSYLVANIA

CRIMEAN KHANAT Perekop Cetatea Alba (Bielgorod, Dniestrowski)

Kilia 0

Caffa

250 km

WALLACHIA 14th century 1st half of 15th century 2nd half of 15th and 16th centuries

Fig. 5.7a Map of north-eastern Europe showing the oak exploitation zones for export to the west from the 14th to the 16th century, after archival texts (after Waz˙ny 2005; graphics E. vander Sloot for Fraiture 2007)

significant challenge. Nevertheless, some chronologies exist for different regions in Poland32 and a few fifteenth century oaks from Lithuania have also been studied by dendrochronologists.33 In present day Latvia, however, researchers so far have found no oak contemporaneous with the period of the Ghent Altarpiece.34

Comparisons of the chronologies from the altarpiece with the available local chronologies are difficult to interpret (fig. 5.7b). If results given by some Polish chronologies such as those from Gdansk and Podlasie are globally better than those given by Lithuanian chronologies, they are not good enough to be conclusively perceived as strong

85

results of three campaigns of dendrochronological analysis

? ? ? ? ? ? Riga

Kalmar

COURLAND (polish in 1561)

Dau

Grunwald (Stębark)

OV

IA

ToruĔ (Thorn)

AZ

Gniezno

ProznaĔ

Minsk

g Bu

M

Wloclawek

Warsaw

S

GREATER POLAND

Vi st

Nowogródek (Novogroudok)

Pinsk

I

BrzeĞü (Brest)

ul

Sieradz

L

a

Sandomierz

Horodlo

KINGDOM OF POLAND

KINGDOM OF BOHEMIA

VOLHYNIA

Czernihów (Tchernihiv)

Luck

WiĞlica Kraków

PrzemyĞl

Kijów (Kiev)

UKRAINE

Lwów (Lemberg)

RED RUTHENIA

Brno (Brünn)

PODOLIA

LITHUANIAN PODOLIA

Kamieniec Podolski (Khmelnytskyï)

Bou

g

Siret

Pest t

tr

ies

e

Laúi

KINGDOM OF HONGRIA

Dn

ou Pr

Falticeni

Danub

ts

epr

Braclaw (Brastlav)

BUKOVINA Esztergom (Gran)

Done

Dni

Halicz

Koszyce (Kassa)

Buda

Kursk (Koursk)

t ia

a

A

Chelm

ip Pr

I

rt Wa

S

Lublin

Nowogród Siewierski (Novgorod-Severski)

Turów

E

LESSER POLAND

BraĔsk (Briansk)

GRAND DUCHY OF LITHUANIA

PODLASIE

a Ok

Nieszawa (Nessau)

Krewo

na Des

Chelmno (Kulm)

Toula

pr

PRUSSIA Malbork (Marienburg)

SmoleĔVk

Wilno (Vilnius)

Dnie

POMERANIA

Kowno (Kaunas)

WiaĨma

Witebsk

Nièmen

Królewiec (Kaliningrad)

Moscow

a

Polock

GdaĔsk

GRAND DUCHY OF MOSCOW

gav

Miedniki Königsberg

Tver

Livonian Brothers of the Sword

SAMOGITIA

BALTIC SEA

Wroclaw (Breslau)

a

NOVGOROD LAND

LIVONIA (polish in 1561)

PiltyĔ

Vol g

KINGDOM OF SWEDEN

Kalocsa

PRINCIPALITY OF MOLDAVIA

Cluj (Kolozsvár)

TRANSYLVANIA

CRIMEAN KHANAT Perekop Cetatea Alba (Bielgorod, Dniestrowski)

Kilia 0

Caffa

250 km

WALLACHIA t>7 6 felling after 1374 P22BEP. 1377 > felling after 1383 1379 > felling after 1385 P466/01/1 P463-01-P466-02-ech 1389 > felling after 1395 1391 > felling after 1397 P21-B 1394 > felling after 1400 P466-03-ech 1395 > felling after 1401 P26A 1397 > felling after 1403 P455-02-2 ?1397 > felling after 1403? 1328? P452-01-2 1399 > felling after 1405 P466/04/1 1400 > felling after 1406 P464/02/2 P465/03/1b 1400 > felling after 1406 1402 > felling after 1408 P29 1403 > felling after 1409 P464/03/2ab 1404 >felling after 1410 P23B

Singing Angels & Musician Angels Hermits & Eve Knights of Christ & John the Baptist Enthroned Knights of Christ & Pilgrims Pilgrims & Virgin Enthroned Adam John the Baptist Enthroned & Pilgrims Virgin Enthroned John the Baptist Enthroned

Deity Enthroned Singing Angels Adoration of the Lamb Deity Enthroned & Adoration of the Lamb Singing Angels & Musician Angels Adoration of the Lamb Knights of Christ & Hermits Eve Adam Adoration of the Lamb John the Baptist Enthroned Virgin Enthroned Pilgrims John the Baptist Enthroned

Figs 5.8a-b Combined dendrochronological bar diagrams for the dated ring series from the altarpiece, with estimation of the felling period for each tree: (a) in chronological order of the last ring measured; (b) panel by panel (in dark orange, sapwood remnants)

1073

pascale fraiture

1023

1023

1036 1038

1036

1050

Fig. 5.8b

1073

1100

1136

1150

1147

1137

1136

1137

1177

1173

1177

1200

1196

1196

1207

1205

1204

1226

1233

1267

1250

1251

1260

1260

1250

1258

1267

1300

1298

1298

1400 > felling after 1406 1406 felling 1433 1408 > felling after 1414

1377 > felling after 1383 1391 > felling after 1397 1404 > felling after 1410

?1397 > felling after 1403? 1406 > felling after 1412

1405 > felling after 1411 1406 > felling after 1412 1395 > felling after 1401

P25A P27C P26A

P29 P27C P28C 1350

P26A P30-P455-01-2-ech

1400

1402 > felling after 1408 1406 > felling after 1412 felling 1440 1407

1395 > felling after 1401 1404 felling 1431

1379 > felling after 1385 P466/01/1 1389 > felling after 1395 P463-01-P466-02-ech 1394 > felling after 1400 P466-03-ech 1399 > felling after 1405 P466/04/1

felling 1440 1407 1405 > felling after 1411

1397 > felling after 1403 1404 felling 1431

P455-02-2 P30-P455-01-2-ech

P28C P25A

1391 > felling after 1397 1404 > felling after 1410 1406 felling 1433

1400 > felling after 1406 1403 > felling after 1409 felling 1412 1442

P21-B P23B P24-P465-01-ech

P464/02/2 P464/03/2ab 1317 P464-01-ech

P463-02-03-ech 1368 > felling after 1374 1389 > felling after 1395 P463-01-P466-02-ech

P465/03/1b P24-P465-01-ech P465/02/1

P22BEP. P21-B P23B

1328? P452-01-2 P452-02-2

Pilgrims

Hermits

Adoration of the Lamb

Knights of Christ

John the Baptist Enthroned

Eve

Musician Angels

John the Baptist Enthroned

Deity Enthroned

Virgin Enthroned

Singing Angels

Adam

Lower section of the altarpiece

Upper section of the alrarpiece

results of three campaigns of dendrochronological analysis

90

pascale fraiture

together for sale to a single wood merchant who in turn supplied a particular panel maker. It is even less surprising to find planks from the same tree in large-scale commissioned works for which the commissioner(s) ordered all the material required for his project at the same time.46 Secondly, assemblages of planks from the same tree guarantee uniform behaviour of the support in response to changes in environmental conditions.47 In the context of this study, given the debate among specialists as to whether the Ghent Altarpiece was devised as one, two, or more distinct projects – or at least phases of project(s) – the repartition in the altarpiece of pairs and trios of planks from the same logs is worth discussing (fig. 5.1). First, it is obvious that the two Angel panels are linked, and that one of them is linked to the Virgin Enthroned. Second, the four wings of the lower part are clearly linked by the elements they include. Third, two pairs of planks demonstrate a relationship between the upper and the lower parts of the altarpiece: one plank of the Eve panel and one plank of the Hermits are from a single log and one plank of the Deity Enthroned and one of the Adoration of the Lamb panel come from another trunk. It should be recalled that ‘Baltic’ oaks were exported from the Hanseatic ports as semi-finished products – quartered trunks obtained by splitting, termed waynscots (fig. 5.3).48 The different pieces of a single trunk could thus have been sold to different wood merchants, joinery workshops or painting workshops. Such waynscots would have then been split or sawn into two, three or four planks before being used in panels.49 In practice, the stocks of imported wood could thus have been distributed between workshops. So it is not so surprising to find planks from the same tree in panels painted by different artists.50 But it would be extremely unlikely to find two planks from the same tree, prepared by different workshops for two distinct projects, finally reunited in the same altarpiece. Consequently, since the centre panels and wings are linked by planks originating from

common logs, as are the lower and upper sections (fig. 5.1), it can be stated that all parts of the altarpiece were produced from a homogeneous lot of wood. It should be added that to date no other painted panel by the Van Eyck family or by any other contemporaneous artist has been shown to be made of planks coming from one of those trunks.51 At another level, dendrochronological comparisons between ring series from the different trees used in the altarpiece show two homogeneous groups of woods (figs 5.9a-b). The first group includes planks from all the wings of the altarpiece (except perhaps the Adam panel which shows weaker correlations). The second group includes mostly ring series from the centre Deity Enthroned and Adoration of the Lamb panels. These two subgroups are not independent of each other, that is to say that certain woods match equally well within one sub-group as in the other (fig. 5.9a).52 Moreover, examination of the repartition of the dendrochronological series between the two sub-groups reveals that the second one systematically includes the longest ring series (all but one have more than 300 rings), coming from the widest planks (more than 30 centimetres) and from trees with the slowest growth rhythm (fig. 5.9b).53 This dendrochronological repartition of the woods between wings and centre panels would therefore be linked mostly to differences in the dimensions of the panels rather than to differences in procurement. Provenance(s) versus Typology(ies) Comparisons of the ring series from the altarpiece with the available European chronologies demonstrate their imported Baltic provenance. However, the lack of local chronologies from this wide exporting area does not enable precise location (fig. 5.7b). Nevertheless, it is possible to go further in this provenance research by means of the study of the dendrochronological series of the altarpiece via a typological approach.54 It is important to stress the fundamental difference that exists between western master chronologies and the so-called ‘Baltic’ chronologies. The

results of three campaigns of dendrochronological analysis

91

Fig. 5.9a Square correlation matrix of t-values between the ring series dated from the altarpiece and the master chronology BALTIC0. All the ring series are represented in abscissa and in ordinate; they are compared individually to each other and the correlation rate between two series is given in the intersection cell of the corresponding line and column. The dark cells represent correlation rates higher than 4 between two series. The graph shows two homogeneous groups of series (encircled in two green rectangles), which are not independent of each other (see the blue circle): certain series match equally well in one sub-group as in the other. This matrix also shows that most of the ring series from the altarpiece give high correlation rates with the BALTIC0 typology: this is translated in the graph by all the grey cells in the first line and first column of the matrix (encircled in red), which correspond to the results of comparison between each series of the altarpiece and BALTIC0

former are built from architectural or archaeological material in situ, which carries a geographical component. The latter – except for the local Polish chronologies, which are also built from in situ material – have been uniquely built from moveable objects (panels, furniture, sculptures, etc.). Thus, chronologies used to date panel paintings have not been built on the basis of geographical criteria, but on dendrochronological ones. If the dated ring series from the altarpiece are compared with the Baltic reference chronologies, most of them (eighteen out of twenty-two) show strongest resemblance to a chronology called BALTIC0, which was built from archaeological material from England (fig. 5.9a).55 BALTIC0 was built with the same methodology as chronologies BALTIC1 and BALTIC2, based on dendro-typological criteria

described by J. Hillam and I. Tyers.56 Each chronology, therefore, reflects a different typology, which is interpreted as being linked to a possibly specific geographical origin and which remains imprecise. BALTIC0 would represent the initial zone exploited at the beginning of the Baltic exports.57 The fact that all the woods from the altarpiece show strongest similarities with BALTIC0 would therefore indicate a general common provenance.58 Consequently, it seems to suggest that the same regions were exploited in the ‘Baltic’ area to supply England and the Southern Netherlands in the fifteenth century. However, this degree of ‘proximity’ remains relative in zones where climatic conditions and soils are the same for hundreds of square kilometres.59

1023

1038

1036

1100

1150

1147

1173

1177

1207

1205

1200

1196

1250

1251

1267 1260

1300

1298

1350

1400

1394 P466-03-ech 1379 P466/01/1

1389 P463-01-P466-02-ech 1368 P463-02-03-ech

1399 P466/04/1

1404 P30-P455-01-2-ech

1395 P26A

1403 P464/03/2ab

1377 P22BEP.

1406 P27C

1407 P28C

1402 P29

1405 P25A

1404 P23B

1406 P24-P465-01-ech

Fig. 5.9b Drawing of the ring series from the two sub-groups highlighted in fig. 5.9a (with corresponding colours): the second group (blue and dark green) systematically includes the longest ring series (all except one count more than 300 rings), coming from trees with the slowest growth rhythm. Add to this the fact that the matrix shows high correlation rates for all the ring series with BALTIC0, this dendrochronological repartition would therefore be linked mostly to different characteristics of the trees used in the altarpiece rather than to differences in procurement

1050

1073

1136

pascale fraiture

results of three campaigns of dendrochronological analysis

Conclusions and Perspectives The thirty-five planks of the Ghent Altarpiece analysed by kik-irpa came from twenty-three Baltic oaks at most, which are remarkable for their slow and regular growth, their age (the oldest reaching 400 years) and their size. Comparisons amongst all the ring series recorded on the altarpiece demonstrate several groups of planks coming from a single tree. Furthermore, repartition of these pairs or trios of planks within the altarpiece reveals that all its parts are linked: wings, wings and centre panels, upper and lower sections. This suggests that the planks used to manufacture the whole altarpiece came from the same procurement network. The three campaigns of dendrochronological study identify a felling period for each panel of the altarpiece. The miniscule differences in dates from one panel to another do not allow for the identification of a staggering in tree felling. However, these results enable us to conclude that the trees must have been felled at the same time. Consequently, on the basis of dendrochronology, it can be stated that the whole altarpiece must have been manufactured with trees sourced from one procurement phase. Using sapwood estimations available to date for the presumed provenance zone of these oaks, this felling period can be situated between 1414 and 1431. Assuming that Hubert van Eyck began the work, followed after his death in 1426 by his brother Jan, the felling period of the tree would have taken place between 1414 and some time before 1426. Comparisons between tree ring series from the altarpiece reveal two dendrochronological subgroups of woods, which are close together. Planks from the wings show a strong resemblance between them. Most planks from the two centre panels also attest to a very coherent dendrochronological ensemble, but many similarities link both subgroups. The observation of the ring series within the sub-groups reveals that they represent two categories of oaks with slightly different growth typologies,

93

the second including the oldest trees characterized by the slowest growth rhythm. Dendrochronology demonstrates that oaks used in the altarpiece originated from the Baltic coastal region. However, to precisely locate this provenance within this enormous sourcing space is not straightforward, given the lack of dendrochronological coverage regarding the possible provenance zones for oak (present-day Poland, Lithuania, Latvia, Belarus…). However, the few pieces of dendrochronological data available for these regions in combination with archival data would suggest a possible Polish provenance (Gdansk or Podlasie), though that is still to be proven dendrochronologically. Furthermore, comparisons of the ring series from the panels of the Ghent Altarpiece with the generic Baltic typologies available from imported archaeological or artistic material show the strongest resemblance with BALTIC0 typology described by Ian and Cathy Tyers. These generic typologies are interpreted as reflecting general provenance zones in the Baltic area, although they cannot be specified. Consequently, oaks from the altarpiece are representative of two sub-typologies of growth within a generic typology, which indicates a similar provenance of the trees used. In conclusion, the dendrochronological analysis of the Ghent Altarpiece in its entirety brings together several indices showing the homogeneity of the batch of trees used to produce its panels and argues in favour of a unique material procurement dedicated to a unique initial project. Whether or not the original project was carried out as planned or modified by the commissioner or the artist(s) along the way cannot be determined by the study of tree rings. NOTES * I would like to thank my predecessor Jozeph Vynckier at the KIK-IRPA for the precision of his dendrochronological work. Many thanks also go to Ian Tyers (Dendrochronological Consultancy Ltd, Retford, UK) for his judicious advice throughout this research, and to Tomasz Ważny (Nicolaus Copernicus University, Toruń, PL; Cornell University, USA) and Rūtilė Pukienė (National Museum – Palace of the Grand Dukes of Lithuania, Vilnius, LT) for the Polish and Lithuanian data they shared with me. Finally, warm thanks to Rebecca Miller

94

pascale fraiture

(Liège, B), Alistair Watkins (Edinburgh, UK/Utrecht, NL) and Christina Currie (KIK-IRPA) for their help in translating and editing this text. 1 See, for example, Coremans 1953; Glatigny et al. 2010 on http://closertovaneyck.kikirpa.be and Verougstraete in this volume. 2 Bauch, Eckstein 1970; Bauch et al. 1978; Bauch, Eckstein 1981; Eckstein et al. 1975; Fletcher 1976; Fletcher 1982; Klein 1983. 3 The first campaign took place in 1986, when the KIK-IRPA carried out the dendrochronological analysis of six lateral wings from the altarpiece (Vynckier 1999-2000). In 2010, a second campaign was undertaken by the KIK-IRPA as part of an international project for the research and conservation of the Ghent Altarpiece initiated by the council of St Bavo’s Cathedral. The centre panels were then examined: the Deity Enthroned, the Virgin Enthroned and John the Baptist Enthroned for the upper section of the altarpiece and the Adoration of the Lamb for the lower section (the scientific report of the analysis is published: Fraiture 2011 on http://closertovaneyck.kikirpa.be). This 2010 scientific analysis was financed by the Getty Foundation (Getty Grant Program, Panel Paintings Initiative) and coordinated by Ron Spronk (Queen’s University, Canada). Finally, the dendrochronology of the two remaining wings, Adam and Eve, was studied in April-May 2013 by the author (see Fraiture 2013 on http://closertovaneyck.kikirpa.be; these two wings were also analysed in 2010 by P. Klein using X-rays undertaken at the KIK-IRPA in 1986; those results will not be discussed here; see Klein 2010 on http://closertovaneyck.kikirpa.be). 4 Vynckier 2002. 5 Dimensions are exceptional, particularly with regard to width: a study of around 300 Baltic planks from fifteenth- to seventeenthcentury Flemish painted panels demonstrates that only 6% of the planks are wider than 32 cm and that 46% are narrower than 25 cm; for the fifteenth century these proportions fall to 4% for planks wider than 32 cm, with a maximum of 34.5 cm (Fraiture 2007). Planks from the altarpiece are also quite long although they are still shorter than the maximum length found in the literature as well as in practice: planks are known that reach 12 feet – around 3.3 to 3.8 m according to the length of the foot, which varies from 28 to 32 cm depending on the production centre (see for example Verougstraete-Marcq, Van Schoute 1989). 6 Fraiture 2007. 7 Medullary rays are ribbon-like aggregates of storage and conducting cells extending radially in the xylem and phloem (www.wsl.ch/ dienstleistungen/produkte/glossare/dendro_glossary/index_EN; Glossary of Dendrochronology adapted from Kaennel, Schweingruber 1995. Multilingual Glossary of Dendrochronology, Bern, Stuttgart, Vienna). 8 Tangential planks with splitting traces have also been encountered (Fraiture, Crémer 2013). 9 Generally, it is considered that 70-80 rings are required to get reliable dendrochronological results (Hillam, Morgan, Tyers 1987; Fraiture 2009a). 10 Sapwood is the peripheral part of the trunk through which the raw sap rises from the roots to the branches; it is thus rich in nutritional substances and very easily degradable. Sapwood contains the most recent rings in the life of the tree. 11 The location of the sapwood on this plank marks the place broken during an accident in 1978, in part due to the poor state of the wood (Marijnissen, Grosemans 1982-1983). 12 A mean ring width of around 1 mm represents very slow growth, while a mean ring width greater than 2 mm reflects rapid growth (Fraiture 2007; Beuting 2011). 13 A ring is composed of earlywood, which is tender and porous, and latewood, which is much denser. In rings of a slow-growing oak, earlywood is proportionally more important than latewood, which gives softness to the wood; although in a rapid-growth oak the proportion of latewood prevails, thus producing a much denser wood.

14 Fraiture 2007; Hoffsummer 1995; Tyers, Tyers 2007. 15 Measuring is done by a specific procedure in AdobeTM PhotoshopTM. Data are then reconstituted and converted into a dendrochronological format. In 2010, this step was done by the Dendron software version II (Lambert 2006; Fraiture 2007); in 2013 by the TakeMeasuresFromAdobePhotoshop tool, version 20120113 (Lambert, KIKIRPA unpublished). 16 The development of this system was made possible by subsidies granted by the Fonds de la Recherche Fondamentale Collective (Fonds National de la Recherche Scientifique) and the Fonds spéciaux pour la recherche de l’Université de Liège. Development of the technical characteristics of the laser was done by Lasea (Brevet BE 9900367). Fraiture 1998 and Fraiture 2009b. 17 Fraiture 2011 and Fraiture 2013 on http://closertovaneyck. kikirpa.be. 18 The Dendron software has been developed and is regularly updated by G.-N. Lambert. For Dendron II (used in 2010), see Lambert 2006. Dendron IV (used since 2012) is currently unpublished (Lambert, former CNRS-Laboratoire de Chrono-Écologie de l’Université de Franche-Comté – UMR 6249, scientific collaborator of the University of Liège). 19 Reference databases built by J. Vynckier who worked at the KIK-IRPA from the end of the 1970s up until 1998, mainly on fifteenth-century Flemish paintings, and by the author at the University of Liège/Centre européen d’Archéométrie from 1998 to 2003 (see Fraiture 2007) and at the KIK-IRPA from 2003. 20 See, for example, the Digital Collaboratory for Cultural Dendrochronology (DCCD). A digital data library for dendrochronology project, initiated by Dr Prof. E. Jansma (Rijksdienst voor het Cultureel Erfgoed, Amersfoort), which made available to its members thousands of dendrochronological data and metadata from around Europe. This project is also the place for methodological research and debates between dendrochronologists. Jansma et al. 2012. 21 Fraiture 2007; Fraiture 2009a; Beuting 2004; Beuting 2011. 22 See above, ‘Description of the Wood used for the Altarpiece’. 23 These two criteria nevertheless suffer from imprecision since, even within a single oak, they can vary depending on the location of the sample in the trunk. Moreover, the number of rings in a sample, while it gives an idea of the age of the tree, also depends on the work involved in wood preparation. 24 Mean ring width of 1.05 mm for one plank and 1.11 mm for the other. 25 The fact that one of the planks was not entirely measured explains the difference in sequence length and from this, the dating for the first rings. 26 See, for example, Salzman 1979; Sosson 1977; Baillie 1984; Tossavainen 1994; Zunde 1998-1999. 27 Regarding the discovery of the use of ‘Baltic’ wood by dendrochronology, see Baillie, Hillam, Briffa 1985 and Eckstein et al. 1986. For the use of ’Baltic’ oak in Flemish works of art, see, for example, Vynckier 1992; Kemperdick, Klein 1997; Fraiture 2009b. For a status quaestionis, see Eckstein, Wrobel 2007; Fraiture 2007. 28 The rigorous climate and the poor soil of these regions produce soft slow-growing trees, unsusceptible to deformation. Moreover, competition between trees in these dense forests generates a rectilinear growth of the oaks, an important quality for the production of long planks. 29 Ważny 2002. Many documents mention wood rafted along the River Vistula. Moreover, with the dense river system in this area of Europe, Gdansk and the Vistula were connected to a huge area of thousands of square kilometres, including the basin of the Neman River, where timber could be collected for export through western Europe. M. Zunde describes exports from Riga through western Europe mostly from the second half of the fifteenth century (Zunde 19981999). On the basis of archival and dendrochronological evidence,

results of three campaigns of dendrochronological analysis

this author suggests that Riga’s imports of timber were done via the River Daugava from regions as distant as the forests located within the Dnieper River basin (Belarus). 30 Chronologies used to date the altarpiece have been constructed from archaeological and artistic material by J. Fletcher (revised by Hillam and Tyers 1995), I. and C. Tyers (unpublished), J. Vynckier (unpublished) and P. Fraiture (unpublished). 31 Fraiture 2009b. 32 Chronologies are, for example, available for Gdansk-Pomerania, Podlasie and the Vistula basin, which cover the fourteenth and fifteenth centuries (Bielsk Podlaski, Bransk, Tykocin and Rusiec). Ważny 2002. 33 These exceptions represent, however, less than 20 individual series for the thirteenth to fifteenth centuries (R. Pukienė, National Museum – Palace of the Grand Dukes of Lithuania, Vilnius, LT). 34 See Zunde 1998-1999; Zunde 2011: dendrochronological work in Latvia led to construction of long Scots pine chronologies but at present, no oak chronology is available for historical periods. 35 Different projects in which the KIK-IRPA Dendrochronology Laboratory is involved work in this field. For example the DendroProvenancing Meetings, historic timber trade between the Baltic area and western Europe initiated by Dr Prof. D. Eckstein (Department of Wood Science of the Hamburg University), or the DCCD project initiated by Dr Prof. E. Jansma (Rijksdienst voor het Cultureel Erfgoed, Amersfoort). 36 Ring series from the different planks which come from the same tree have been grouped to provide one individual series per tree and the most recent ring from all the planks from one tree gives the date which then leads to interpretation. 37 The cambium is the source of reproductive cells that form the sapwood toward the inner trunk. It is found on the periphery of the trunk between the sapwood and the bark. 38 The heartwood is the biologically inactive part of the wood. A tree ring is formed in the sapwood; after a few years, it is transformed into heartwood. 39 This date-range is more or less restricted by the number of preserved sapwood rings on the wooden samples and by the number of samples that still have sapwood present. 40 Eckstein et al. 1986, revised by Ważny, Eckstein 1991 (in a recent personal communication, the author explained that generally he uses a range of 8-23 sapwood rings for Polish oaks, save for exceptional trees, that is for very old and slow growing oaks; pers. comm., February 2013). The wide interval is used here precisely because these oaks are of exceptional age. 41 Sohar et al. 2012. 42 See, for example, Kemperdick, Klein 1997. In this article, the authors propose adding twenty-five years to the date of the last ring measured. They specify, however, that this interval of twenty-five years is not appropriate for all of the works analysed – it is sometimes too long. Fraiture 2008; Vandekerchove et al. 2012. 43 Even in favourable cases where some sapwood rings are still present, the lapse of time between felling and use of the wood is still unknown.

95

44 1413 corresponds to ‘felling after 1412’ and 1431 is the earliest terminus ante quem obtained. 45 Indeed, examination of the dates of all the planks of the altarpiece taken individually shows that other planks give comparable early results, but they are combined with planks in the same panel that have a more recent last ring. That is why considering this terminus post quem of 1395 alone would not be sensible. Furthermore, given the extremely narrow last rings of the ring series recorded on this panel, a difference of 15 rings represents less than 1 cm of wood! 46 See for example the published archives concerning the Raising of the Cross triptych by Rubens (Vynckier 1992). 47 For example, Depuydt-Elbaum et al. 2010; Houbrechts and Vandervellen 2011. 48 This explains why, to date, Baltic oak has only been found in the form of planks (or parts of planks, as in sculpted altarpieces), in the importing countries of western Europe. Based on sources, waynscots were mainly used for architectural or naval construction; from them also came the planks destined for panel manufacture. See, for example, Salzman 1979; Bonde et al. 1997; Zunde 1998-1999; Tossavainen 1994; Ważny 2005; Eckstein, Wrobel 2007. 49 Bonde et al. 1997; Fraiture 2007. 50 Fraiture 2012. 51 This assumption is based on a comparative research between ring series of the altarpiece and reference databases from different laboratories working with Baltic oak (KIK-IRPA and ULg/CEA; Flemish Heritage via DCCD February 2013; Dendrochronological Consultancy, Retford, 22 May 2013; P. Klein, pers. comm., 15 December 2010). 52 For example the chronology of the left plank of the Eve panel and the right plank of the Hermits, which come from the same tree, or for the top plank of the Adoration of the Lamb. 53 These growth differences are more visible in the cumulative growth graphs than the mean ring widths. 54 See, for example, Billamboz 2011; Fraiture 2009b; Hillam, Tyers 1995; Weitz 2012. 55 This chronology, constructed by C. and I. Tyers, is unpublished. 56 Hillam, Tyers 1995. The fourth typology – BALTIC3 – does not concern this study because it does not overlap the period of the woods from the altarpiece (C. and I. Tyers, unpublished). 57 Pers. comm. I. Tyers, 24 October 2005. 58 For the few planks that differ from the majority of the others within the altarpiece or from the BALTIC0 typology, other reasons than a different geographical provenance could be mentioned, for instance a specific location in the forest (on the edge, better light) or strong genetic characteristics of the tree which lower the influence of the climatic signal. The case of the left plank of John the Baptist Enthroned, for example, is easily explained by its rapid growth, which is quite different from all the others and could be due to a more favourable location in the forest. 59 These similar conditions induce a dendrochronological signal that can be common to oaks quite distant from each other in these vast zones (Ważny, Eckstein 1991).

3 1 3

3

VIII

II

III

IX

IV

V

VI

VII

XII

XI

X

2

3

2

1

2

2

2

I

1

1

XIV

XIII

XVIII

XVII

2

2

2

XV

XVI

XIX

2

XX

2

2

Fig. 6.1 Ghent Altarpiece: I Adam, II Singing Angels, III Virgin Enthroned, IV Deity Enthroned, V John the Baptist Enthroned, VI Musician Angels, VII Eve, VIII Just Judges, IX Knights of Christ, X Adoration of the Lamb, XI Pilgrims, XII Hermits, XIII Angel Annunciate, XIV Interior with City View, XV Interior with Lavabo, XVI Virgin Annunciate, XVII Joos Vijd, XVIII John the Baptist (grisaille), XIX John the Evangelist (grisaille), XX Elisabeth Borluut

6

Research into the Structural Condition and Insights as to the Original Appearance of the Panels and Frames of the Ghent Altarpiece Aline Genbrugge and Jessica Roeders

ABSTRACT: Between March and November 2010 intensive research and documentation was carried out on the condition of the panels and frames of the Ghent Altarpiece in St Bavo’s Cathedral in Ghent. The investigation was led by Jean-Albert Glatigny (panel painting and sculpture conservator), assisted by a team of panel conservators: Renzo Meurs (furniture conservator, Amsterdam), Aline Genbrugge (paintings conservator, Brussels, KIK-IRPA) and Jessica Roeders (paintings conservator, Haarlem, Frans Hals Museum). Research into the original tool marks, the evidence of restoration treatments, and the modifications to the support provide additional insight into some of Van Eyck’s working methods, the possible original appearance of the assembly, and the visible effects of previous treatments.

—o— Between March and November 2010, in-depth research and documentation was undertaken to investigate the structural condition of the panels and frames of the Ghent Altarpiece in the St Bavo’s Cathedral in Ghent. The investigation was led by Jean-Albert Glatigny in collaboration with Renzo Meurs, Aline Genbrugge, and Jessica Roeder. During the research, different teams of investigators and panel painting specialists came together to contribute their expertise.1 Jean-Albert Glatigny and his team of panel conservators were responsible for the documentation of all of the findings.2

This paper focuses on the research into the structural condition of the panels and the frames of the Ghent Altarpiece. Some of the research into the original tool marks, the evidence of later restoration treatments, and modifications to the support will be explained. This will give insights into the original appearance of the assembly and the effects of the different treatments that have been undertaken in the past. The Altarpiece and the Conservation History of the Supports The Ghent Altarpiece consists of eighteen panels, seventeen of which were studied in situ in the Villa Chapel in St Bavo’s Cathedral. The only nonoriginal panel of the altarpiece is the copy of the Just Judges, in the lower left register,3 which was not examined in the cathedral during the project. Since its completion in 1432, the altarpiece has had a turbulent history. The panels were separated from each other on several occasions and endured wars, thefts, fires, and diverse treatments. In a broad sense, the altarpiece can be divided into three separate structural groups, each of which has experienced a distinct conservation history (indicated by numbers 1, 2, 3 in fig. 6.1). Within these groups, there are also some further differences in

98

aline genbrugge and jessica roeders

the material history of the various panels (fig. 6.1). The first group (1) consists of the Adam and Eve panels, which are the only panels that have maintained their structural integrity and have not been substantially altered. They are painted on both sides and are still in their original engaged frames. They survived in a relatively unaltered state, partly because the images were considered too indecent to be on display in the church. As a result, the original Adam and Eve panels were removed and put into storage. They were replaced by reconstructions of the panels, in which the figures were more appropriately covered for display in the church environment.4 The second group of panels (2) consists of the other wing panels. These were purchased by Friedrich Wilhelm III of Prussia in 1821 and displayed in the Gemäldegalerie in Berlin from 1830-1920.5 In 1894, all the wing panels except for Adam and Eve were then split lengthwise and cradled, so that the fronts and backs of the paintings could be exhibited alongside each other. Their original frames were also split lengthwise and were adapted by adding a wooden lath on the reverse of the rebate of the frame, so that the cradled panels would be surrounded by something that resembled their original framing. The third group of panels (3) are the centre panels. These are the only panels in the altarpiece that were originally painted on one side only. They are also the only panels that were originally framed in rebate frames. The original frames were lost in the eighteenth century. As part of the treatment in the 1950s, the panels were reframed, and those frames are still in place today. The Original Construction of the Altarpiece: Study of the Upper Register The Adam and Eve panels were the most suitable to focus on in order to obtain a clearer view of the production, construction and the original appearance of the altarpiece. They are the only ones that are still in their original frames, although they have been somewhat modified and altered over time.

The examination of Adam and Eve provided information about the appearance of the lost frames of the three centre panels of the upper register. Adam (with Interior with City View on the reverse) and Eve (with Interior with Lavabo on the reverse) are both painted on panels consisting of two planks of Baltic oak, butt-joined with dowels. For this recent study the panels were not taken out of their frames. Information on the separated panels and their frames was obtained from the extensive documentation of the 1950-1951 examination and conservation treatment by Paul Coremans and Albert Philippot.6 The frames of the Adam and Eve panels have a remarkable construction. In the 1950-1951 photograph of the unframed Eve panel, dowel holes are visible in the unpainted top edge of the panel. The top curved frame member is attached to the panel with dowels; however, the left, right and bottom frame members have grooves on the inner edges which allow the panel to slide in (fig. 6.2). Technically, if this were the case for all four of the frame members, the frame would be called an engaged frame. Instead, the frame of the Adam and Eve panels are partly engaged because the left, right and bottom frame members are engaged but the curved top members are fixed onto the front and reverse of the panel with dowels, a rather unusual combination.7 Further interesting observations were made on the outer sides of the frames. The frame of the Adam panel has a rebate on the left framing element, whereas Eve’s frame shows a counterrebate on the right frame member.8 Today these two rebates are filled with wooden inserts (fig. 6.3). Originally, the unfilled rebate of the Eve frame probably fitted into the rebate of the Adam frame when the polyptych was closed. Closer examination shows that the Adam frame indeed has a form that fits exactly into the unfilled Eve frame. Furthermore, the right frame member of the Adam panel has a curved moulding that is mirrored by an identical moulding on the left frame member of the Eve panel. It is likely that this moulding had the

99

the original appearance of the panels and frames of the ghent altarpiece

1 2 1

2

2

2 1

2

2

Fig. 6.2 Unframed Eve panel, documentation of the 1950 conservation programme. Number 1 indicates the top curved frame element that is attached to the panel with dowels; 2 indicates the left, right and bottom frame elements with grooves

negative shape of a column that would have been present on the frame of the centre panels. This column, placed between the frames of the Virgin Enthroned and the Deity Enthroned panels, would have fitted precisely into the negative moulded form of the Adam frame when the polyptych was closed. The same applies to the column on the frame on the other edge of the Deity Enthroned panel adjacent to the John the Baptist Enthroned panel, which fits into the negative shape of the Eve frame (fig. 6.4). It has long been debated whether the curved tops were actually intended to be rounded or whether they originally had a rectangular form and were modified in the past. When the altarpiece wings are closed they do not cover the centre panels completely but leave the blue corners of the interior panels uncovered. Furthermore, the construction of the partly engaged frames is unusual, as discussed above. Nevertheless, the research suggests that the curved wing panels are indeed original, for several reasons. First, the lengths of the shorter vertical member of the frame of the Adam and Eve panels do not appear to have been modified. Second, the bevel on the curved profile of the frame continues onto the top of the vertical

member. If the format was originally rectangular, changes would be visible, especially in this part. Traces of original tool marks probably caused by the use of a knife or chisel can be found on the top edge of the panel and the frame. No saw marks are present, which would have suggested reworking. Third, remnants of ground layers and red paint can also be seen on the parts on the top of the frame (fig. 6.5). During the conservation treatment, further research will be done on the polychromy and original format. A hypothesis on this divergent format is that the frames of the wing panels were not originally different from the centre panels, but that the outer top corners of the Virgin Enthroned, Deity Enthroned and John the Baptist Enthroned were altered at some point. As stated above, the blue paint in the corners of these three centre panels remains exposed when the altarpiece is closed. This light-blue paint, covering the areas above the gilded niches, was applied during the conservation treatment in the 1950s. During recent cleaning tests on the blue paint of the Virgin Enthroned panel, remnants of an original, darker blue azurite was discovered. Furthermore, traces of some kind of possibly original fixation holes are visible in the

100

aline genbrugge and jessica roeders

1 2

Fig. 6.3 Number 1 shows the rebate on the left frame element of the Adam panel that is nowadays filled with a wooden inset; 2 shows the curved hollow moulding in the right frame element of the Adam panel

1

XIII.

XIV.

II. III.

I.

2

XV.

1

IV.

VII.

XVI.

2

VI. V.

Fig. 6.4 Diagram representing the original fitting of the closed altarpiece, seen from the bottom (cross-section). Number 1 indicates the fitting of the rebates in the frames in the closed position; 2 indicates the curved moulding corresponding to the negative shape of the two columns

2 1 1

2

Fig. 6.5 Number 1 indicates the bevel in the curved part of the frame; 2 indicates the original tool traces on the top edge of the panel

paint layer on the top of the Virgin Enthroned panel. These traces could be a remnant of some sort of decorative element (fig. 6.6). Looking at other early altarpieces from that time, it is possible that the blue parts in the corners of the centre panels originally contained matt blue azurite paint and decorative elements (fig. 6.7).9 Even though the Adam and Eve panels and frames do not fit the form of the centre panels in its closed position, they fit exactly into the painted gilded niche in which the Deity Enthroned is depicted. These initial observations suggest that the altarpiece may have combined three-dimensional woodwork and threedimensional illusionistic effects in the painting

the original appearance of the panels and frames of the ghent altarpiece

101

Fig. 6.6 The arrow in the diagram suggests the applied decorative patterns. The arrows on the X-radiograph indicate the traces of fixation holes in the Virgin Enthroned panel.

itself. On this basis we can propose a preliminary hypothesis for the original appearance of the top register. Further research into the blue paint and possible traces of decorative elements will be carried out during the 2012-2017 conservation treatment in order to find out more about the original construction.

Later Modifications: the Adoration of the Lamb The panels of the central section of the altarpiece contain many traces of previous modifications and past interventions, especially the Adoration of the Lamb panel in the lower register. Like the other panels, the Adoration is made from Baltic oak.10 All the panels in the central section are painted on

Fig. 6.7 Reconstruction of decorative elements, with cleaning test in the blue paint in the corner of the Deity Enthroned panel.

102

aline genbrugge and jessica roeders

one side only, but the Adoration is the only one where the planks have a horizontal orientation. On the reverse of the panel, and around the rebate edges, the presence of original tool marks suggest that the rebates are original features.11 This could indicate that the panel was originally placed in an embedded frame (fig. 6.8). There are several intriguing dark areas on the reverse of the panel (fig. 6.8).12 Whether these are original or were caused by a previous intervention is somewhat unclear. They have long been explained as traces of the major fire in the cathedral that occurred in 1822.13A more likely hypothesis, discussed during the expert meetings, is that they were caused by surface dirt and oxidation of the wood on the forest floor, after the wood had been cut and split. Another hypothesis is that they may be original marks related to panel construction. The stains may relate to the preparation of the wood before the panel was sold. They might be part of a fifteenthcentury panel preparation procedure. This would involve ‘sealing’ the reverse of a panel by charring which in combination with a layer of read lead, creates a moisture barrier that could achieve equilibrium with the front of the panel (where a ground and paint layer was applied).14 Another remarkable feature of the reverse of the Adoration of the Lamb is the wooden blocks and laths, applied during later interventions. There are twelve laths glued and screwed onto the panel, perpendicular to the wood grain, and twenty-four inserted wooden blocks. Also, thick metal clamps span every joint to support them. To understand why and when these interventions took place, the 1950 X-radiographs were studied and compared to those taken in 1986. Archival findings and documentation from the 1950s treatment were also consulted (fig. 6.9). The twelve screwed and glued laths and the metal clamps predate the1950s conservation treatment. After examining them in greater detail, it appears that that the screws of the perpendicular laths and the clamps were applied at a similar time, probably in the early nineteenth century. It is

2

1

1

2

Fig. 6.8 Reverse of the Adoration of the Lamb panel, with detail of the rebate on the edge of the reverse (number 1) and detail of a black stain (number 2)

possible that this intervention took place after the fire in 1822. From the archival findings, we know that the twenty-four inserted wooden blocks were applied during the 1950s treatment in order to provide further reinforcement to the joins.

the original appearance of the panels and frames of the ghent altarpiece

2

1

2

1

103

2

3

3

1

Fig. 6.9 Left: X-radiograph, 1950; centre: X-radiograph, 1986; right: photo, 2010

Looking more closely at the reverse of the Adoration of the Lamb, traces of green and red paint are also visible. It seems that at some unknown point in history the reverse of all of the centre panels was covered with a layer of red lead. Photographs dating from the Second World War show the reverse of the centre panels: on the reverse of the Adoration of the Lamb, the red lead had already been partially scraped off. The red lead layer was removed completely in the conservation treatment in the 1950s (fig. 6.10). Remnants present in the wood grain suggest that the red lead layer extended over the reverse of the whole panel; however, it is not present in the wood grain of the rebates, except for some large drops. Perhaps the red lead was

applied when the painting was still in its (original) rebate frame. On the front of the panels, some of the effects of previous conservation treatments are visible, in combination with the effects of the climate conditions in the church. There is a slight vertical concave warp; this had already been observed in the 1950s and the tailor-made curved profile of the frame accommodated the deformation. Today, this profile no longer fits the curvature of the panel. The panel has straightened out, and is restricted by the curved profile. Moreover, the metal framing construction around the altarpiece applied after 1953 constrains the movements of the centre panel (fig. 6.11).

Fig. 6.10 During the 1950 restoration treatment; scraping of the red-lead layer

104

aline genbrugge and jessica roeders

Fig. 6.11 Picture on the left shows the concave warp of the Adoration of the Lamb panel; picture on the right shows the reverse of the Adoration of the Lamb panel blocked due to the metal framing system

On the front, the three joins are clearly visible in the paint layer, especially along the outer edges of the planks where they have opened slightly. The opening up of the joins is probably due to the constraints imposed by the old interventions and the framing system, which would have prevented the panel from moving in response to the natural fluctuations in relative humidity and temperature in the church. The panel will be treated and stabilized during the current conservation treatment. Conclusion The main goals of the research carried out on the Ghent Altarpiece between March and November 2010 were to understand the original construction and appearance of the panels and frames and to research the modifications of the different treatments of the past. This was done through close observation of original tool marks and later interventions. We also considered how these former interventions, along with the fluctuations in temperature and relative humidity, might have affected the condition of the panels and the paint layers.

Especially through the study of the outer and inner sides of the frame members of the Adam and Eve panels, important clues were found about the original frames of the centre panels, which were lost in the eighteenth century. We discovered that the rounded top members of the wing panels and the frames of the upper register are connected to the rectangular forms of the inner centre panels. The precise way in which these were linked will be studied during the current restoration campaign. We were also able to propose a preliminary hypothesis for the original appearance of the top register, which we hope to refine during the 2012-2017 conservation treatment. Another aim of this six-month research project was to establish the material condition of the Ghent Altarpiece panels and to make a treatment proposal for its structural conservation. This research will serve as a valuable starting point for the treatment and further technical research on the triptych.

the original appearance of the panels and frames of the ghent altarpiece

NOTES * We thank Simon Bobak, Livia Depuydt, Hélène Dubois, JeanAlbert Glatigny, Anne van Grevenstein, Ray Marchant, Renzo Meurs, Devy Ormond, Ron Spronk, Abby Vandivere, Hélène Verougstraete, and Antoine Wilmering. 1 George Bisacca (New York, Metropolitan Museum of Art), Al Brewer (London, Royal Collection), Sue Ann Chui (J. Paul Getty Museum), Livia Depuydt (KIK-IRPA), Hélène Dubois (KIK-IRPA), Paul van Duin (Amsterdam, Rijksmuseum), José de la Fuente (Madrid, Prado), Aline Genbrugge (KIK-IRPA), Anna van Grevenstein (UvA), Ingrid Hopfner (Vienna, Kunsthistorisches Museum), Ray Marchant (Cambridge, Hamilton Kerr Institute), Renzo Meurs (Amsterdam), Alan Miller (New York, Metropolitan Museum of Art), Béla Nagy (Budapest), Britta New (London, National Gallery,, Anne-Cathérine Olbrechts (Monumentenwacht Vlaanderen), Jessica Roeders (Haarlem, Frans Hals Museum), Andrea Santacesaria (Florence), Ron Spronk (Ontario, Queen’s University; Nijmegen, Radboud University,), HélèneVerougstraete (Université Catholique de Louvain-la-Neuve), Jørgen Wadum (Copenhagen, National Gallery of Denmark), and as an observer Antoine Wilmering (The Getty Foundation). 2 The research into the structural condition of the supports and frames can be accessed on http://closertovaneyck.kikirpa.be. 3 In 1934 the original panel was stolen; it was replaced in 1941 by a copy by Jeff van der Veken.

105

4 Copies of Adam and Eve (clad in animal skins) were made by Lagye, and the church received the copies of the wings by Cocxie, which are now on view at the entrance of the St Bavo’s Cathedral. (DeSchrijver, Marijnissen 1953, nn. 67, 68) . 5 See von Roenne in this volume. 6 Coremans 1953. 7 No other examples of this kind of partly engaged frame have so far been discovered. 8 A rebate is a recess on the sides of the frame members. The rebates on the outside of the Adam and Eve frames are L-shaped recesses (rebate and one counter-rebate). These L shapes are designed so that the two frame members fit into each other. 9 As for example, the Miraflores Altarpiece (Staatliche Museum, Berlin). 10 See Fraiture in this volume. 11 A rebate, or L shaped recess on the reverse of a painting, was used to fit the panel into the embedded frame. 12 These black stains are also present on the reverse of the other central panels. 13 The 1822 fire caused damage to the centre panels. The Adoration of the Lamb panel appears to have been split when it was pulled out of the altarpiece. Cinders and molten lead fell in the chapel. De Schryver, Marijnissen, 1953, nn. 44, 45, 46. 14 These hypotheses were formulated by various panel-painting conservators, namely Jean-Albert Glatigny, Allen Miller, George Bissaca, Al Brewer and Ray Marchant.

Fig. 7.1 Ghent Altarpiece, Adam and Eve panels

7

Small Hairs. Meaning and Material of a Multiple Detail in the Ghent Altarpiece’s Adam and Eve Panels Ann-Sophie Lehmann

ABSTRACT: Based on recent close-up examination of the Adam and Eve panels of the Ghent Altarpiece, this article examines the intricate and innovative painting techniques Jan van Eyck employed for the depiction of their body hair against the background of the appreciation of diligently painted hair as a trademark of Northern painting in numerous art theoretical sources on the one hand and the possible iconographical meaning of the representation of body hair in Early Netherlandish painting on the other. It is proposed that the technical sophistication invested in the ‘small hairs’ should not only be seen as the result of the observant eye of the painter or as touchstone of artistic excellence, but also as a pictorial means that – in tune with contemporary iconographies – supported the physical identification with the naked, human body and therefore with the deeds, sufferings and salvation of the biblical figures represented. Such a bodily identification constituted one of the main aims of devotional imagery in general, and of the foundation of which the altarpiece was part in particular.

—o— The golden tresses of angels, the bluish stubble of beards, the soft fur collars of precious gowns, the bristly pelt of beasts – admiration for the life-like rendering of hair has been a commonplace in descriptions of Jan van Eyck’s paintings from the sixteenth century to the present day. Not so when it comes to the body hair of the Adam and Eve figures in the Ghent Altarpiece (fig. 7.1). In fact, these hairs, which appear in great quantity on the

legs, pubis and chest of Adam and the pubis of Eve, have drawn out more criticism than any other aspect of the proverbial Eyckian realism, ranging from surprise and disbelief all the way to outright disgust. They also played a part in the attribution of the panels to Jan van Eyck, as they fitted the nineteenth-century construction of his image as uncompromising recorder of nature and therefore the ultimately more ‘modern’ painter of the two brothers.1 While a careful reconstruction of the reception of Adam and Eve is provided elsewhere,2 this contribution focuses on the initial motivations for the presence of the minutely painted body hair and investigates material and iconographical aspects in relation to the first and only written account that drew positive attention to its striking presence. De cleene aerkins The historian Marcus van Vaernewyck (15181569) was a prominent Ghent citizen, alderman, churchwarden, head of the guild, and member and secretary of the Maria ’t Eeren chamber of rhetoric.3 In his writings he repeatedly mentions the Ghent Altarpiece in different contexts, inspiring Ludwig Scheewe, one of the first scholars to study the reception of the Van Eyck brothers, to describe him as the Van Eyck scholar of the second half of

108

ann-sophie lehmann

the sixteenth century.4 The first description in Spieghel der Nederlandsche audtheyt (1568) is closely related to Lucas de Heere’s famous Ode to the Ghent Altarpiece, which may have served as a source.5 The second mention, in Van die beroerlicke tijden in die Nederlanden en voornamelick in Ghendt, 1566-1568, an account of the iconoclastic riots in Ghent that was not published until the nineteenth century, was motivated by the removal of the Ghent Altarpiece to safety and contains a number of interesting details.6 Van Vaernewyck first gives a brief explanation of all the panels in the upper register, including Adam and Eve, before expressing his admiration for the two nudes again in more detail: Maer van Adam ende Eva waer wonder te zegghen, niemant en zoude wel zonder twijffelen connen jugieren, tzelve anziende, of den eenen voet van Adam uuten platten tafereele steect of niet, ende zijnen rechter aerme ende handt, die hij up zijn burst lecht, schijnt van zijnen lijve duerluchtich zijnde. Dlichaem es oock zoo vleeschachtich dat schijnt vleesch te wesen. Niet alleen de aderen en zijn daer inne gheconterfeet zeer levende, maer ooc die cleene aerkins die een meinsche uuten lichame groijen.7 But wondrous things can be said about Adam and Eve. No one who looks at them could say for certain whether one of Adam’s feet extends beyond the flat panel or not, and his right [sic] arm and hand, which he rests upon his chest, seem to radiate light from within his body. The body is so flesh-like that it appears to be flesh. Not only the veins are drawn in it after life, but also the small hairs, which in a man grow from the body. (author’s translation) The passage ends with a brief contemplation of the nature of the fruit Eve is holding, which Van Vaernewyck, surprisingly, thought to be a fig; an interpretation that is in line with scripture but not with the visual appearance of the fruit, which looks mostly like a lemon, or a generically ‘bitter fruit’.8 The description of the bodies, however, is full of

acute and unique observations. Not only is Van Vaernewyck the first author to mention the trompel’oeil effect of Adam’s foot,9 but following the phrasing of Lucas de Heere and in tune with the contemporary sensitivity for the artistic rendering of skin, he praises the flesh-like appearance of the body, yet even elucidates the lifelike effect through a description of the peculiar visual effect of the semi-transparency of human skin, which Van Eyck effectively imitated by the layering of oil paint and the highlighted edges of the shadows on Adam’s arms and hands; an effect that also occurs in Eve.10 Apart from these painterly effects, Van Vaernewyck also singles out the more graphic elements that enhance the mimetic appearance of the body surface: the veins and the ‘small hairs’ which grow on Adam as they do on real men. His positive acknowledgement of the body hair remains exceptional and is a good starting point to investigate the possible meanings of the depiction of body hair for earlymodern viewers by way of a small cultural history of this literally tiny subject as well as a description of the materials and techniques involved to create the hairs in the paintings. Touching Hair In the summer of 2010 I was able to view the Adam and Eve panels up close during the photographic documentation of the Ghent Altarpiece that preceded the conservation campaign. Having studied the commentaries on the figures in depth, I was well aware of the effect the hairs had had on previous writers, who had seen the panels in their original setting in the chapel or, closer, during their display in the Musées royaux des Beaux-Arts in Brussels between 1863 and 1920, which prompted Eugène Fromentin, for instance, to characterize them as ‘two horribly hairy savage beings’ in his Les maîtres d’autrefois of 1877.11 Nevertheless, to observe the sheer endless amount of individual little hairs on their bodies was amazing, as is the preservation of the hairs over the centuries.12 Many very fine, single brushstrokes have been painted on the lower arms, legs, as well as breast and pubis of

109

small hairs

Adam – Eve only has pubic hair – using different shades ranging from black to light brown. Some hairs have a slight relief and others are flatter, possibly due to variations in the thickness of the paint. In some areas, especially in the hairs on the breast and the legs near the fig leaf, the paint strokes have beaded (fig. 7.2). During the prior conservation campaigns in 1951/1952 and 1986/1987, this effect led Coremans and Brinkman to speculate respectively about an emulsion or water-based paint medium for this particular element.13 The consistent use of oil throughout the altarpiece on the whole makes these theories appear unlikely. In fact, a contemporary German painting of St Veronica with the Sudarium in the National Gallery London shows the exact same phenomenon in Christ’s beard. As the authors of a recent technical study of the painting explain, the beading occurred because the hairs were applied onto a dry paint layer.14 If this is correct, it can be inferred that Van Eyck also painted those hairs when the flesh colour had already dried, suggesting that the hairs were painted in different stages. While the hairs on arms and legs were painted onto the skin one by one, which can only have been achieved with a very fine brush, an additional technique was used for the denser pubic area of Adam. Next to brush strokes, fine lines have been scratched into the wet paint, revealing the lighter ground of the flesh colour or the underpaint, thereby creating the impression of individually highlighted hairs (figs 7.3a-b). The same technique has been employed in the hair of the Singing Angels and in the book the Prophet Micah holds above the Annunciation on the right outer panel of the altarpiece. Here, a thick yellow paint has been applied over a brown base tone – inverting the strategy employed for the hair – the scratches indicating the individual pages of the top edge of the closed book (fig. 7.4). The technique of scratching has also been observed in the works of other Early Netherlandish painters,15 but Rembrandt is known for its most virtuoso use, such as in his early self-portrait of around

Fig. 7.2 Ghent Altarpiece, Adam, detail, Adam’s leg

1628 in which his unruly curls are boldly enhanced with long scrapes into the wet paint (fig. 7.5a).16 The reason that scratches are particularly evocative of hair is due to the oil medium’s particular affordance of modelling and layering, that allows not only for additive but also subtractive techniques: when the darker paint layer has dried slightly, the paint will not flow back into the scratches that reveal the lighter ground (or vice versa), but form soft etches at its border, creating a surface relief as well as a highlighting effect that together approximate the texture and the reflection pattern of the depicted object (fig. 7.5b). These details show that the knowledge and artistic command of the various properties of oil were fully developed already in this early phase of the new medium.17 Van Vaernewyck employs a specific term to describe the depiction of the veins and hairs – conterfeyten – that in fifteenth and sixteenth century sources is generally used in the context of drawing from life and the making of portraits in particular.18 On a technical level, the term highlights the pictorial creation of the hairs lines with single, linear brush strokes and scratches; techniques, which seem to have been transferred from the domain of drawing to that of painting. Vaernewyck’s specification of this ‘counterfeit’, the portrayal of veins and hairs as ‘very alive’ (zeer levende), may suggest drawing after a life model, but it first and foremost tells us that they literally draw the figure to life;

110

ann-sophie lehmann

Fig. 7.3a Ghent Altarpiece, Adam, detail, Adam’s pubic hair

small hairs

Fig. 7.3b Ghent Altarpiece, Adam, close-up detail, Adam’s pubic hair

under the skin (veins) and on top of it (hairs).19 While they were certainly observed in real life (though their execution may have been a matter of painting uyt den gheest, as Karel van Mander characterized the repetitive action involved in the

111

painting of hair and also leaves), the hair in particular also draws the viewer in by triggering a perceptive response that exceeds vision. Cognitive research into the phenomenon of multi-sensorial perception and in particular the relation between optic and haptic perception has demonstrated that when humans look at soft, furry surfaces their haptic sensation is triggered more strongly than when observing smooth, hard surfaces.20 Such research hands us an explanation for the common notion that the things in a painting look ‘so real we want to touch them’, brilliantly evoked in Dürer’s long-haired Self-Portrait of 1500 (Munich, Alte Pinakothek), in which he pinches a piece of his furry collar with two fingers, as if to accentuate the tactile longing he engenders in the viewer (fig. 7.6).21 It also offers scientific support for the differentiation between haptic and optic styles in painting, developed by Alois Riegl and

Fig. 7.4 Ghent Altarpiece, the Prophet Micah, detail, the prophet’s book

112

ann-sophie lehmann

Figs 7.5a-b Rembrandt Harmensz. van Rijn, Self-Portrait, c.1628, oil on panel, 22.6 cm x 18.7 cm, Amsterdam, Rijksmuseum (object no. SK-A-4691)

small hairs

Fig. 7.5b Detail of Fig. 7.5a

Heinrich Wölfflin around 1900 and revived in recent phenomenological approaches to art history and media studies.22 In the case of Adam and Eve, the plasticity of the oil medium enhances the haptic effect also on a material level as it makes the tiny brushstrokes literally extend from the picture plane into the viewer’s space, an effect that is even stronger when the hairs are painted so as to indicate their rising off the body surface as on Adam’s right lower arm and wrist (fig. 7.7). Acute observation and its mimetic translation into depiction through treating the apparently repetitive motif of hundreds of single hairs with a subtle variation in tone, thickness of medium and the additive and subtractive application of paint, make for hairs that evoke life-likeness in every respect. So why did Van Eyck concentrate his mimetic forces in such a manner that no viewer could escape the hairiness? Hair as a Distinction between the Naked and the Nude At first sight, body hair seems rather out of place in the canon of Western art. From antique sculpture to Botticelli and Michelangelo, Rubens and Ingres, male bodies usually have only pubic hair but no other body hair, while women’s bodies appear

113

to have no body hair at all, and are, as George Didi-Huberman has phrased it, successfully ‘denuded’ by the absence of any direct sign of sexuality.23 Not surprisingly, the phenomenon is rarely considered and those few who have given it a thought – among them Diderot and Hogarth – explain the absence of body hair in painting as a logical continuation of the hairless appearance of antique statues, which in turn is taken to reflect the actual custom of depilation in antiquity. In addition to such conflated anthropological and aesthetic considerations, socio-biological and psychological reasons for the lack of body hair have been noted.24 Sources dealing with cosmetic hair removal in antiquity, the Middle Ages and the Renaissance support the idea that hairlessness was a common ideal, enforced by the smooth, hairless representations of youth in art.25 However, one notes variations within the range of that ideal, and at times body hair could be integrated into the aesthetic parameters of a work of art. Female pubic hair, for instance, can be seen on red-figured vases under the transparent chitons of Helen or Penthesilea. In the Christian tradition, early medieval miniatures of the Fall of Man often show Adam and Eve with pubic hair.26 Though nineteenthcentury critics perceived the body hair of the Adam and Eve panels as an extremely unusual feature, it is in fact not the exception but the rule in Netherlandish and German nudes of the fifteenth and first half of the sixteenth century. Many naked or partly naked male protagonists in early Netherlandish painting, ranging from John the Baptist to St Sebastian, are adorned with hairs on their legs, chests, armpits and pubis. Christ, too, is shown with body hair, for example in the Descent from the Cross in Rogier van der Weyden’s Miraflores Altarpiece, where he has fine, individual hairs painted around his nipples and on his chest and legs. Also mythological males, such as Jan Gossaert’s Neptune in Neptune and Amphitrite (1516, Berlin, Gemäldegalerie) have pubic hair,27 as do female nudes, religious as well as mythological, and in some cases they even have hair on their legs, like Hans

114

ann-sophie lehmann

Fig. 7.6 Albrecht Dürer, Self-Portrait with Fur-Trimmed Robe, 1500, limewood, 66.7.1 x 48.9 cm, Munich, Alte Pinakothek (inv. no. 537)

small hairs

115

Fig. 7.7 Ghent Altarpiece, Adam, detail, Adam’s wrist

Memling’s Eve on the wings of the St John Altarpiece (c.1485, Vienna, Kunsthistorisches Museum). Following Kenneth Clark’s famous and much criticized distinction of the ‘naked’ and the ‘nude’, the appearance of body hair in German and Netherlandish art could be interpreted as a national distinction and expression of a more ‘realistic’ approach to the human body in the North as opposed to the ideal renderings of antiquity and the Italian Renaissance.28 Yet not only is this interpretation wrong for a number of reasons (many Italian nudes also have body hair29 and its presence in the North by no means excludes idealization, as will be explained below), it also offers no satisfying explanation. It is more fruitful to try and understand the depiction of body hair in conjunction with other, contemporary iconographies. Empathic, Wild and Erotic Hairs At least three iconographic categories can be distinguished, in each of which body hair has a different meaning and accordingly a potentially different effect on the viewer. The first encourages identification between the image and the viewer and could be called the empathic category. In images of religious figures, and especially of Christ, the hair

intensifies the impression that the body depicted is that of a real human being and evokes, as in the case of the realistic depiction of wounds, the viewer’s empathy. Van Vaernewyck’s wonderment at Adam’s ‘little hairs which in a man grow from the body’ contains a strong element of recognition, which is expressed in the fact that he concentrates his scrutinizing observations on Adam, who appears to him as ‘a man’, in other words a convincing representative of the category of the human male, to which also Vaernewyck himself belongs. Like Adam, Van Vaernewyck and every man grows hairs on his legs. That the mimetic rendering of the human body in religious imagery assists in the mental and physical identification with the deeds, sufferings or fate of the figures represented has been claimed as one of the central aims of early modern religious imagery.30 Identification is indeed greatly enhanced by the way that Van Eyck chose to paint the hairs and emphasize their haptic appeal described above. In fact, both the material act of painting the hair, brushstroke by brushstroke, and the perceptual act of experiencing the hair with visual and tactile senses are necessarily embodied as both painter and onlooker work with their bodies. It is in this literal networking between pictures

116

ann-sophie lehmann

and bodies that belief was mediated and is still accessible today.31 The historical belief in bodily resurrection that Adam and Eve embody for the donors in fifteenth-century Ghent and which informed the life-like rendering of the panels, may very well be experienced through a ‘modern’ belief in the power of art when observing the detailed body surface. The second iconographic tradition associated with body hair is that of wild men and women.32 Here, hair growth in an excessive form does not create proximity but distance, because the hirsute condition signifies the otherness of human beings who are still in a primordial state. In the earlymodern period, however, this particular wildness is not necessarily associated with primitivism in the sense of a lower stage of human evolution but rather with the paradisiacal state of the noble savage, to which we may trace our ancestry.33 In its less specific appearances, excessive hairiness has a negative connotation as an attribute of members of lower social classes such as slaves, servants or ‘bad’ men, for example the bad thief or the soldiers at crucifixions, who apart from darker skins or minor deformities like warts or crooked noses, frequently have a lot of body hair.34 The ambivalent iconography of innocence as well as primitivism was still being echoed in Eugène Fromentin’s characterization of Adam and Eve as ‘hairy savages’. The third iconography associates male and especially female pubic and auxiliary hair with sexuality. Comparison with erotic poetry written in France and the Netherlands in the fourteenth and fifteenth centuries shows that body hair formed an important part of the sensual body. A late fourteenth-century Brabantine poem, for example, celebrates a girl for her beautiful yellow braids, red lips, white teeth and snow-white body. Only in two places is she ‘brown’: under her arms and ‘down below’.35 The border between sensual description and pornographic directness is more fluid in poems than in the visual arts, where the latter is mostly restricted to the medium of the print. Explicit engravings by Aldegrever, the Beham brothers or

Marcantonio Raimondi often feature pubic hair in an overtly erotic fashion. However, female nudes in biblical, allegorical and mythical paintings, such as works by Lucas Cranach and Hans Baldung Grien, often display pubic hair too, but in an idealized fashion, neatly coiffured and under a see-through veil, as if to stimulate a voyeuristic gaze.36 In smallscale sculptures of Adam and Eve by artists like Konrad Meit and Christoph Wydyz, as well as in Dürer’s engraving of 1504, Eve has nicely curled pubic hair corresponding to the equally orderly presence of the hair of her male pendant. It is crucial to note that sensuality and idealization do not exclude but complement each other in this particular detail. This is also the case for Jan van Eyck’s Eve: though she has been deemed anything but ideal by many, the distribution of body hair, which she has on her pubis but not on her legs, is a form of idealization. The presence of body hair can thus emphasize different and apparently opposing features of the empathic, wild or sensual body. In Van Eyck’s Adam and Eve panels all three of these aspects work together. First of all, the hairs invite identification with the figures as real humans; they underline the ancestral relation of the first parents with the present viewer; and last but not least they enhance the sensual presence of the art works. Hair in Theory Body hair not only appeared in works of art as a bearer of such manifold connotations, but in arttheoretical discourses as well. In the middle of the fifteenth century, Angelo Decembrio, in a fictive dialogue about medals between two art lovers at the court of Leonello d’Este, has his main character, the duke, make a plea for the artistic relevance of the naked body down to the last detail. The human body would always draw more attention than any garments, jewellery or ornament, and the viewer would appreciate the correct representation of sinews, muscles and veins, skin, hair and body hair.37 In the sixteenth century, body hair features in opinions about the paragone between

small hairs

painting and sculpture collected by Benedetto Varchi.38 Giorgio Vasari, for instance, in his letters to Varchi, takes the side of painting because it can represent, among other things, lifelike human flesh, the piumosità (downiness) of hair and the morbidezza (softness) of beards.39 North of the Alps, critics admired Albrecht Dürer for his lifelike rendering of hair.40 Van Mander’s description of the Ghent Altarpiece, which is based on De Heere and Van Vaernewyck, takes up the admiration for the hairs: Ook de hayrkens in de beelden, in de peerdt-sterren [sterten] en manen soudemen schier mogen tellen en soo dunne en aerdich gedaen dat het alle Constenaers verwondert (‘Also, one could almost count the little hairs in the figures, the horses’ tails and manes, so finely and skilfully painted that they are admired by all artists’).41 In the seventeenth century, finally, Joachim von Sandrart perceived this virtue as an element of a national style, traceable to Jan van Eyck: Unsere Teutschen haben mit so sonderbarer Arbeitsamkeit ihre Werke vollbracht: wie zu sehen in den Stucken Albrecht Dürers, Holbeins, Lucas van Eyck und anderer in welchen auch die geringsten Haare ganz klar und rein ausgebildet erscheinen; das dann in der Nähe wohl zu sehen ist. (‘Our Germans have completed their works with such remarkable diligence, as can be seen in the pieces by Albrecht Dürer, Holbein, Lucas [sic] van Eyck and others, in which even the smallest hairs seem to be clearly and purely depicted, as can be seen from close to’).42 To depict body hair, then, was not a token of a proto-scientific understanding of nature, which inspired the painter to faithfully reproduce it as it appeared to his eye, which is how Max Friedländer made sense of Adam and Eve’s detailed bodies and his reading persists in contemporary interpretations.43 Rather, the hair can be seen as a small, but nonetheless meaningful element in the dense field of religious patronage, material innovation, precise observation, iconographical traditions, and aesthetic appeal in which Jan van Eyck’s art emerged.

117

NOTES * I would like to thank Anne van Grevenstein for welcoming the team of the Impact of Oil Research Project (Netherlands Organization for Scientific Research 2007-2012) during the start of the conservation campaign of the Ghent Altarpiece in 2010 and Abbie Vandivere and Esther van Duin for sharing their insights into Van Eyck’s painting techniques with me. I would also like to thank Xochitl Flores-Marcial who copy-edited the manuscript for me during my stay as a Getty Scholar at the Getty Research Institute. 1 Recent scholarship into the genesis of the Ghent Altarpiece confirms this attribution, though based on the more solid grounds of archival and technical research, which places the paintings among the last panels to be completed. 2 Lehmann 2004 and 2017. 3 Van Nuffel 1979, pp. 795-809; Lamont 2005. 4 Scheewe 1933, pp. 63-66; Van Vaernewyck 1965, pp. 107-115. 5 Van Vaernewyck 1568, fols. 117r-119r ; Scheewe 1933, p. 64; Dhanens 1965, p. 110; Waterschoot 1966, pp. 113-115; Melion 1991, pp. 139-141. 6 Van Vaernewyck 1566-1568 (1872), pp. 143-146. 7 Ibid., p. 143. 8 Art historians have identified the fruit as a small pineapple (Leclerq 1861, p. 285), a lemon (Kaemmerer 1898, p. 26), an etrog (Dequeker 1986, pp. 98-110), a so-called pomus adami (Snyder 1976, pp. 511-515; Harbison 1991, p. 37), and as a Seville or bitter orange (Segal 1984, pp. 403-420). 9 X-ray photographs taken in the 1950s revealed that Van Eyck had originally planned the foot to be parallel to the picture plane; see Coremans 1953, p. 98. On the historiography and interpretation of this detail, see Lehmann 2012, pp. 8-13. 10 See Lehmann 2008 for an interpretation of the term duerluchtich and its relationship to Van Mander’s use of the term gloeien to describe Hendrik Goltzius’s manner of painting flesh. 11 Fromentin 1948, p. 237. 12 On the material state of the painting surface of the panels, see the report by Van Grevenstein-Kruse et al. 2011, http://closertovaneyck.kikirpa.be/#home/sub=documents. 13 Coremans 1953, p. 80 and Brinkman 1993, p. 219 and n. 69. 14 Spring et al. 2012, p. 90 and fig. 8 for a detail of the beard. 15 Sander 1992, p. 75. 16 Bruyn, Ernst van de Wetering et al. 2005, p. 158 ff. 17 See the papers on transparency and texturing by respectively Bol and Vandivere in this volume. 18 Parshall 1993, pp. 554-579. 19 This distinction between drawing after life and drawing from life has been convincingly developed by Turel 2011, pp.163-182. 20 Lederman, Klatzky 2004, pp. 107-122. 21 For an in-depth discussion see Körner 1993. The early-modern relationship between vision and touch is further investigated by Scribner 1998, pp. 93-117. 22 See Marks 2000, Sobchack 2000, pp. 1-29. 23 Didi-Huberman 2006, pp. 37-52. 24 See Endres 2004, pp. 17–38; Hollander 1978; Lehmann 2009. 25 See for instance Kilmer 1982, pp. 104-112 and Bain 1982, pp. 7-10. 26 For a strikingly early discussion see Müller 1906. 27 Schrader 2010, pp. 57-67. 28 Clark 1956. For a recent reappraisal of the nude in Northern art see De Clippel, Van Cauteren, Van der Stighelen 2011. 29 For example Agnolo Bronzino’s infamous Venus in the Allegory with Venus and Cupid (c.1545, London, National Gallery) has pubic hair, although it has been rendered almost invisible by overpainting and successive cleanings, and is only discernible in the original.

118

ann-sophie lehmann

30 This argument was first developed by Ringbom 1965 and Marrow 1979. For more recent investigations into the empathic and tactile appeal of Early Netherlandish paintings, see Baert 2010, pp. 65-90 and Knight Powell 2012, esp. ch. 6. 31 On the relation between art work and spectator, see for instance Shearman 1992 and Frangenberg 2006. 32 See Husband, Gilmore-House 1980. 33 Leitch 2008, pp. 283-302; Moser 1998; Berman 1999, pp. 288-304. A special iconographic case is Mary Magdalene covered with body hair; see Baert 2003, pp. 46-60. 34 See Mellinkoff 1993; Vandenbroeck 1987. 35 Het es brun te selker steden Onder die ocxselen ende beneden. Pornographic poetry features many metaphorical references to the female sex and pubic hair, for example François Villon’s ‘silky hunting net suspended between two strong thighs’; both cited in Uytven 1998, pp. 61-63. 36 See Hollander 1993, esp. pp. 136-148; Werner 2007, pp. 89-109; Kopp-Schmidt 2010, pp. 41-48. 37 Baxandall 1963, pp. 304-326, esp. p. 316: ‘So that if you saw an eagle crowned, or two-headed and looking out on both sides, or an

elephant carrying a castle, or an unusually beautiful stag with gold collar and garlanded antlers, or leopards and tigers bridled and harnessed to a chariot with the triumphant Bacchus, and him half-naked, you would pay more attention to the subtlety of the features of the face and bare body than to the clothes and trappings. You would study the way in which sinews or muscles fit together, the circuits and tensions of the veins, the representation of skin, hair or plumage’. Baxandall’s translations ‘hair or plumage’ is not entirely correct, because man has no plumage. The original, quae cutis pilorum plumarumque, can better be translated as hairs (pilus denoting the individual hair) and down, as of a beard or body hair. 38 Mendelsohn 1982. 39 Che dirò io della piumosità de’ capegli e della morbidezza delle barbe, quoted from Barocchi 1978, p. 496. 40 Körner 1993. 41 Van Mander 1994-1999, vol. 1, fol. 200v; see also vol. 2, p. 207. 42 Von Sandrart 1994, bk. 3, pt. 1, ch. 7, p. 72. 43 Friedländer 1967, pp. 76-77; Seidel 2008, pp. 2-18.

Fig. 8.1 Retable de l’Agneau Mystique, volets d’Adam et Eve, Gand, Cathédrale Saint-Bavon

8

Le rôle du dessin sous-jacent et de l’ébauche préparatoire au lavis dans la genèse des peintures de l’Agneau Mystique Catheline Périer-D’Ieteren

ABSTRACT: The Role of the Underdrawing and the Preparatory Wash ébauche in the Genesis of the Paintings making up the Ghent Altarpiece  Paintings by Jan van Eyck are the result of a complex creative process that starts with a very elaborate underdrawing and continues right up to the final painting stage. The composition of the Ghent Altarpiece is so perfect that without scientific methods of examination it is impossible to guess at the variety in the drawing and the multiple modifications made by the master in the course of execution. The panels depicting Adam and Eve are the focus of this study. An initial linear drawing of placement lines and hatching for modelling is superposed by a wash phase, which helps establish the sense of volume. A few unusual examples of this manner of working have been brought to light since the publication of L’Agneau Mystique au laboratoire in 1953. The new infrared reflectograms have shown that this method was more widespread.These images also reveal Van Eyck’s systematic use of a vigorous black line correcting or reinforcing the forms of the first sketches as well as differences in execution that betray the intervention of assistants. The Portrait of Giovanni(?) Arnolfini and his Wife and the Virgin and Child with Canon Joris van der Paele show the same propensity for making several compositional changes. Finally, in the latter painting, the probable use of a pounced drawing is shown for the decorative motifs on the dais and orphrey as well as an astonishingly dynamic style in the areas of hatching. The study also reveals unexpected facets in the artistic personality of Van Eyck through a new comparative examination of the underdrawings of several autograph paintings.

—o—

Le dessin sous-jacent de Jan van Eyck a déjà fait l’objet de multiples études depuis l’Agneau Mystique au laboratoire en 1953.1 Aussi, l’objectif de cette communication est de se limiter à exploiter les données nouvelles fournies en 2010/11 par l’imagerie scientifique et en particulier par la réflectographie infra rouge.2 La publication de Van Asperen de Boer A Scientific Re-Examination of the Ghent Altarpiece en 1979 constitue naturellement, pour l’Agneau Mystique, le deuxième ouvrage de référence qui a alimenté les réflexions comparatives sur l’ensemble des panneaux. Un bref survol de la fortune critique des publications sur le dessin sous-jacent révèle des démarches méthodologiques différentes. L’ouvrage collectif sous la direction de Paul Coremans, l’Agneau Mystique au laboratoire, campe les problèmes en caractérisant le dessin de Van Eyck. Johannes Taubert,3 partant des infrarouges alors publiés, compare le stade du dessin à celui de l’exécution picturale et introduit ainsi pour la première fois une réflexion sur la démarche créative du maître. Van Asperen de Boer,4 quant à lui, examine la totalité du retable en réflectographie infrarouge qui était, dans les années septante, une nouvelle méthode d’investigation du dessin sous-jacent, et offre aux chercheurs un ensemble important de documents inédits. La richesse de l’information

122

catheline périer-d ’ ieteren

ainsi apportée est grande, mais reste jusqu’à ce jour étrangement peu exploitée. Enfin, depuis 1979, une série d’études ponctuelles sur des œuvres majeures du peintre ont vu le jour à l’occasion de restaurations, de catalogues critiques des collections muséales ou d’expositions.5 Notre démarche sera double. Premièrement, étudier le dessin du retable de Gand dans le détail, par comparaison avec celui des peintures autographes de Jan van Eyck afin de réaliser un relevé de ses particularités et de mieux appréhender les phases de sa genèse. Deuxièmement, examiner ce dessin en tant que partie d’un tout pour tenter de déterminer sur la base de ses caractéristiques d’écriture la part prise par le maître de celle de ses collaborateurs. Nous centrerons notre démonstration sur les panneaux d’Adam et Eve (fig. 8.1), les plus représentatifs de la manière de Jan van Eyck et dont le caractère autographe est communément accepté. Nous confronterons leurs particularités d’écriture d’abord à celles de deux peintures signées et datées du maître, les mieux documentées à ce jour : le Mariage des Arnolfini de la National Gallery de Londres6 et la Vierge au Chanoine Van der Paele du musée Groeninge à Bruges7 pour montrer les analogies qu’elles partagent. Ensuite, nous les comparerons à celles des dessins sous-jacents des autres panneaux de l’Agneau Mystique afin de mettre en exergue les différents styles de dessin qui coexistent au sein même du retable. Le dessin des deux figures d’Adam et Eve (figs 8.2a-b) est réalisé au pinceau avec un médium liquide. Abondant, il est soumis à plusieurs corrections de mise en place des formes, elles-mêmes encore reprises en cours d’exécution picturale. Le dessin des corps est aussi élaboré dans les deux panneaux. Par contre, la préfiguration du modelé est peu poussée dans le visage d’Adam à la différence de celui d’Eve. Ce dernier est plus travaillé et s’apparente dans le tracé et la distribution des hachures au visage de la Vierge au chanoine Van der

Paele (fig. 8.2c) sans toutefois, vu les différences de dimensions, être aussi soigné dans l’exécution. Au sein du retable, le visage d’Eve fait d’ailleurs exception, le dessin de ceux de la Vierge, de Dieu le Père et du donateur étant réduit selon une tendance « minimaliste » relevée dans la plupart des portraits de Van Eyck, en particulier les portraits autonomes.8 Dans le visage d’Adam, seule la mise en place des paupières, des yeux et du nez a vraiment retenu l’attention du maître selon un processus qui se généralisera dans ses peintures. Le dessin du visage d’Eve (fig. 8.2b) se particularise par des plans de longues hachures parallèles qui préparent le modelé des ombres. Elles suivent le sens des formes sur la joue droite tandis qu’elles se superposent en se croisant sur le menton. Un vigoureux trait foncé cerne la partie droite du visage en resserrant un peu son volume. Il se prolonge le long du corps à hauteur du cou, de l’épaule, de la courbe du sein, du bras et de la jambe droite en corrigeant un premier tracé moins appuyé (fig. 8.3b). Il reprend aussi les contours des doigts des deux mains. Un réseau de hachures parallèles très souples dans leur tracé au pinceau module l’ombre de la poitrine et du ventre. Un léger lavis paraît s’y superposer pour situer l’ombre la plus profonde et préparer l’effet pictural final. Un style de dessin tout à fait analogue se retrouve dans la figure d’Adam. La partie droite du corps est en effet ombrée par de longues hachures serrées qui s’espacent à gauche dans les demi-teintes du buste (fig. 8.3a). Un lavis, étendu à hauteur du bras gauche sur les hachures disposées en faisceaux vient compléter les modulations de l’ombre. Un large trait noir marque aussi les contours en les soumettant à de légers glissements. La position de l’épaule (fig. 8.4a) a été revue comme l’indiquent les lignes successives à la fois nerveuses et libres dans la recherche créative. La position des pieds est totalement changée tandis que le dessin des doigts de pied est renforcé par les mêmes lignes noires énergiques (fig. 8.4b), la volonté évidente du peintre étant ici de souligner le surgissement du volume dans l’espace. Cette étude poussée du mouvement

le rôle du dessin sous-jacent et de l ’ ébauche préparatoire au lavis

123

Figs 8.2a-c (a-b) Retable de l’Agneau Mystique, détails des visages d’Adam et Eve, RIR, (c) Jan van Eyck, Vierge au chanoine Van der Paele, détail visage de la Vierge, RIR, Bruges, Groeningemuseum

est une caractéristique majeure de la genèse des compositions eyckiennes qui prolonge sa réflexion en continuum de la première phase du dessin sousjacent au dernier stade de l’exécution picturale. Par ces modifications constantes, Van Eyck cherche à améliorer les rapports entre figures et espace, ce qu’illustre bien la position d’Adam, et à figer le mouvement au sein de ses compositions immobiles à l’instant où l’attitude est la plus proche de la réalité « du vécu ». Cette préoccupation spatiale semble l’emporter au niveau du dessin sur le détail des formes, qui n’est, le plus souvent, précisé qu’au stade pictural. C’est le cas du volume des mains et des doigts d’Adam et Eve dont les contours sont repris par un large trait pour fixer leur emplacement après plusieurs ébauches de mise en place mais sans qu’un soin particulier soit accordé à leur configuration. Ce même type de corrections s’applique aux autres personnages du retable. L’élément neuf essentiel qui ressort de cette première analyse est le recourt systématique de Van Eyck à ce large trait noir assuré qui corrige et confirme les contours ou l’emplacement des formes par rapport à la première ébauche. Ce procédé se répète invariablement dans toutes ses compositions. Souvent liés à des modifications préalables

successives de mise en place des formes, nous interpréterions ces tracés comme la marque d’un choix arrêté par Van Eyck au dernier stade du dessin sousjacent. Cela n’exclut toutefois pas que ces mêmes parties soient parfois encore soumises à corrections au stade du dessin comme de l’exécution picturale. On observe un deuxième procédé récurrent celui de l’usage de pigments très dilués que nous désignerons comme un lavis, apparaissant soit sous la forme d’une large ligne soit d’une surface grisée. Dans notre livre sur Colyn de Coter et la technique picturale des peintres flamands,9 nous avions relevé cette particularité technique, mais sans nous y attarder à défaut d’un nombre suffisant de documents de référence. Aujourd’hui, nous pensons pouvoir confirmer cette pratique non seulement dans l’Agneau Mystique, mais aussi dans les autres peintures autographes de Van Eyck. Si le terme de lavis demande peut-être à être revu, la fonction de ces lignes et de ces plages pourrait être, comme nous l’avions suggéré en 1985, de densifier les ombres des modelés translucides et nuancés de Van Eyck, sans perturber leur aspect lisse final. Cet usage généralisé s’inscrit parfaitement dans l’adroit jeu croisé des transparences et des densités de la matière recherché par le maître. Dans le retable, les lignes en lavis marquent les creux des plis des

124

catheline périer-d ’ ieteren

Figs 8.3a-b Retable de l’Agneau Mystique, détails des lignes en lavis sur les corps d’Adam et Eve, RIR

vêtements, comme dans celui de la Vierge de l’Annonciation, et certains contours, tandis que les plages grisées notamment dans les carnations d’Adam (figs 8.5a-b), sont étendues comme un léger voile sur le réseau sous-jacent de hachures parallèles ou entrecroisées. Ces lavis participent de cette manière à la gradation progressive de l’ombre à la lumière et donc au rendu d’un volume parfait. On constate que le degré d’élaboration du dessin sous-jacent d’Adam et Eve est plus poussé que celui de la majorité des autres parties du retable. Il

convient dès lors de tenter de l’expliquer. Nous l’attribuerions au caractère tout à fait novateur dans la peinture flamande du premier tiers du XVe siècle de ces nus grandeur nature – qui ont d’ailleurs choqué à l’époque moderne – encadrant sur la face du retable la composition traditionnelle de la Deisis et des Anges Chanteurs et Musiciens. Van Eyck se devait donc de tendre à la perfection dans le rendu anatomique des figures. C’est pourquoi, il en précise les contours d’un trait noir correctif et prépare leur modelé avec un soin particulier. Il insuffle ainsi au registre supérieur du retable,

le rôle du dessin sous-jacent et de l ’ ébauche préparatoire au lavis

125

Figs 8.4a-b Retable de l’Agneau Mystique, détails de l’épaule et du pied d’Adam, RIR

Figs 8.5a-b Retable de l’Agneau Mystique, détail du bras d’Adam, lumière ordinaire et RIR

d’aspect hiératique, une modernité surprenante parvenant à traduire avec la même maîtrise le mouvement volontaire d’Adam – d’où la correction du pied – et son expression de culpabilité interrogative. Cette expression, préfigurée dans le dessin

n’est cependant réellement donnée qu’au stade final de l’exécution picturale par de petits changements ponctuels (figs 8.6a-b). Ainsi, il peint le rebord inférieur de l’œil gauche, en recule le coin vers la gauche et ne conserve qu’une partie de

126

catheline périer-d ’ ieteren

Figs 8.6a-b Retable de l’Agneau Mystique, détails des yeux, (a) Adam ; (b) Joos Vijd, lumière ordinaire et RIR

l’ombre prévue qu’il avait marquée d’un lavis pour y ajouter un léger rehaut blanc. Il pose enfin une lumière sur la pupille. Cette démarche « de donner vie » par de petites touches de matière est vraie pour tous les personnages. Comparons maintenant le dessin sous-jacent d’Adam et Eve à celui de deux peintures autographes quasi contemporaines : le Mariage des Arnolfini (1434) et la Vierge au Chanoine Van der Paele (1436). Dans ces œuvres, les modifications dans le dessin de mise en place sont particulièrement nombreuses et portent notamment, comme dans l’Agneau Mystique, sur les visages, les mains et les pieds des personnages. Les plus remarquables se relèvent dans la figure d’Arnolfini (figs 8.7a-c). Je renvoie pour leur étude détaillée à Lorne Campbell10 pour ne commenter

que le détail de la main droite. La morphologie de la première main dessinée, la paume ouverte, avait été précisée lors d’une deuxième phase de dessin par des traits noirs, comparables à ceux relevés dans l’Agneau Mystique, mais cette seconde ébauche a été abandonnée lors de l’exécution picturale au profit d’une main finale décalée vers la droite et peinte sans l’aide d’un nouveau dessin préalable (fig. 8.7a). Ceci montre le processus eyckien continu de corrections et ses différentes séquences de visualisation des formes dans l’espace. Dans la Vierge au Chanoine Van der Paele, des changements de composition importants pour resserrer les liens entre Marie et son Fils ont également été apportés à la position de leurs mains.11 Toutes ces modifications, comme dans le retable de Gand, s’opèrent à différents moments : au stade

le rôle du dessin sous-jacent et de l ’ ébauche préparatoire au lavis

b

a

c

127

Figs 8.7a-c Jan van Eyck, Mariage des Arnolfini, (a) détail de la main, (b) des pieds et (c) du visage de Giovanni Arnolfini, RIR, Londres, National Gallery

du dessin initial avec une première affirmation des contours retenus par une ligne noire, ceux-ci pouvant encore être modifiés lors d’une première phase d’exécution picturale, suivie éventuellement d’une nouvelle reprise de la forme peinte dans la version finale. Cette démarche est illustrée autant par les pieds d’Arnolfini (fig. 8.7b) que par les sabots des chevaux sur le panneau des Chevaliers du Christ.12 Les visages sont aussi soumis à repentir (fig. 8.7c), en particulier les yeux, pour trouver la juste direction du regard afin de rapprocher psychologiquement les personnages et pour individualiser les expressions. Dans cette optique, la bouche est souvent retravaillée. Les hachures de modelé sont similaires à celles d’Adam et Eve dans leur tracé et leur distribution : longues et parallèles, serrées en bordure des formes et plus espacées dans les ombres moyennes, superposées ou entrecroisées pour indiquer les zones d’ombre intenses souvent renforcées par un lavis. Le dessin des drapés (fig. 8.8) de l’Agneau Mystique, comme ceux des tableaux de Londres et de Bruges, préfigure clairement les volumes en changeant les plans de hachures de direction en fonction des lumières et en y apposant un lavis. De larges lignes diluées soulignent de la même manière le creux de plusieurs plis, constituant une première ébauche de mise en place que Van Eyck, s’il la maintient, confirme par le trait

128

catheline périer-d ’ ieteren

Fig. 8.8 Jan van Eyck, Vierge au chanoine Van der Paele, détail du drapé de la Vierge, RIR, Bruges, Groeningemuseum

le rôle du dessin sous-jacent et de l ’ ébauche préparatoire au lavis

foncé qui lui est propre. Il travaille de la même façon pour les contours des visages et des linéaments. Celui du Chanoine Van der Paele en offre un bel exemple. Enfin, on notera deux particularités qui, à notre connaissance, n’avaient pas encore été relevées dans cette œuvre. Premièrement, l’emploi d’un report mécanique pour le décor floral du dais en brocart derrière la Vierge. Les grenades et les fleurs se répètent en effet à l’identique sur l’ensemble du tissu, comme le prouve le montage photographique qui a permis de superposer ces deux motifs du dessin sous jacent reconstitués graphiquement à tous les motifs peints (figs 8.9a-b). Seul un très léger décalage s’observe entre les contours au lavis du dessin sous-jacent, qui reprennent et fixent les formes initiales du dessin de mise en place reportées probablement au poncif, et les tracés noirs définitifs de l’exécution picturale posés sur la surface texturée du tissu. L’examen des RIR révèle, en effet, des points dessinant des motifs là où le dessin sous-jacent n’a pas suivi le report des formes. Ces points ne doivent pas être confondus avec les petits traits et/ou points en surface résultant d’une interruption de tracé des motifs posés au stade peint sur les lignes obliques en léger relief qui imitent les fibres du tissu. Le trait au pinceau marque le dos de la trame et s’interrompt dans le creux. Le décor semble avoir été mis en place une fois la Madone dessinée et peinte comme le révèle le dessin des pétales de fleurs se superposant par endroit à la chevelure de Marie. Un autre procédé de report mécanique, le pochoir, semble avoir été utilisé pour les motifs décoratifs du brocart de saint Donatien reproduits de façon similaire sur l’ensemble du vêtement. Les ornements des orfrois, par contre, paraissent avoir été réalisés au poncif, quelques points marquant les contours des petits édifices abritant les figurines étant encore perceptibles (fig. 8.10). Ces observations incitent à penser que la réalisation des ornements de tissus était confiée à un assistant, au moins pour la phase de report, les quelques modifications relevées dans les tracés noirs finaux qui simplifient le plus souvent la forme dessinée, pou-

129

vant être de la main de Van Eyck ? Une restauration importante ayant été menée par Van der Veken en 1933-1934,13 il faudrait être certain, pour pouvoir confirmer les hypothèses émises sur les phases d’exécution et leur auteur, qu’il n’ait pas repris ou renforcé lui-même les tracés noirs sur la surface picturale.14 Seul un examen au microscope binoculaire du tableau pourrait apporter la réponse. L’examen de l’exécution peinte du triptyque de Dresde, malgré son petit format qui empêche l’utilisation d’un poncif, a révélé une démarche comparable : répétition, probablement au pochoir, d’un même type de motif mis sur un tissu texturé, les fils d’or étant peints en diagonale et s’arrêtant aux formes, et surtout une même écriture pour les tracés noirs de surface. On note aussi sur le dais une alternance identique de parties foncées et claires, la source de lumière étant située à gauche, et une étroite parenté dans la gamme chromatique à dominante de vert enrichie de couleurs or, rouge, blanche et d’aplats bleu.15 La deuxième particularité est le caractère étonnement dynamique du dessin de modelé dans les vêtements (fig. 8.8). Les hachures sont exécutées rapidement, parfois en traits continus sans lever la main. Elles sont posées un peu dans le désordre avec des réseaux de traits qui se croisent et même qui enjambent par endroits le dos des plis. Ce type d’écriture, qui s’observe aussi dans le visage de saint Georges, est inattendu et en contradiction apparente avec les compositions picturales équilibrées et immobiles de Van Eyck. Il nous reste à rapprocher les dessins sous-jacents des autres panneaux de l’Agneau Mystique, de celui des dessins relevés dans les peintures autographes. Nous nous concentrerons en particulier sur les volets des Anges Chanteurs et Musiciens. Le soin apporté au dessin comme à l’exécution picturale des Anges Chanteurs est plus grand et le tracé de mise en place des linéaments des visages aussi plus sûr que celui de leurs homologues musiciens

130

catheline périer-d ’ ieteren

Figs 8.9a-b Vierge au chanoine Van der Paele,  détail du dessin au poncif dans le dais de brocart et superposition graphique d’un motif aux autres, lumière ordinaire et RIR

le rôle du dessin sous-jacent et de l ’ ébauche préparatoire au lavis

131

Figs 8.10a-b Vierge au chanoine Van der Paele, détail du brocart de saint Donatien, orfroi avec dessin au poncif, lumière ordinaire et RIR

(figs 8.11a-b). Ces derniers, d’une exécution nettement plus faible, ne dénotent pas le même souci du détail dans le rendu des cheveux et dans la préparation du modelé. Le dessin ne présente que de rares reprises hésitantes au trait noir à imputer aux assistants qui ont adopté la manière du maître mais sans la dominer. Seul le visage de l’ange chantant derrière le lutrin (fig. 8.11a) offre, selon nous, l’ensemble des caractéristiques d’écriture de Van Eyck et une démarche créative qui lui est propre, comparable notamment à celle de la Vierge au Chanoine Van der Paele (fig. 8.2c). On reconnaît les longues hachures au pinceau serrées et à la pointe effilée mises par plans dans l’ombre de la tempe et de la joue, les corrections dans la position et le dessin des yeux, des narines et de la bouche et les traits vigoureux noirs soulignant l’essentiel des informations pour le rendu du menton, des paupières et des boucles enlevées des cheveux. Autant de particularités qui ne s’observent pas dans le visage de l’ange à l’extrême droite. Les trois anges placés à sa droite évoquent partiellement sa manière dans les modifications apportées à la forme des yeux et des lèvres qui leur confère une expression autre que celle prévue initialement. L’exécution des quatre anges restants est plus relâchée.

Ces dessins pourraient, nous semble-t-il, avoir été réalisés par Van Eyck comme exemple à l’attention de ses assistants en charge des figures qu’il ne peint pas lui même et pour lesquelles il se serait contenté d’apporter des corrections aux tracés initiaux. Il confirme en particulier la position des doigts de la main de l’ange au clavecin (fig. 8.12b) après plusieurs essais dessinés qui seront encore modifiés dans la phase picturale pour se rapprocher le plus possible de la réalité. Dans le reste de la composition ses assistants empruntent sa technique de reprise de formes mais les lignes lourdes et incertaines trahissent clairement leur intervention sans doute sous la tutelle du maître (figs 8.12a, 8.12c). Ce dernier ne néglige aucun détail et n’hésite pas à reprendre le dessin autant dans les parties qu’il exécute que dans celles confiées à ses collaborateurs. L’examen du dessin sous-jacent des panneaux de Dieu le Père et de la Vierge n’apporte aucune donnée nouvelle, mais confirme certaines parentés d’écriture avec les peintures autographes, notamment, le dessin du drapé de Marie, comparable à celui de la Vierge au chanoine Van der Paele. Le dessin de saint Jean, quant à lui, peut être rapproché de celui de l’ange chanteur autographe dans

132

catheline périer-d ’ ieteren

Figs 8.11a-b Retable de l’Agneau Mystique, détails, (a) ange chanteur ; (b) ange musicien, RIR

l’exécution très libre des mèches de cheveux et dans les plis tracés en lavis de sa robe nouée à la taille. Dans la partie inférieure, l’Adoration de l’Agneau, les Chevaliers du Christ, les Ermites et les Pèlerins montrent un dessin très variable dans sa qualité d’exécution et dans le soin accordé aux détails. Dans le groupe représentant la hiérarchie

de l’Eglise, seuls les deux premiers religieux, l’évêque Lievin, saint patron de Gand, et le diacre saint Etienne, présentent un dessin fouillé de type eyckien. Dans les autres figures, les traits de contours et les lignes noires de reprise des formes, comme la pose du réseau de hachures d’ombre sont moins assurés que dans les parties autographes.

Figs 8.12a-c Retable de l’Agneau Mystique, (a-b) détails des mains des anges musiciens ; (c) main d’un évêque, RIR

le rôle du dessin sous-jacent et de l ’ ébauche préparatoire au lavis

En ce qui concerne le revers du retable, nous ne reviendrons pas sur le changement de composition radical du registre supérieur connu de tous où la série d’arcs gothiques trilobés a été remplacée par une charpente. Cette modification aurait été apportée par Jan van Eyck à la mise en page originale conçue par son frère Hubert.16 En situant l’Annonciation dans un intérieur bourgeois, Jan innovait comme il l’avait fait sur les faces du polyptique avec les figures d’Adam et Eve. Les visages de Joos Vijd et de son épouse sont soumis au même type de modification « d’expression » des yeux et des lèvres que celles notifiées pour Adam et Eve et les œuvres autographes précitées. On y retrouve des lignes de contour similaires au lavis, de minces et brèves hachures pour marquer les petites plages d’ombres sur le front et les tempes et les fameux traits noirs énergiques de confirmation de mise en place des formes autour des yeux du donateur (fig. 8.6b). Le dessin des deux saints Jean est peu affirmé et sans reprise. Il s’écarte, selon nous, de celui de Van Eyck au même titre que l’exécution picturale et le style très lourd et raide des drapés. Dans l’Annonciation, la comparaison du dessin sous-jacent de l’Archange et de la Vierge révèle un dessin fort différent qui trahirait la participation de deux mains. Celui de l’archange n’est élaboré ni dans le visage, ni dans les vêtements. Il se réduit à quelques plans de hachures dans le drapé et à de simples lignes de mise en place des plis qui campent les formes sans que Van Eyck y appose les très noirs de correction si caractéristiques à sa manière. Les mains, entre autres, s’avèrent être maladroites autant dans le dessin que dans leur exécution picturale. Le dessin préparatoire de la Vierge, par contre, est très étudié. On reconnaît la main du maître dans la reprise du contour gauche du visage ainsi que dans la belle organisation des plis de la robe. Ceux-ci sont campés par de larges lignes en lavis, plus ou moins dilué, qui marquent le creux des plis dont le dessin encore est confirmé ou corrigé par un trait noir libre et énergique. Des séries

133

de hachures légères viennent ensuite moduler les ombres aussi préparées par des plages en lavis. Le volume du livre d’Heures est adroitement situé dans l’espace par quelques traits noirs assurés et le texte des pages entrouvertes est simulé par de fines hachures parallèles. Une autre différence flagrante de dessin trahissant l’intervention d’un aide s’observe dans les lunettes occupées par les prophètes et les sibylles. Ainsi, le style du dessin de Zacharie s’apparente à première vue à celui de Van Eyck par le traitement fluide des poils de la barbe, les repentirs dans la forme des yeux et les types de hachures parallèles. Toutefois, il est relâché dans l’exécution. La comparaison du rendu du livre au volume moins bien défini que celui de l’Annonciation et l’écriture plus raide des hachures dans les vêtements sont significatives. Le dessin enlevé de Micha, plein d’accents forts et de reprises spontanées, est par contre certainement de la main du maître. On reconnaît la souplesse des hachures dans les zones d’ombre et leur orientation variée en fonction des effets de lumière recherchés, les petites lignes parallèles sur le dos des plis et les lavis dans les creux, les corrections apportées aux doigts, et les reprises de contours en noir. Toutes les caractéristiques de ce dessin superbe évoquent celles de la Vierge au Chanoine Van der Paele. De la même façon, le dessin de la Sibylle d’Erythrée se démarque de celui de la Sibylle de Cumes. Le premier est plus pictural parce que constitué presque essentiellement de lignes au lavis mais la mise en place des plis est hésitante et manque de netteté alors que le dessin du second est linéaire et plus ferme. L’aspect un peu lourd et brouillon du dessin d’Erythrée, dû à l’irrégularité des tracés, ne nous semble pas correspondre au style assuré de Van Eyck tout comme celui, assez pauvre, de la Sibylle de Cumes. Les corrections apportées aux yeux et aux mains des deux prophétesses révèlent par contre une attention analogue portée au rendu des visages, dans la ligne des préoccupations de Van Eyck. Nous verrions donc dans le dessin de ces lunettes l’intervention de plusieurs assistants.17

134

catheline périer-d ’ ieteren

La problématique des collaborations dans l’exécution du retable a été soulevée dès les années 1953 en se centrant plus particulièrement sur l’intervention de Jan par rapport à celle de son frère Hubert. Elle a resurgi en l’élargissant à l’intervention d’autres mains après la publication de Van Asperen de Boer en 1979, mais sans que de vraies réponses ne soient formulées.18 L’imagerie scientifique actuelle, complémentaire du réexamen de visu des panneaux, a fourni de nouvelles données qui révèlent à l’évidence la collaboration de plusieurs peintres. Ces informations demandent maintenant à être exploitées au niveau de l’ensemble du polyptyque, envisagé en tant que « Gesamtkunstwerk » pour paraphraser Taubert.19 Une meilleure caractérisation des dessins autographes de Van Eyck, telle que nous avons tenté de le faire, permettrait de mieux comprendre la distribution des tâches au sein du retable et d’expliquer les différences manifestes qui s’observent dans le dessin sous-jacent autant que dans l’exécution picturale dont l’étude devrait être approfondie. À côté de changements de composition spectaculaires, comme celui du revers du retable, Van Eyck exploite un éventail complexe de moyens pour aboutir à l’état final « parfait » propre à ses peintures. Ces interventions visent toutes à cerner la réalité au plus près : la vérité anatomique, celle des expressions individuelles ou des rapports affectifs entre les personnages, la véracité des paysages et du rendu de toutes les textures de surface jouant à la lumière. Le maître apporte des corrections constantes à ses propres tracés avant de les arrêter au dernier stade du dessin sous-jacent par une ligne noire énergique et souvent, avant d’encore reprendre la forme ou la position des éléments à différents stades de l’exécution picturale, peaufinant la composition en ajoutant ci et là des détails infimes que seul un examen rapproché permet de discerner. La qualité de ses modelés se prépare dès le dessin sous-jacent par la variété des hachures appliquées essentiellement au pinceau et par leur adroite distribution. L’utilisation quasi généralisée de lavis lui permet d’obtenir des effets de densité qui

contribuent au rendu des volumes et aux transitions imperceptibles de l’ombre à la lumière.20 Van Eyck nous apparaît enfin sous un jour nouveau comme un chef d’atelier réalisant des dessins à titre d’exemple et reprenant les tracés de ses assistants comme Rubens le fera plus tard.21 Son dessin toujours décrit en termes de « subtilité » et de « perfection arrêtée » revêt un caractère inattendu, passé le plus souvent inaperçu par les accents linéaires forts qu’il y apporte et parfois par le dynamisme de sa facture dans la pose des hachures. Il ne s’abstient pas de faire recourir ou de recourir luimême à des procédés de reproduction mécanique (poncif et/ou pochoir) fournissant, avec la Vierge au Chanoine Van der Paele, peut-être un des exemples les plus précoces de ces pratiques dans la peinture flamande du XVe siècle.22 Sa technique de dessin au lavis, si spécifique, fera aussi des émules, le procédé étant repris par Rogier van der Weyden, Dirk Bouts et Gérard David pour ne citer que les peintres les plus représentatifs qui lui succèdent.23 En réalisant cette étude, nous nous sommes rendu compte après avoir écrit en 1985 : « le dessin de Van Eyck est d’une exécution si fouillée qu’il préfigure l’état final de la peinture » que cette assertion, bien qu’exacte, s’avère aujourd’hui réductrice à la lumière de ces nouvelles observations. En effet, elle est loin de couvrir toute l’étendue des recherches et actions mises en œuvre par le maître pour que sa peinture réponde à sa conception esthétique et traduise simultanément sa foi en Dieu et en l’homme.24 NOTES * Nous voudrions exprimer nos remerciements à Christina Ceulemans et Christina Currie (KIK-IRPA), Rachel Billinge (National Gallery de Londres) et à Till Holger Borchert (Groeningemuseum de Bruges) pour nous avoir autorisée à utiliser les photographies et RIR de l’Agneau Mystique, du Mariage des Arnolfini et de la Vierge au Chanoine Van der Paele, à Sophie Potter (KIK-IRPA) et Nathalie Bloch (ULB) pour l’aide précieuse qu’elles ont apporté à la réalisation du Power Point, à nos collègues : Mélanie Gifford, Rachel Billinge et Lorne Campbell. Une reconnaissance particulière va à Valentine Henderiks et Françoise Rosier pour les échanges d’idées sur le dessin sous-jacent de Van Eyck et pour leur lecture critique du texte qui a permis d’approfondir la réflexion sur la démarche du maître et de revoir la formulation écrite des idées nouvelles. Merci aussi à Sacha Zdanov pour l’aide apportée à l’élaboration des notes.

le rôle du dessin sous-jacent et de l ’ ébauche préparatoire au lavis 1 Coremans 1953. 2 De nouveaux documents en haute définition ont été réalisés par l’IRPA pour le site internet consacré à l’Agneau Mystique et sont disponibles à l’adresse suivante : http://closertovaneyck.kikirpa.be. 3 Taubert 1975, p. 41-72. 4 Van Asperen de Boer 1979, p. 141-214. Voir également les études antérieures de Davies 1954, p. 119-135; Desneux 1958, p. 13-21 ; Comblen-Sonkes 1970, p. 195-223. 5 Voir notamment Janssens de Bisthoven 1983 ; Périer-D’Ieteren 1985 ; Hand, Wolff 1986, p. 75-86 ; Van Asperen de Boer, Faries 1990, p. 37-49 ; Van Asperen de Boer, Ridderbos, Zeldenrust 1991, p. 8-35 ; Bosshard 1992, p. 4-11 ; Van Asperen de Boer 1992, p. 9-13 ; Sander 1993, p. 245-263 ; New York 1994 ; Billinge, Campbell 1995, p. 47-60 ; Comblen-Sonkes, Lorentz 1995 ; Verougstraete, Van Schoute 1995 ; Butler 1997, p. 28-46 ; Campbell 1998, p. 174-223 ; Gifford 1999, p. 108-116 ; Verougstraete, Van Schoute 1999 ; Borchert 2000, p. 40-42 ; Foister, Jones, Cool 2000 ; Washington/Antwerp 2006, n°s 8 et 9, p. 70-81 et p. 282 ; Gifford et al. 2013. 6 Jan van Eyck, Mariage des Arnolfini, 1434, Londres, National Gallery, inv. NG 186. 7 Jan van Eyck, Vierge au Chanoine Van der Paele, vers 1434-1436, Bruges, Groeningemuseum, inv. 0000.GRO0161.I. 8 Voir notamment Jan van Eyck, Homme au turban et Léal Souvenir, Londres, National Gallery, inv. nos NG 222 et NG 290 ainsi que le Portrait de Marguerite van Eyck, Bruges, Groeninge Museum, inv. n° 0000.GRO0162.I, où seuls des plans de très légères hachures à peine perceptibles préparent l’ombre comme la démontré R. Billinge dans cette même journée d’étude. Campbell 1998, p. 212-217 et p. 218-223. 9 Périer-D’Ieteren 1985, p. 23 et Périer-D’Ieteren 2005, p. 99. 10 Voir analyse de Campbell 1998, p.182. Voir également sur le Mariage des Arnolfini et sur la Vierge au Chanoine Van der Paele Taubert 1975, p. 47-54. 11 Janssens de Bisthoven 1983, p. 195 et Taubert 1975, p. 51-52. 12 Taubert 1975, p. 44-45. 13 Janssens de Bisthoven 1983, pp. 209-210. 14 Nous adressons nos remerciements les plus chaleureux à Yvo Mohrmann de la Hochschule fûr Bildende Kunste Dresden et à son assistante Kerstin Riße pour l’aide apportée en juin 2013 à l’examen

135

des RIR et pour les simulations graphiques qu’elle a réalisées à cette occasion, à Uta Neidhart, conservatrice à la galerie des maîtres anciens à Dresde et Christoph Schölzel, conservateur de peintures au Département conservation restauration du même musée pour les discussions très enrichissantes en présence du triptyque et enfin à Till-Holgert Borchert, conservateur du Groeningemuseum et Anne Van Oostenwyck pour l’aide apportée sur place à l’examen comparatif entre les documents RIR et la surface picturale de la Vierge au Chanoine Van der Paele. 15 On observe encore un même type de décor de fleurs et de feuilles sur le dais de la Madone de Lucques (Francfort, Städel Museum) ce qui confirmerait l’intervention d’un assistant spécialisé dans l’exécution de ces parties décoratives. Il devait probablement aussi exister des répertoires reprenant les orfrois du genre de ceux portés par saint Donatien ceux-ci se retrouvant à la même époque dans tous les domaines d’expression artistique. 16 Cette modification aurait été apportée par Jan van Eyck à la mise en page originale conçue par son frère Hubert. 17 Pour la comparaison de ces détails d’ecriture du dessin nous renvoyons au site http://closertovaneyck.kikirpa.be. 18 Van Asperen de Boer 1979, p. 141-214. 19 Taubert 1978, p. 11-18 et 30-37. 20 On note chez Van Eyck une nette volonté de marquer les ombres au niveau du dessin sous-jacent, ce qui, à première vue, paraît en contradiction avec le modelé final fluide sans effet de clair-obscur qui lui est propre. 21 Bruxelles 2007. 22 Nous avions émis l’hypothèse, déjà en 1982, de l’usage d’un poncif dans la figure du chanoine Van der Paele pour la mise en place du col du surplis mais aussi dans la chape de saint Donatien : PérierD’Ieteren, 1982-1983, p.75, note 5. Cette observation a été confirmée par Bart Fransen en 2012, p.122. La problématique de l’usage de moyens mécaniques de reproduction dans les peintures de Van Eyck s’avère cependant beaucoup plus complexe. Une étude approfondie menée en collaboration avec plusieurs collègues détenteurs de peintures du maître est en cours et fera l’objet d’un article dans un futur proche. 23 Cette même technique se relève dans des peintures des PaysBas du Nord . Voir Filedt Kok et al. 2011, p. 90-92. 24 Périer-D’Ieteren, 1982-1983, p. 22.

Figs 9.1a-b Ghent Altarpiece

9

Art and Compensation. Joos Vijd and the Programme of the Ghent Altarpiece Bernhard Ridderbos

ABSTRACT: Although the literature on the Ghent Altarpiece is overwhelming, the question of why a wealthy and respectable but rather nondescript person like Joos Vijd ordered the most important altarpiece in Early Netherlandish art has not received the attention it should have had. It is even doubted whether he was the original commissioner of all the panels. In this paper it is argued that the programme of the altarpiece, although complex, is a consistent whole and can be related very well to Joos Vijd and the history of his family. Therefore, both this history and the iconography of the painting are analysed and it is suggested that the form and content of the Ghent Altarpiece compensated for a lack of social self-confidence from which Vijd may have suffered.

—o— The quatrain on the frames of the Ghent Altarpiece has given rise to a good deal of controversy, not only as regards its authenticity but also its content.1 Scholars who consider this text to be authentic are divided over the question of how it should be interpreted. The inscription tells us that Hubert van Eyck began the work and that his brother, Jan van Eyck, completed the ‘weighty task’ at the request of Joos Vijd. According to Panofsky, this means that Vijd was not the commissioner from the start.2 He argued that the lower part of the altarpiece would originally have been intended as a separate painting, destined for the Ghent town council. Joos Vijd would have acquired this painting as well as other

unfinished works by Hubert after the painter’s death in 1426, and would have asked Jan to complete them and to combine them with new images – the Adam and Eve on the interior and the depictions on the exterior – in one altarpiece. Whereas Panofsky’s ideas are being revived by Hugo van der Velden, I will explore the opposite view, which is that Vijd was the commissioner of all the panels and that the work was conceived as a whole from the start, even if it was begun by Hubert and finished by Jan.3 Thus, I will try to answer the following questions: what motives could a respectable and wealthy but rather nondescript person like Vijd have had for ordering this extremely ambitious work of art, and can the subject matter of the altarpiece be related to those motives? To answer these questions, first Joos Vijd’s family background will be discussed, then his social position will be examined, and finally the programme of the painting will be analysed. The Fall of Clais Vijd For many years Joos’s father, Clais Vijd, held a high position in Flanders.4 In 1355 Clais had succeeded his brother Jan as castellan, bailiff and receiver of the Land of Beveren, which meant that he bore full responsibility for the management and administration of one of the largest seigniories of the county.

138

bernhard ridderbos

One of his two daughters was married to Joos Triest, who was bailiff of the Land of Waas, and his eldest son, also named Clais, became Triest’s successor. It would not have been surprising, therefore, had his other sons, Christoffel and Joos, continued the family tradition of official service and made a successful career in the employ of the Duke of Burgundy, Philip the Bold. Things, however, took a different direction. In 1390 Clais senior was found guilty of fraud: he had left out of the accounts revenues that he had collected in the name of Philip the Bold but had appropriated for himself and had also put in fraudulent claims for fictitious expenses. The duke, though declaring that he preferred ‘grace, doulceur et misericorde’ to ‘rigueur de justice’, fined him an enormous amount of money and dismissed him.5 Philip the Bold may have wanted to make an example of Clais as a warning to other, likewise corrupt officials. After succeeding his father-in-law, Louis of Male, as Count of Flanders in 1384, Philip had created the office of Controller of Flanders and this functionary had examined Clais Vijd’s management. The longer the office of the Controller existed, the more difficult it proved to combat misconduct. In the course of the fifteenth century the penalties were reduced and sometimes the duke – by then Philip the Bold’s grandson, Philip the Good – even remitted part of the fine.6 Thus, it looks as if Clais Vijd had simply had bad luck. That did not prevent him from amassing a fortune. He was married to a rich lady from the Land of Waas, Amelberga van der Elst, and held an important fief – the Hof ter Saksen – in the Land of Beveren. After his dismissal he acquired the seigniory of Pamel and Ledeberg, near Ninove; such a possession was largely a privilege of the nobility.7 At the time of his death, in 1412, his income from rural estates, peatlands, houses and annuities was so large that according to the historians Frederik Buylaert and Erik Verroken, who recently carried out extensive research into Clais Vijd and his family, he must have been one of the richest men of Flanders.8

Nevertheless, the disgraceful termination of Vijd’s employment as an official in the service of the Duke of Burgundy in 1390 must have affected his social status and may have influenced the careers of his sons Christoffel and Joos (Clais junior had died in the same year in which his father was sentenced9). Christoffel, who, like his eldest brother, had been knighted, was captain of a militia in the Land of Waas. Only after his father’s death and more than twenty-five years after the latter’s dramatic dismissal was he appointed to offices similar to those held by Clais senior and Clais junior: from 1414 to 1416 he was bailiff – and for a while also castellan – of the Land of Beveren. His death before October 1418 presumably ended this short career as an official.10 The third son, Joos, never entered ducal service but lived as a patrician in Ghent, where he married Elisabeth Borluut, a descendant of one of the oldest and most prestigious families of the town.11 The Social Position of Joos Vijd Like his father-in-law, Gerem Borluut, but less frequently, Joos Vijd fulfilled the year-long office of alderman of the city of Ghent more than once. The first time was in 1395, when he occupied the lowest patrician seat on the Gedeele, the lower of the two aldermanic benches that formed Ghent’s municipal administration. Either he made little impression or he was insufficiently interested in the affairs of the city council, for although an alderman of the Gedeele could be re-elected after one year, Vijd did not sit on the bench again until 1415, when he got the same low seat.12 Another long period elapsed before he was elected for the third time, in 1425, and now he was given the middle patrician seat on the upper bench, the Keure. As soon as he was installed he took part in Philip the Good’s campaign against Jacqueline of Bavaria, Countess of Holland and Hainaut, who had been imprisoned in Ghent but had made her escape. At the duke’s request, three Ghent aldermen – of whom Joos Vijd, as the patricians’ representative, was one – and the town pensionary joined his retinue and were entrusted with the task of negotiating with Jacqueline. Instead of

art and compensation

negotiations only clashes between the two forces took place; in the meantime the four delegates stayed for nearly six months in Utrecht. In 1430 Joos Vijd, who as an alderman of the Keure could be re-elected after two years, again received the middle seat, but in 1433 he was invested with the highest rank within the town council: the first seat of the Keure. This happened under specific political circumstances. The year before, a serious riot had broken out in Ghent due to Philip the Good’s plan to introduce a new monetary system, with an unfavourable rate of exchange for the old money as a veiled tax.13 Popular fury was directed primarily against the city’s dignitaries, who had approved the ducal decree. Feelings cooled down after the election of the new aldermen, among whom were sympathizers of the rebels, and a delegation was sent to Philip to ask his pardon. However, tensions did not abate and a number of ex-governors were banished from Ghent. At the beginning of 1433 the duke, realizing that he had to tread carefully, granted privileges, and in the spring he paid a visit to the town. Around Ascension Day the people’s dissatisfaction flared up again and finally, in the autumn of the same year, Philip moderated the hated measure of the unfair exchange rate considerably. Vijd’s appointment as first alderman in the summer of that year, when the situation in Ghent was still very unstable, suggests that he was regarded more as the decent, neutral person that was needed at the time than as an outspoken representative of the establishment who could easily arouse negative emotions among the lower classes. This is in line with the rather inconspicuous role he played in the public life of Ghent during a great part of his life, given the long intervals between his first, second and third elections. The slow development of Joos Vijd’s career as a member of Ghent’s town council may have been due to the fact that he was more involved in the countryside than the city. Already during his father’s lifetime he had received an extensive fief from him, but he was also active himself in acquiring land and bringing it into cultivation: together

139

with others, among whom was his brother Christoffel, he bought salt marshes from John the Fearless and invested in their diking. After Christoffel’s death, Joos’s position as a large landowner considerably increased: neither of his brothers had legitimate children, and thus he was the main heir of the enormous family property. Unfortunately, his marriage had remained childless; in 1433, therefore, he had his possessions registered for the benefit of his heirs – his nephew Joos Triest, son of his sister Elisabeth, and his niece, Goedele Raes, daughter of his sister Mabelie.14 To conclude: the humiliation of his father, his respectable but not very remarkable position in Ghent, his involvement in the countryside without possessing the status of a high-ranking official, his ownership of a seigniory even though not of noble descent, the absence of direct descendants, all these factors together may have caused Joos Vijd to suffer from uncertainty about his social identity. In order to see whether such an uncertainty can be related to his commission of the Ghent Altarpiece we have to turn to the painting itself. The Programme of the Ghent Altarpiece The lower zone of the exterior of the Ghent Altarpiece (fig. 9.1a) shows Joos Vijd and Elisabeth Borluut in prayer beside pseudo-sculptures (grisailles) of John the Baptist, the patron saint of the Church of St John (which would later become St Bavo’s Cathedral), and John the Evangelist, whose particular significance for the donors appears from the fact that 6 May – the date mentioned in the quatrain – was one of his feast days. The scene of the Annunciation above, accompanied in the lunettes by prophets and sibyls who foretell the coming of Christ, proclaims the beginning of salvation. The fulfilment of the history of salvation is revealed when the wings are opened (fig. 9.1b). The interior has often been interpreted as an All Saints picture, but Volker Herzner has convincingly argued that while motifs have been derived from the iconography of All Saints, a different theme is represented.15 The feast of All Saints is devoted to the saints in

140

bernhard ridderbos

heaven, who intercede with God on behalf of mankind as long as the end of time has not yet come. The Ghent Altarpiece, however, shows the final redemption from the Fall – of which the figures of Adam and Eve are reminders – when the Last Judgement has taken place and the first heaven and the first earth are gone. Then, according to the last two chapters of the Revelation to John, which was ascribed to John the Evangelist, there shall be a new heaven and a new earth; the new Jerusalem shall come down out of heaven from God; and the Lord God Almighty and the Lamb shall be the temple of the new Jerusalem (Rev 21: 1, 2, 22). The inscription in the Adoration of the Lamb, on the fountain, hic est fons aque vite procedens de sede dei + agni (This is the fountain of the water of life proceeding from the throne of God and of the Lamb), is derived from these chapters, combining two passages. In Revelation 21: 6, God says: ‘To him that thirsteth, I will give of the fountain of the water of life, freely’, and chapter 22:1 speaks of ‘a river of water of life, clear as crystal, proceeding from the throne of God and of the Lamb’. God, the Lamb and the new Jerusalem have each been given their own place in the Ghent Altarpiece. In the upper register, God – described in Revelation as ‘he that sat on the throne’ (Rev 21: 5) and accordingly depicted here without characteristics which define him as either the Father, the Son or the Trinity – is praised in heaven by singing and playing angels. In the lower register, the Lamb, placed on an altar, is adored by throngs of saints in a flowering meadow symbolizing the new earth. In the background the new Jerusalem is visible. Since God and the Lamb are represented in separate panels, the motif of the river proceeding from their throne, which has a long iconographical tradition, had to be omitted. As technical research has demonstrated, the fountain was painted at a late stage of the execution of the Adoration of the Lamb.16 Here I must argue against Hugo van der Velden’s hypothesis that the fountain was added in honour of the baptism of Josse of Burgundy in Ghent on 6 May 1432, the day on

which the Ghent Altarpiece was presented to the public, according to the quatrain.17 In Van der Velden’s view, only the lower register was finished at that time and the quatrain was inscribed on a canvas below the painting, which was expanded in the following years with the upper register. After the completion of the whole work, the quatrain – despite the date no longer being applicable – would have been transferred to the frames of the exterior. However, the idea that the lower zone was originally an autonomous altarpiece is implausible. God and the Lamb are mentioned together in the relevant chapters of Revelation; therefore, Herzner is right in observing that they are part of one and the same iconographical programme and must have been conceived at the same time. This is confirmed by the inscription on the fountain, referring to both God and the Lamb, and accords with the results of recent dendrochronological analysis, which has identified planks from the same tree in the panels used for God and the Adoration of the Lamb.18 As for a possible relationship between the fountain and the baptism of the newborn prince: although Van der Velden is convinced that this event took place in the Church of St John there is no documentary information about the location at all nor is there any reason to assume that St John’s was chosen as the place where Josse was held at the font.19 True, three generations later Charles V was baptized in this church, but then the political situation was different and the festivity was a manifestation in which the whole city participated.20 Of course, the birth of Josse was celebrated in Ghent but little is known about specific festivities in honour of his baptism, which was attended neither by Philip the Good, since he was staying in Dijon at the time, nor Isabella of Portugal, who was still confined to her rooms in accordance with the rules of etiquette.21 The baptisms of Josse’s brothers, Anthony and Charles (the later Charles the Bold), demonstrate that Philip and Isabella preferred the surroundings of the ducal palace for that ceremony. Anthony was born in Brussels and baptized in the Church of St James on the Coudenberg; Charles was born in

art and compensation

Dijon and baptized in the Sainte-Chapelle. Most probably, Josse was baptized in St Veerle’s, the court church of the counts of Flanders, situated in the immediate vicinity of the ducal residence, the Hof Ten Walle, where he was born. In this church he was buried three months later. Another possibility is the court chapel of Ten Walle. Instead of linking the fountain to a contemporary occasion it should be regarded as an alternative to the motif of the river of the water of life and as a means of coupling the liturgy celebrated in the Adoration with the liturgy in the Vijd Chapel. The liturgical character of the represented eschatological reality is expressed by the Lamb standing on an altar, while blood pours from its breast into a chalice. Because the fountain was not planned from the start, the question arises of why the altar with the Lamb, as the focus of all the surrounding groups, was positioned high in the composition, leaving an empty space between the two crowds in the foreground. The answer could be that this arrangement took account of the presence of a priest in front of the altarpiece during the Eucharist. Filling the space between the two groups he would have seemed to join in the celestial liturgy and have emphasized its foretaste in the terrestrial liturgy. This idea may have been abandoned as the execution of the altarpiece proceeded, since apart from such momentary situations the composition of the Adoration would have looked somewhat unbalanced. Thus, the fountain was added and a permanent connection between the celestial and the terrestrial liturgy was realized by providing the basin with an opening through which a small stream of living water flows towards the altar in the Vijd Chapel. Besides representing the new heaven and the new earth at the end of time as a visual accompaniment to the Sacrifice of the Mass in the donors’ chapel, the programme of the interior concerns, as Elisabeth Dhanens has emphasized, the church and parish of St John. This is not only because St John the Baptist has an honoured place in the upper zone but also because the Lamb is his attribute.22 Whereas the saint and his attribute are divided

141

over the two registers, their relationship is indicated both by the saint’s pointing gesture, which is directed to God but refers to the Lamb, and by the words Ecce Agnus Dei qui tollis peccata mundi inscribed on the antependium of the altar on which the Lamb is standing. These words are appropriate within the context of the depicted liturgical celebration, but they also allude to the Baptist who spoke them. It can be concluded that the programme of the interior is a complex but consistent iconographical whole: themes, motifs and inscriptions related to John the Evangelist as the author of the Book of Revelation and to John the Baptist as the patron saint of the Church of St John are combined with the theme of the Eucharist. At the same time, there is an inextricable link between the exterior and the interior, since the two Johns, under whose protection Joos Vijd and Elisabeth Borluut place themselves, relate the donors to the programme of the inside. Another connection between the donors on the exterior and the programme of the interior may have been realized by means of the groups on the four panels flanking the scene of the Adoration, at least if we follow the valuable observation made by Dhanens that all these groups refer to the Vijd family: the Just Judges to Joos as an alderman of Ghent, because the aldermen administered justice; the Knights of Christ to Clais junior and Christoffel, who had both been knighted; the Hermits to Clais senior, who was a benefactor of the Charterhouse of Rooigem near Ghent and was buried there; the Pilgrims to Christoffel because St Christopher heads this group.23 Through these representatives the Vijd family is allowed to join the prophets, apostles, martyrs, virgins and confessors in the adoration of the Lamb. The Ghent Altarpiece offered Joos Vijd and Elisabeth Borluut the opportunity to manifest themselves as members of various communities: by paying homage to St John the Baptist and his attribute, the Lamb, they situated themselves within the context of the parish of St John; by

142

bernhard ridderbos

appealing to the connection between the celestial and the terrestrial liturgy, they situated themselves within the Church as the mystical body of Christ;24 by testifying to their hopes of eternal salvation, they situated themselves within the company of all the saints and blessed souls who will be united with Christ at the end of time. The wish to present themselves as belonging to these communities with the aid of an artwork of an unprecedented size and inimitable evocative power can only partly be explained by the piety and wealth of the donors. Another strong motive is found in Joos Vijd’s urge to compensate for a lack of social identity caused by his family history, his own inconspicuous career and the absence of direct descendants. NOTES * With special thanks to Hans Bloemsma for correcting and improving my text, and to Frederik Buylaert and Erik Verroken for generously sharing their important article with me before publication. 1 For recent discussions of the quatrain, see Van der Velden 2011a; Herzner 2011c; Van der Velden 2011b; Herzner 2013-2014; Kemperdick 2014, pp. 19-28; Meckelnborg 2014. 2 Panofsky 1953, pp. 205-230. 3 Van der Velden 2011b, pp. 140-141. An earlier, much different version of my text was published in Ridderbos 2008. I have also expressed my views on the Ghent Altarpiece in Ridderbos 2014, pp. 53-70. 4 On Clais Vijd and his family, see especially Buylaert, Verroken, forthcoming; on his dismissal, see also Van Rompaey 1967, pp. 420-426. See further: Fris 1907, pp. 84-89; De Maesschalck 1919-1920; Van Herreweghen 1964; pp. 172-177; Goodgal 1981, pp. 103-121; Goodgal 1985; Buylaert, De Clerq, Dumolyn 2011, pp. 714-716. 5 Van Rompaey 1967, pp. 425-426. 6 Van Rompaey 1967, pp. 175-177, 420-470. 7 Buylaert, De Clerq, Dumolyn 2011, pp. 406-414. Buylaert, Verroken, forthcoming, relate the commission of the Ghent Altarpiece to the idea that the Vijd family would have acquired noble status. 8 Buylaert, Verroken, 2013. 9 In January 1390 – a month before his father was found guilty – Clais junior had been transferred to Vier Ambachten in the capacity of bailiff; he died at the end of that year; Buylaert, Verroken 2013. 10 On his functions in the Land of Beveren between 1414 and 1416, see Verelst 1984, pp. 332-333, 340-341. On his death before October 1418, see Buylaert, Verroken, 2013. 11 On Joos Vijd, see Fris 1907, pp. 84-89; De Maesschalck 1919/20; Dhanens 1965, pp. 27-32; 85-95; Goodgal 1981, pp. 103-121; Goodgal 1985; Buylaert, Verroken 2013. 12 On the benches of aldermen in Ghent, see Van Werveke 1947, pp. 55-58; Blockmans 1987; Boone 1990, pp. 33-57; Van Leeuwen 2004, pp. 21-123.

13 On this rebellion, see Van Dixmude 1835, pp. 137-139; Serrure, Blommaert 1839-1840, vol. 1, pp. 33-34; De Smet 1856, pp. 42-43; Monstrelet 1861, pp. 36-38, 50, 68; Fris 1900; Goodgal 1981, pp. 116-120; Goodgal 1985, pp. 30-31. 14 Goodgal 1985, pp. 29, 40-49. See Herzner 1995, pp. 154-155, n. 8, for the correct dating of this registration. 15 Herzner 1995, pp. 30-50. The similarities between All Saints pictures and the Adoration of the Lamb explain why this panel influenced two miniatures, from 1480-1490 and c.1500, illustrating the hours of All Saints, but they do not prove that the Adoration of the Lamb shows all the saints assembled in heaven, as Kemperdick thinks, instead of at the end of time; Kemperdick 2014, p. 16; see also Herzner 1995, p. 44, n. 68. Kemperdick’s argument that the Ghent Altarpiece cannot represent the end of time, since on the Last Day social differences shall be removed and all the resurrected shall be the same age of 33 years is contradicted by his own example of Memling’s Last Judgement, in which the blessed are different ages and are being clothed in accordance with their former functions on earth; Białostocki 1966, p. 67. In Roger van der Weyden’s Last Judgement persons of various ages and ecclesiastical or secular ranks find themselves behind the apostles in heaven; Veronee-Verhaegen 1973, p. 38. 16 Van Asperen de Boer 1979, pp. 194, 201. 17 Van der Velden 2011b, pp. 140-141. 18 Fraiture 2011, pp. 6, 10, 12. 19 Dequeker has already pointed to the fact that there is no evidence for this assumption; he thinks Josse was baptized in the chapel of the Hof Ten Walle; Dequeker 2011, pp. 194-195. 20 On the festivities in honour of the birth and baptism of Charles V, see Pleij 1999, pp. 123-124. 21 On the ceremonies on the occasion of the birth and death of Josse and his brothers, see Sommé 1998. On public festivities in Ghent in honour of Josse’s birth, see Mareel 2010, pp. 21-32. In the invitation to attend Josse’s baptism that Isabella of Portugal sent to Thierry le Roy the location is not mentioned. For the transcription of this invitation, see Devillers 1892, pp. 530-532; this text is also reproduced in Lemaire, Henri, Rouzet 1991, p. 94. Philip the Good left Dijon on 11 May and, after staying at various places, he was in Ghent on 28 May; Vander Linden 1940, p. 100. 22 Dhanens 1965, p. 47. 23 Dhanens 1965, p. 31. Dhanens relates the Knights of Christ to Clais senior instead of Clais junior, whose existence was only recently uncovered and whose knighthood has been erroneously ascribed to his father. On Clais junior, see Buylaert, Verroken 2013. According to Dhanens, also Joos Vijd’s name saint is present in the panel of the Pilgrims, apparently because one of the figures to the left of St Christopher wears a pilgrim’s hat, just like St Josse. However, this figure has a beard, whereas St Josse, when he is depicted as a pilgrim, is a beardless youth. For the iconography of St Josse, see Werner-Tschochner 1974. Moreover, Dhanens connects the Pilgrims to a foundation by Joos Vijd of a hospice in Beveren; on this foundation, which was only effected after the death of Joos Vijd and Elisabeth Borluut by their heirs, see Buylaert, Verroken 2013. 24 Goodgal 1981 gives an extensive interpretation of the interior of the Ghent Altarpiece as ‘the communion of the mystical body with its head, through the sacrament of the eucharist’ (p. 234).

Fig. 10.1 Ghent Altarpiece, John the Baptist Enthroned, detail, book

10

John the Baptist and the Book of Isaiah in the Ghent Altarpiece Patricia Stirnemann, Claudine A. Chavannes-Mazel and Henry Dwarswaard

ABSTRACT: This paper discusses the type of book held by John the Baptist, its text and liturgical importance. Attention is drawn to the excellent examples of pipes and tooling on the gold fore-edges of the books; the tooling is traced back at least to the fourteenth century, in Italy and France. The Old Testament prophets in the left foreground are observed to be holding the four essences gathered at Sukkot, including the etrog, which Van Eyck would have known from his visit to Portugal in 1428-1429. It is noted that on the left, three of the crusaders are quite possibly portrayed as John the Fearless, Maréchal Boucicaut, and Emperor Sigismund, and that the patron saints of Philip the Good’s children are represented in the right hand panels. It is suggested that, overwhelmed by grief at the death of both sons in 1432, the ducal family abandoned the altarpiece.

—o— This paper will discuss the book held by John the Baptist in the Ghent Altarpiece. Various problems were originally posed by Claudine Chavannes, who passed the inquiry on to Henry Dwarswaard for an MA paper and continued to help with the research, and I [Patricia Stirnemann] joined both of them later. The root question was simple. The text on John’s book has long been identified as the opening of chapter 40 in the book of Isaiah, and the meaning of the text within the altarpiece has been well observed. But the type of book and its singularity have drawn less attention. Although there are twenty books represented on the entire altarpiece, this is the only one with a legible text word on any

of the fifteen liturgical books visible when the triptych is open. So there seems to be some meaning behind its display. Below I will lay out the questions, the complementary answers and proposed solutions. What book is John holding? (fig. 10.1) Liturgically, chapter 40 of Isaiah is the reading for the second nocturn at Christmas. In this case, the book would be a breviary, and indeed the frequency of the rubrics that break up the text would seem to confirm the identification. Are the Van Eycks painting an actual manuscript? Can we date and place it? In order to answer this question we used a variety of approaches. We all agreed that the initial in the book was thirteenth-century. The ornamental initial on a coloured, rectangular ground comes into use in the later twelfth century with school books such as Gratian’s Decretum and luxury liturgical books such as bibles. The practice of mounting the initial and opening word on the rectangular ground continues through the first half of the thirteenth century and then disappears, except in England, where it continues into the fourteenth century. Comparisons for the widely looping tendrils inside the decorated letter can be found in Bruges and elsewhere, from about 1240 onwards. So, is the book thirteenth century? When you look closely at it, the penwork initials are late thirteenth or fourteenth century, with the large empty loops and the straight hair-

146

patricia stirnemann, claudine a. chavannes-mazel and henry dwarswaard

lines. What is neither thirteenth- nor fourteenthcentury in John’s initial is the small painted polylobed flower. This rosette is widely found in French penwork during the last third of the thirteenth century, but will only be found as painted border decoration in the fifteenth century. Furthermore, the mosaic background of the field of the initial, where gold, and blue and red paint break up the spaces between the tendrils, is an eleventh- and twelfth-century aesthetic, transposed onto the thirteenth-century initial. The initial, then, is a composite creation, a hybrid. Nonetheless, could a real book have been used as a model? We used two approaches here to answer the question. Henry Dwarswaard made a statistical analysis of the word lengths and found that the word length distribution does not match with a real language.1 We then compared the capitals on the open book with breviaries and found that the sequence of initials did not match the succession of initials, that is to say texts, in any real breviary. Furthermore, the lection Consolamini in all the breviaries consulted (mostly from northern French libraries) began with a penwork initial, not a painted decorated letter. Just as one would not find a decorated letter at the head of chapter 40 in the book of Isaiah in a medieval bible, neither would one expect to find one at the head of this minor text in a breviary. A decorated letter with an adjacent word panel can be found in a breviary, such as Cambrai, MS 93 (first half of the thirteenth century), where the letter, a Beatus, heads the ferial Psalter, a biblical book that is often introduced with a decorated or historiated letter and accompanying word panel. From this accumulated evidence we concluded that the Van Eycks had no specific model for the book. They had certainly examined liturgical manuscripts in search of ideas, but they fused their impressions and chose to mix their models, opting for a thirteenth-century initial and word-panel in order to enhance the legibility of the word from a distance and indicating that the book was probably a breviary by imitating the frequency of the rubrics

on the page. In other words, the Van Eycks were using impressionistic allusions to situate the medieval viewer spontaneously.2 In so doing, they also changed St John’s closed book, which is his usual attribute in western iconography, to an open book.3 Turning to the books as objects, one notes that all the liturgical books on the open triptych have gold punched fore-edges; the Virgin’s book has a green liseuse or wrapper with tassels from which hang the initials M (on the back cover) and A (on the front cover) for Ave Maria and a pearl cluster decorates the pipe (also pippe) or page marker. The pipe is even more pronounced on the Singing Angels’ book. The noted breviaries held by the popes have green bindings and punched gold foreedges, one with a rosette pattern. The gold foreedge of John’s book has a punched diamond pattern, the pipe has a pearl cluster and the leather ties have metal clasps and hanging pearled tassels. The lower tie has been caught up between the folios. Few of these pearled tassels and book-markers still exist.4 Most of our evidence for these objects comes from paintings like this one and from inventories. But it was another archaeological problem that vexed us: when did binders begin to gild and pattern the fore-edges of books with punch marks? When we asked an expert in the field, Marie-Pierre Laffitte, she acknowledged that the subject was largely unexplored and sent us to an article by Albert Derolez on the bindings in the seminary at Ghent.5 Unfortunately all the examples are sixteenth-century. But, using an article by Mme Laffitte on the vocabulary of bindings, we were able to trace the practice back to Charles V in the inventories published by Léopold Delisle.6 The only entry referring to gilt fore-edges relates to a breviary for the use of Rome that the king kept in his chamber, a breviary of the highest royal quality, with gilt edges punched in a lozenge pattern, just like John’s book: Vol. II, n° 120 – bréviaire à l’usage de Romme et sont les fueilletz dudit bréviaire dorez bezancées à lozanges. En la grant chambre du roy a Vincennes.

john the baptist and the book of isaiah in the ghent altarpiece

The references to pippes are equally edifying: 1208 – une pipe d’or à troys pommelles d’or à deux fleurs de lys aux deux boutz 3058 – une pippe … où il a ung petit dyamant et deux perles 3282 – [bréviaire] la pippe d’or à ung ballay et à six perles en ung estuy fort Not only did we find out that gilding goes back at least to the second half of the fourteenth century, we also saw that pearled book-markers were quite in fashion. While most book-binding specialists are rightfully cautious and usually consider gilding a late-fifteenth-century invention, the inventory of Charles V and the untapped visual evidence from Flemish painting resoundingly testifies to a much earlier dating. An attendant question is: how did they do it? And for this, Miriam Foot offers possible evidence:7 The earliest written recipe I have found occurs in a late-fifteenth-century manuscript, a mixture of secrets, chemical, medical and technical tracts and recipes, partly in Latin, partly in Dutch, including one for gilding bindings: ‘Take clean egg white and put it on, then while it is still wet lay it on the gold and let it dry. Then burnish it and work the gold with an engraving tool’ which may refer to gauffering the edges. ‘For old books,’ the recipe continues, ‘add a little saffron and a little bole to the eggwhite.’ (Wellcome Library, MS 517, fol. 219r. The recipe is headed ‘Om boeken te lusten buten mit goude’ – to embellish the outside of books with gold). It would indeed seem that the recipe refers to edgegilding, because the folio edges were the only parts of a bound book that were fully gilded. To return to the original question: why is Consolamini the only legible word on the open altarpiece? Isaiah 40: 1-10 reads:8

147

1 Comfort, comfort my people, says your God. 2 Speak tenderly to Jerusalem, and cry to her that her warfare is ended, that her iniquity is pardoned, that she has received from the Lord’s hand double for all her sins. 3 A voice cries: ‘In the wilderness prepare the way of the Lord make straight in the desert a highway for your God. 4 Every valley shall be lifted up, and every mountain and hill be made low; the uneven ground shall become level, and the rough places a plane. 5 And the glory of the Lord shall be revealed, and all flesh shall see it together, for the mouth of the Lord has spoken.’ 6 A voice says, ‘Cry!’ And I said, ‘What shall I cry?’ All flesh is grass, and all its beauty is like the flower of the field. 7 The grass withers, the flower fades, when the breath of the Lord blows upon it; surely the people is grass. 8 The grass withers, the flower fades; but the word of our God will stand for ever. 9 Get you up to a high mountain O Zion, herald of good tidings; Lift up your voice with strength, O Jerusalem, herald of good tidings, lift it up, fear not; say to the cities of Judah, ‘Behold your God’. This is the second, not the first lesson at Christmas matins, and the third verse is cited in all four Gospels and refers to John the Baptist. Moreover, it is the lesson at Lauds on the Sunday preceding the feast of the birth of John the Baptist on 24 September. These verses in the context of the altarpiece explain, as has been often noted, the pastoral wilderness, the mutability of its grass, flowers and

148

patricia stirnemann, claudine a. chavannes-mazel and henry dwarswaard

people, compared to the durability of the altar of the Lamb and Fountain of Life and of baptism. (fig. 10.2) The last line of verse 9 would explain John’s gesture toward God the Father: ‘Behold your God’. The interaction between the book of Isaiah and John the Baptist continues downward to the altar bearing the Lamb, which is inscribed with the Baptist’s words when he first sees Christ in the Gospel of John (1: 29): ‘Behold the Lamb of God which taketh away the sin of the world’. Isaiah 40:11 reads: He [God] shall feed his flock like a shepherd: he shall gather the lambs with his arm, and carry them in his bosom, and shall gently lead those that are with young. Jesus is indeed the Lamb of God. The relation between Isaiah and the Gospel of John, echoing one another, continues in the inscription around the fountain: ‘this is the fountain of life that flows from the seat of God’, evoking the well-known verse from Isaiah (55: 1) ‘everyone

that thirsteth, come ye to the waters’ and its reflection in John (7: 37) ‘If any man thirst, let him come unto me and drink’ and ‘out of his heart shall flow rivers of living water’. Furthermore, the (repainted) fountain has eight sides, the symbolic number of baptism. In the Gospel, Christ utters the words about ‘those who thirst’ at the end of the feast of Sukkot. This feast – also known as the Feast of Tabernacles or Feast of Booths – is described in Leviticus (23: 40). It falls on the fifteenth day of the seventh month, according to Hebrew calculation, and lasts seven days with an additional day of rest. Sukkot follows Yom Kippur by four days. Yom Kippur is a day devoted to repentance and fasting, the holiest day in the Jewish year. Sukkot is a week of thanksgiving and prayer for all the nations ‘in the wilderness’, when the Hebrews leave behind their man-made homes and possessions and go into the wilderness and live in tents, meditating on the book of Ecclesiastes and recognizing the greatness and durability of God and his creation, the tent of the sky and the work of nature. Sukkot celebrates

Fig. 10.2 Ghent Altarpiece, Adoration of the Lamb

john the baptist and the book of isaiah in the ghent altarpiece

Zechariah’s prophecy (14: 16) of the gathering of all the nations of the world in Jerusalem in the new age to celebrate the Lord. And we find the same invitation in Revelation (21: 6): ‘And he said unto me, It is done. I am the Alpha and Omega, the beginning and the end. I will give unto him that is athirst of the fountain of the water of life.’ At Sukkot, four essences of wood are gathered, including a beautiful tree which is a kind of lemon tree called an etrog, myrtle branches, palm fronds and willow branches, each with a symbolic meaning for Hebrews: Etrog is for the righteous Myrtle is for doers of good deeds Palm is for the learned Willow is for the simple folk

149

In the lower left corner of the lower scene of the altarpiece (fig. 10.3), there are several men facing outward in the foreground: one is bald, two bear branches and another holds a branch with a fruit. These appear to be the great prophets of the Old Testament. The one holding the fruit, which is an etrog, has very long peyots or sidelocks and he wears a crown of myrtle. The man in blue holds a branch of myrtle and the other man behind him holds a branch of willow.9 The man in a red mantle is bald and I would identify him as Isaiah and compare him to the figure of Isaiah in Dijon on Claus Sluter’s Well of Moses. Why has Van Eyck taken such pains to represent the Prophets during Sukkot? Not only does the week of Sukkot take the Old Testament Jews into the wilderness, just as John the Baptist withdrew to

Fig. 10.3 Ghent Altarpiece, Adoration of the Lamb, detail, the four major prophets of the Old Testament

150

patricia stirnemann, claudine a. chavannes-mazel and henry dwarswaard

the wilderness, but in the Jewish calendar Sukkot falls in the seventh month, which is usually the month of September, around the third week; it lasts seven days in Israel and eight days in the diaspora. As the month is determined by a lunar calendar, the dates are moveable. In virtually every Christian year, this week of prayer and feasting in the Hebrew community encompasses the Feast of the Nativity of John the Baptist on 24 September, when, on the Sunday preceding the feast, the lesson at Lauds is Isaiah 40. John, of course, is not only the patron saint of Ghent and the church where the altarpiece was to be housed, he is also the patron saint of Jan Van Eyck, which may help to explain the exceptional interweaving of the prophecies of Isaiah, the Gospel of John, a major Jewish feast and the feast of the Nativity of John the Baptist. It is worth reflecting on where and how Van Eyck became familiar with the feast of Sukkot or where he saw an etrog, a tropical fruit. It is unlikely that he would have seen imported etrogs or Jews building booths in Liège, Ghent or Bruges, since by the beginning of the fifteenth century the Jewish communities in the Southern Netherlands had been reduced to just a few households.10 When Van Eyck went to Portugal to paint the portrait of Isabella, however, he would have met Jews holding highly respected positions within the royal household. Since the thirteenth century, the crown had recognized the Jewish community as a distinct legal entity whose leader, a royal appointment, presided over seven district governors. Local rabbis were also approved and paid by the crown. Jews frequently served as royal treasurer and held other prominent positions.11 It hardly seems likely that Joos Vijd and his wife would have recognized and understood the ceremonial attributes of Sukkot held by the Jews in the altarpiece. Nor is it probable that Van Eyck would have included these details as a remembrance of a personal spiritual experience and discovery while on his diplomatic mission in Portugal. It has been recognized that Isabella is the model for the Erythrean Sibyl on the topmost central panel on the right

when the altarpiece is closed – on the ladies’ side, so to speak, with Eve and the Virgin annunciate.12 The portrait-sibyl would have been painted after Van Eyck’s trip to Portugal, as were almost certainly the great Jewish prophets celebrating Sukkot. Their inclusion would appear to be intended for Isabella’s understanding. Furthermore, Luc Dequeker has pointed out that the date, 6 May, in the quatrain naming the artists is the date of the baptism of Isabella and Philippe’s second son Josse in 1432, and has argued that the altarpiece was made at the behest of the Duke of Burgundy. His argument becomes even stronger when it is noted that on the open altarpiece, St Anthony, patron saint of the duke and duchess’s first son, recognizable by the Tau on his robe, leads the Hermits, and that St Josse, the patron saint of pilgrims and of their second son, follows St Christopher who protects the faithful from sudden death (figs 10.4, 10.5). The altarpiece appears to have been adapted (from a six- to eightsided fountain) to celebrate the baptisms of Anthony and Josse. But the year 1432 was full of tragedy. Anthony, born 30 December 1430, and for whose baptism Gilles Binchois composed the motet Nove cantum (hence the chorus of singers?), died in February of 1432. Josse was born on 24 April in 1432, baptized on 6 May, but died in August. It is surely overwhelming grief that led to the ducal family’s abandonment of the altarpiece. It would appear that the altarpiece evolved rapidly in 1431-1432, with the addition of the Erythrean Sibyl, the eight-sided fountain, the four prophets at Sukkot, the Iberian tiles, perhaps the singers and music-makers, and probably the panels with St Anthony and St Josse. The grievous family events of 1432 and the evolution of the altarpiece throw the word Consolamini on John’s book into yet another and even more immediate light. The word ‘console’ or ‘comfort’ in the imperative plural could have been added after the deaths of the children, not only as a biblical, liturgical reference, but also as an injunction to console the ducal family. There are perhaps other references in the altarpiece to Philip’s or Isabella’s family. One such case

john the baptist and the book of isaiah in the ghent altarpiece

151

Fig. 10.4 Ghent Altarpiece, Hermits, detail, St Anthony

Fig. 10.5 Ghent Altarpiece, Pilgrims, detail, St Josse

may be the crowned figure in the last row of the Knights of Christ (fig. 10.6). This youthful face with its very pronounced nose and slightly hooded eyes evokes an adolescent portrait of John the Fearless (like that in Lille, Musée des Beaux Arts), the only crusader among the Burgundian dukes. In September 1396, the Christian army of King Sigismund of Hungary was defeated at Nicopolis in Bulgaria and John the Fearless (then twenty-five years old) and Maréchal Boucicaut were taken prisoner, to be released only in 1398. Philip the Good dreamt of avenging his father’s humiliation… just as he harboured dreams of attaining kingship. It is tempting to see here a portrait of John with a ducal (or prospective royal) crown, accompanied by Boucicaut in the blue hat (on which a crown has

been painted out). The third figure from the left in the second row, wearing a fur hat, resembles a younger version of the 1433 portrait of Sigismund in the Kunsthistorisches Museum in Vienna, attributed to Pisanello. The identifications and interpretations proposed here stem from the conviction that the altarpiece was easily read by the patrons and that the few truly recognizable saints or worshippers stand out as figures of import to the ducal couple. To the great theophany, Van Eyck added a web of details that spoke directly to the duke and duchess, that referred to specific occasions such as the baptism of each of their two sons, details that evoked easily recognizable and venerated liturgical texts read on the most solemn and significant feast days

152

patricia stirnemann, claudine a. chavannes-mazel and henry dwarswaard

Fig. 10.6 Ghent Altarpiece, Knights of Christ, detail

– Christmas and the feast of the Nativity of John the Baptist – details drawn from the yearly liturgical drama of the Ordo prophetarum performed on Christmas Day. And as special acknowledgement to the duchess, Van Eyck aptly included a reference to the feast of Sukkot which Isabella had witnessed, understood and appreciated so often in Portugal. It is precisely this attention to the particular, the fine and intimate interweaving in placement and time of specific and heterogeneous persons, words, gestures, attributes and liturgical and paraliturgical allusions that never ceases to beckon our meditation.

Appendix. The Words of John’s Book: Frequency Analysis (Henry Dwarswaard) Seen at close range, the text on John’s book seems realistic, yet almost nothing can be read. Mary’s book in the Annunciation has been partially read and identified, but all attempts to read the text of John’s book have failed.13 Was Van Eyck imitating words of a real text? Even if we cannot read the text it is possible to extract information from the page by examining the lengths of the words, and in that way we will answer our question. On the recto page the words are clearly separated, and the lengths of the words can be deter-

153

john the baptist and the book of isaiah in the ghent altarpiece

mined (though in some words a minor inaccuracy results from an unclear letter). On the verso page the words are less clearly separated, but in several sentences word and sentence breaks can be seen. To analyse the length of the words only the number of letters must be found; what each letter actually is does not make a difference. So the first line of the recto page, below the C initial could be: argtai githitrum un gadtin quftrim or:

xxxxxx xxxxxxxxx xx xxxxxx xxxxxxx

which would produce the same result in the frequency counts. Having typed out the best guess for both pages,14 I applied a computerized counting method (as used in Statistical Natural Language Processing 15) to determine the frequencies of the word lengths on the recto page and verso page and compared the result with the same method applied to a real text in Latin, French, English and Dutch. Having found the frequencies for both pages it turned out that the result for the verso page is less reliable because the word breaks are unclear and the right edge of the text block cannot be clearly

seen. The following results are therefore based on the content of the recto page. The result shows a clear difference between the word length distribution on the painted page and in a real text. Because the word ‘CONSOLAMIni’ is Latin, I compare with Latin here, but the results remain valid for other languages. The comparison text is the complete Vulgate bible.16 In a real text there are many short words (e.g. ‘et’) in Latin texts,17 and that is clearly missing in John’s text. We also see that the mean word length and the variation in word length (the standard deviation) are larger than normal. The word length distribution on John’s page looks like a standard random distribution (the blue line in fig. 10.7a). From these observations I conclude that the word length distribution does not match the distribution in a real language; the word lengths painted by Van Eyck cannot constitute a real text. Attempts to read the text will therefore fail, although by exception a single word might be readable. The remark by De Baets that even the language cannot be determined is correct, because no text, in any language, was painted.18

20

20

Consolamini page mean L 6.84 std dev 3.10

%

Vulgate Bible mean L 5.37 std dev 2.76

%

15

15

10

10

5

5

0

0 0

2

4

6

8

10

12

14

16

18

L

20

0

2

4

6

8

10

12

14

16

18

L

Figs 10.7a-b Comparison of word lengths between the text on John’s Consolamini page and the Latin text of the Vulgate

20

154

patricia stirnemann, claudine a. chavannes-mazel and henry dwarswaard

NOTES 1 See Appendix. 2 Van Eyck also alludes impressionistically to a textual environment on the left outside wing where Zechariah has opened the Old Testament to chapter 9 of his book and the letter O of Onus verbi Domini is legible on the verso folio. The citation on the phylactery comes from this chapter (Zech 9: 9). 3 Weis 1974. 4 Laffitte 1989, p. 71 notes that only one pipe exists in the collections of the BnF: n.a.l. 722. For a fourteenth-century representation by the Tuscan painter Bulgarini, see Strehlke and Israëls 2015. 5 Derolez 1984. 6 Laffitte 1989, p. 68; Delisle 1907, vol. I, p. 45, nos 1208, 3058, 3282, 3305, vol. II, no. 120. 7 Foot 2006, pp. 65-71; Braekman 1986, p. 41. 8 Although the English is not as clear, the Latin and original Hebrew state that the people must comfort Jerusalem. See Elliger 1970, pp. 1-3, 4-6, 21. 9 Luc Dequeker has identified these species and their importance at Sukkot in several articles, as well as at this Colloquium, where he identified it as the fruit held by Eve in the altarpiece. See Dequeker 1984 and Dequeker 2011. 10 Schwarzfuchs 1971. 11 Lichterstein 1971; Ranferling 1867. Jean-Pierre Rothschild (Institut de Recherche et d’Histoire des Textes) kindly discussed at

length several aspects of the feast. Luc Dequeker notes (Dequeker 1984, p. 347) that Jan van Eyck met Jewish scholars, among them the botanist Josha de Lorca, in Spain while on a diplomatic mission at the court of Aragon in 1427. 12 Jolly 1987. Although the prophecies of the sibyls and prophets ultimately rely on the Pseudo-Augustinian Sermo contra Iudeos, whose first lines constitute the fourth reading at Matins on Christmas, it is far more likely that they have been adapted from a late version of the Ordo prophetarum, a liturgical play performed on Christmas or New Year’s day. 13 De Baets 1959, col. 105. 14 In fact I did the ‘best guess’ twice. The statistical results were not significantly different. 15 Manning 1999, parts I and II. 16 I used the Clementine Vulgate Bible because it is available in one complete file (on www.documenta catholicaomnia.eu). I also used Genesis, Isaiah and Psalms in the Jerome Vulgate (by the Deutsche Bibelgesellschaft on www.biblija.net), and the results were quite the same. 17 In the Clementine Vulgate Bible, ‘et’ constitutes 8.3% of the words. 18 For the sake of completeness I add that any natural Western language is ruled out. Hebrew, or artificial manipulation of the language image such as in the Voynich manuscript or in Perec’s novel La Disparition, for example, are not applicable to John’s book.

b

c

d

a

e

f

g

h

i

j

k

Fig. 11.1 Le XVe siècle

11

La présentation de l’Agneau Mystique dans la chapelle Vijd. Le rapprochement progressif de deux retables Hélène Verougstraete

ABSTRACT: The Presentation of the Ghent Altarpiece in the Vijd Chapel. The Progressive Bringing Together of Two Retables Six diagrams describe the six stages in the historical presentation of Eyckian altarpieces. At the start, two distinct altarpieces decorated the Vijd chapel. Religious wars, iconoclasm and the French Revolution all wrought much destruction. The subsidence of the upper wings led to further mutilations. Today, the original shape of the upper altarpiece is no longer known. It has been sawn all along the top: the wings in the sixteenth century and the central parts during the French Revolution. The central parts had at least five successive frames. After a troubled history, the physical rejoining of the two altarpieces into one was started in the nineteenth century and completed in the twentieth century. The artist painted the shadow of the frame on the panel in two places, on the wings of Adam and Eve and on the Annunciation. This means that there was a pictorial convention relating to the arrangement of the wings. Radiography reveals that the banderoles in the applied brocade behind the Virgin Enthroned and John the Baptist Enthroned carried a legible text (not an invented one, as has been suggested), incised with an instrument in the leaves after their application to the panel. It also suggests that the same text was repeated on each banderole. The proposed reading of this text supports the idea that Jan van Eyck painted the high altar as homage to his deceased brother. The altarpieces were therefore closely linked to one other. Nonetheless, it is important to keep in mind their duality in order to correctly evaluate their iconography.

—o—

Les informations diverses dont on dispose au sujet de l’histoire des retables m’ont aidée à reconstituer, sous forme de schémas, une évolution de leur présentation au cours des temps dans la chapelle Vijd, et en particulier de montrer quand et comment les deux retables se sont rapprochés pour n’en constituer plus qu’un seul. À la base de ces reconstitutions il y a entr’autres les sources d’archives, à partir desquelles Antoine De Schryver et Roger Marijnissen ont décrit l’histoire mouvementée du chef d’œuvre.1 Elisabeth Dhanens propose un schéma de la chapelle Vijd, dans lequel se superposent deux retables distincts sous l’amorce d’un baldaquin.2 Je me suis appuyée sur ces travaux, mais également sur une relecture de certains passages dans les archives, sur les photographies anciennes et sur l’observation des cadres originaux.3 Le présent article est présenté sous la forme de commentaires des figures 1 à 6 qui reconstituent quelques étapes marquantes dans la présentation des retables. La figure 7 résume les étapes évoquées. Pour rappel, les cadres des volets sont originaux, mais sciés dans l’épaisseur à Berlin en 1894 sauf ceux d’Adam et d’Ève restés à Gand. Pour les parties centrales, il y eut au moins cinq cadres successifs : les cadres originaux, les cadres néo-classiques réalisés à Paris pendant la Révolution, des cadres noir et or au retour à Gand, des cadres dorés en 1920 et de

158

hélène verougstraete

nouveaux cadres dorés en 1950. Etroitement liés à l’histoire des cadres, des changements interviennent dans l’environnement: baldaquin, autel et prédelle, encadrement baroque, rapprochement progressif des deux retables… On trouvera dans les figures de 1 à 6, à gauche : les présentations successives des retables et à droite : les schémas et photographies qui fondent l’argumentation. Résumons ce que nous retenons de l’histoire connue des retables vue sous l’angle des cadres et de la présentation. Au XVIe siècle, les troubles religieux puis les iconoclastes ont bouleversé la présentation initiale et endommagé la peinture de la prédelle. Dès la première moitié du même siècle d’autres dégradations sont jugées préoccupantes. Plus tard les commissaires de la Révolution acheminent vers Paris les quatre panneaux centraux, après avoir travaillé pendant deux jours à leur démontage à Gand. À Paris chaque panneau reçut un nouveau cadre, de style néo (-classique ?).4 En 1816, les quatre panneaux furent restitués. Quelques mois après, avant même qu’on ait suspendu les volets, ceux-ci sont vendus et partent en Allemagne (sauf ceux d’Adam et d’Eve). À Gand, en 1861-1865, l’occasion se présente de reconstituer un ensemble grâce à l’achat des volets de la copie par Michiel Coxcie. Le peintre Victor Lagye complète ces volets par une copie d’Adam et d’Ève vêtus de peaux de bêtes. Après la Première guerre mondiale, les volets originaux reviennent d’Allemagne. L’ensemble des originaux et des copies est exposé à Bruxelles en 1920, puis l’original retourne enfin à Gand. Les cadres dorés des parties centrales, fraîchement confectionnés pour l’exposition, sont récupérés pour la présentation à Gand. Le retable voyagera encore pendant la Seconde guerre mondiale. À partir de 1950, le retable est étudié et restauré. Les cadres centraux sont renouvelés une fois encore et tous les éléments placés dans un cerclage métallique. Les volets, qui avaient été sciés dans l’épaisseur à Berlin en 1894, sont remis dos à dos. En 1986, le retable est déplacé vers la chapelle Villa, toujours dans son cerclage métallique.

À diverses époques, on protégea les retables au moyen de tentures. Les charnières, crochets métalliques et autres systèmes de fermeture (serrures, loquets…) ont pour la plupart été enlevés ou camouflés aux XIXe et XXe siècles. La feuillure qui fermait le retable supérieur (fig. 11.2d) est aujourd’hui comblée par des lattes en bois. Ce gommage des éléments fonctionnels complique l’interprétation de l’articulation primitive. Etait-il nécessaire ? La question mérite d’être débattue. Outre la question de la présentation des retables, nous abordons deux autres points : le premier concerne l’ombre du cadre, peinte en trompe-l’œil sur les panneaux d’Adam et d’Ève, et sur ceux de l’Annonciation, et uniquement en ces endroits ; le second point concerne le texte sur les banderoles du Pressbrokat du drap d’honneur derrière Marie et saint Jean. Très endommagé et retouché, ce texte est difficile à lire aujourd’hui à l’œil nu, mais, gravé dans les feuillets du Pressbrokat, il reste parfaitement lisible à la radiographie, aux rares endroits bien conservés (figs 11.8-11.10). Commentaires des figures 11.1 à 11.7 Figure 11.1. Le XVe siècle Les deux retables Nous avons, en 1989, souligné que les deux registres de l’Agneau Mystique avaient des cadres qui ne pouvaient pas avoir été conçus par les mêmes artisans du bois.5 Les deux retables diffèrent, outre par l’épaisseur des planches et leur mode d’élégissement, par leur construction, le profil des moulures, la largeur des volets, les ferrures, la disposition des charnières, l’articulation, le système de fermeture… Au XVIe siècle, des archives parlent de deux pièces : twee stickx.6 La dualité matérielle des retables n’empêche pas un lien étroit entre eux. Comme le dit le célèbre quatrain, Jan termine le travail qu’Hubert laisse inachevé. C’est en faveur d’Hubert van Eyck, décédé en 14277, que Marie et saint Jean intercèdent auprès de Dieu. Le retable haut est peint après la mort d’Hubert par Jan qui l’a conçu en hommage à son frère défunt. La dualité des retables est importante pour la compréhension de

la présentation de l’agneau mystique dans la chapelle vijd

l’iconographie, considérée jusqu’à présent comme un programme unique dont la cohésion supposée a donné lieu à des controverses intarissables. Le lien entre les deux retables est profondément fraternel et pieux, mais le lien iconographique est peut-être plus limité qu’on ne l’a pensé. Les circonstances d’hommage au frère défunt invitent également à accorder à Jan la paternité du savant quatrain peint sur les cadres du retable bas fermé. Seul Jan a pu affirmer qu’Hubert était le plus grand peintre, tandis que lui-même était « second en art ». Dans toute autre bouche, ces paroles ne manqueraient pas d’étonner. 8 La figure 1 montre, sur l’autel, une prédelle ornée d’un sujet peint, sur laquelle est posé le retable inférieur. Au-dessus, un second retable repose sur des crochets ou équerres métalliques ancrés dans la maçonnerie. Comme de coutume, chaque retable était également attaché au mur par le haut, pour l’empêcher de basculer, par exemple lorsqu’on faisait pivoter les volets. L’espace entre les deux retables était probablement limité, pour faciliter l’accès au retable haut. Le mur étroit de la chapelle Vijd où se trouvaient l’autel et les retables ne permettait pas un déploiement plan des volets. Il était prévu qu’à l’ouverture, Adam et Ève soient positionnés à angles droits avec les anges.9 À la fermeture par contre, l’Annonciation épousait le plan du mur. Le baldaquin Selon Elisabeth Dhanens, un baldaquin sculpté surplombait les retables. Les derniers vestiges de ce baldaquin furent enlevés au XIXe siècle, et les ultimes traces lors de la restauration de 19501951.10 L’auteur publie un tracé d’architecte de la chapelle Vijd où figure le départ d’un voûtain. On peut se demander si la Fontaine de Vie (Madrid, Museo del Prado) (fig. 11.1a) ne doit pas sa forme étrange (avec une partie centrale cintrée fortement exhaussée, sur laquelle est peint un baldaquin) à la contrainte du modèle gantois. « Copié-collé » dans notre figure 11.1, à titre d’hypothèse, les proportions de ce baldaquin correspondent à celles du mur de la chapelle Vijd.

159

La forme originale du retable haut Le retable supérieur ne conserve nulle part un bord haut original. Nous ignorons donc sa forme primitive. Ses volets ont été amputés dès le XVIe siècle. Dans la tranche supérieure des cadres des volets cintrés, on voit à l’œil nu le champ des panneaux (fig. 11.1g). Normalement, un panneau était embrevé dans la rainure d’un cadre-bâti, et donc invisible à l’extérieur. Coxcie, en 1555-1558 (59 ?), donne à ses volets une forme cintrée qui s’inspire peut-être des volets amputés (fig. 11.4f). Plus tard, à Paris, à la Révolution française, les trois panneaux centraux du retable supérieur seront amputés à leur tour de leur partie haute (fig. 11.3c), jugée « totalement inutile ». Les mesures (hauteur et largeur) données par les commissaires de la Révolution lors du démontage à Gand en 1794 plaident en faveur d’une forme simple. On ne peut toutefois pas exclure une forme cintrée, par exemple pour le panneau central avec Dieu. Le cintre est présent dans les évocations du chef d’œuvre par divers artistes.11 Le panneau central de Coxcie avec Dieu le père conserve à ce jour un bord non peint cintré dans le haut (fig. 11.4b).12 Pour le panneau gantois qui lui a servi de modèle, ma reconstitution propose un cadre central cintré à l’origine, ne fut-ce que pour rappeler notre ignorance de la forme initiale. Des remplages à l’intérieur et des plates peintures à l’extérieur L’ensemble du retable haut devait présenter, à l’intérieur, des remplages dorés sur fond bleu (exemple dans la fig. 11.1b), comme on peut le déduire de la radiographie et de la description des écoinçons faite en 1953 : « une surface à niveler et […] des traces de la couche bleue ancienne ».13 Un médaillon avec ange, comme en représente Jan Gossaert (fig. 11.1d), a peut-être fait partie des remplages eyckiens. Des exemples existent de médaillons peints insérés dans les remplages de retables (fig. 11.1c). Il n’y avait pas de remplages à l’extérieur des volets: les cadres cintrés, dans leur état actuel, s’y profilent en talus, moulure typique qui borde une plate peinture à l’extérieur d’un volet (fig. 11.1e).

160

hélène verougstraete

L’ouverture et la fermeture des volets L’accès aisé aux volets inférieurs rendait simple leur manipulation. La fermeture se faisait probablement avec un crochet pivotant. Pour atteindre les volets supérieurs on utilisait peut-être une perche, comme on le fait encore aujourd’hui à des fins similaires sur le Mont Athos14 ou bien on montait sur une escabelle. On fermait d’abord les volets à gauche (côté Adam), ensuite ceux à droite (côté Ève). La disposition de la feuillure15 correspond à cette séquence (fig. 11.2d), imposée par les règles de préséance. Toujours pour le retable haut, la fermeture de chaque paire de volets se faisait en deux temps, et par des éléments de menuiserie. Voici comment on procédait. On fermait et attachait d’abord les larges volets cintrés. Creusée dans l’arrête des montants des cadres, à la jonction des deux volets d’une paire, on remarque une moulure en cavet (fig. 11.1i). La gorge qui en résulte s’appliquait, à la fermeture, sur une colonnette au cadre central (fig. 11.1k), comme le suggère très justement l’équipe de JeanAlbert Glatigny.16 Ajoutons que l’élargissement de la gorge dans le bas (fig. 11.1f) devait couvrir la base de la colonnette (fig. 11.1k). Mais surtout, la fermeture des grands volets s’accompagnait d’un chevillage. Des trous dans les cavets des cadres d’Adam et d’Ève sont encore visibles (fig. 11.1i). À ces trous devaient correspondre, dans les colonnettes, des chevilles saillantes (fig. 11.1j). Nous avons rencontré ailleurs ce système de fermeture par chevillage. Ici, il visait à soutenir les grands volets, à empêcher qu’ils ne s’affaissent. Le système a très tôt dysfonctionné suite au poids des volets. La rotation des volets se faisait par des charnières à l’intérieur du retable entre partie centrale et volets. À la jonction des volets entre eux, les charnières étaient ferrées sur les plats extérieurs des cadres (fig. 11.1h). À l’ouverture, l’articulation entre les volets était souple. La fermeture, par contre, entraînait les volets de manière rigide, L’ensemble des deux retables, le retable bas posé sur une prédelle et le retable haut surmonté d’un couronnement de remplages dorés sur fond bleu, le tout surmonté d’un baldaquin, avait un aspect

monumental, verticalisant et gothique que Coxcie semble avoir modernisé. Figure 11.2. Le XVIe siècle La destruction de la peinture ornant la prédelle Marcus van Vaernewyck explique que des « mains de veau » (kalvershanden) effacent la peinture de la prédelle avant 1550.17 Ne s’agirait-il pas de mains de calvinistes, animés d’ardeur religieuse? Nous sommes en pleine guerre de religion. De nombreuses estampes satiriques s’amusent à représenter une tête de veau (kalf) sur l’assiette de Calvin. La destruction de la structure de la prédelle elle-même intervint un peu plus tard, de l’action des iconoclastes. L’affaissement des volets hauts et les troubles iconoclastes Dès le début du XVIe siècle, l’affaissement des volets hauts avait certainement rendu difficile la fermeture du retable (fig. 11.2a). Pour les alléger ils furent amputés dans le haut, – nous l’avons déjà évoqué – et les montants centraux effilés vers le haut, pour que la feuillure puisse encore fonctionner (fig. 11.2b). On comprend l’insistance avec laquelle il fut recommandé à Michel Coxcie de soigner son support : « par Maître Michiel, qu’il est force faire sur tablez de bois bien liéez ».18 Il est possible qu’au XVIe siècle Coxcie ait abandonné les remplages au profit de formes cintrées plus renaissantes et plus légères. Les problèmes du retable s’aggravent dans les dernières décennies du XVe siècle. On prévoit un paiement pour « réparer et aider [soutenir] la table d’Adam et d’Ève ».19 Le montant prévu au contrat dans les comptes de la fabrique est élevé. Par mesure de précaution, on n’ouvrira plus le retable qu’aux jours de fête « pour que son état s’aggrave moins par l’ouverture des volets »20 , et on ne percevra plus de recettes pour les visites.21 Certains cadres du retable gantois s’étaient désarticulés, en particulier ceux des anges chanteurs et musiciens. Une photographie faite à Berlin au 19e siècle (fig. 11.2e) montre trois stades de

la présentation de l’agneau mystique dans la chapelle vijd

a

b

c

d

e

Fig. 11.2 Le XVIe siècle

162

hélène verougstraete

ferrures du côté de l’Annonciation: les originales, puis de petites ferrures à la jonction des cintres, et enfin d’amples ferrures non originales et grossières qui cachent un dégât du cadre (ces dernières en rouge sur le schéma).22 Waagen décrit ces grosses charnières: Da die alten Beschläge, die die Flügel untereinander und mit dem Mittelbilde verbanden, nämlich mit der Zeit schadhaft geworden seyn mochten, so dasz man sie durch die sehr plumpen und rohen zu ersetzen nöthig fand, welche man jetzt noch daran wahrnimmt […]23 Ces charnières sont ferrées sur le cintre qui est une menuiserie appliquée, non portante, ne faisant pas partie du bâti initial. Elles datent probablement du moment où on a raccourci les volets. Elles ont alors pour fonction de remédier à la suppression de la partie haute du bâti. Ce qui est curieux, c’est que ces ferrures se terminent par des charnons dans le prolongement du plat extérieur du cadre. Un nœud de charnière, avec ses charnons (enroulements métalliques destinés à recevoir la goupille), forme l’axe de rotation d’un volet. On doit ici supposer que ces charnières de renfort ne faisaient que soutenir les charnières originales normalement situées à l’intérieur du retable (l’axe de rotation entre la partie centrale et le volet se situe toujours à l’intérieur d’un retable), qui avaient dû fissurer le bois. Les charnières de renfort ne pouvaient fonctionner que par un système de nœuds multipliés, et de fortes pentures devaient être attachées également à l’arrière du cadre central. Les revers des panneaux centraux ne montrent aucune trace de traverses arrière originales en bois qui auraient pu consolider un cadre mis à l’épreuve du poids à porter (celui des volets). D’autres grands retables ont de pareilles traverses arrière, par exemple la Descente de Croix et le Jugement dernier de Rogier van der Weyden. Après une mise à l’abri contre les déprédations des iconoclastes, les retables sont placés dans la chapelle Viglius en 1584-1585, pour permettre le réaménagement de la chapelle Vijd. Un tailleur de

pierres et un menuisier sont payés pour préparer la présentation du retable.24 Le baldaquin, l’autel et la prédelle ont probablement été dévastés par les iconoclastes. Les retables sont replacés dans leur chapelle en 1586 et le nouvel autel (celui-là même qui subsista jusqu’en 1951, désigné comme « auteltombeau à couvercle en bois peint» par le chanoine Van den Gheyn ?) est consacré en 1588.25 En 15861589, les archives mentionnent la construction d’un socle (een voet), peint par J. Cools. Les archives ne le précisent pas, mais ce socle devait être destiné au retable haut. Il devait résoudre les problèmes d’affaissement des volets et porter le retable fermé, volets compris. Précédemment, les crochets/ équerres métalliques portaient uniquement la partie centrale. Sur une photographie ancienne, on observe un loquet sur le volet de l’ange de l’Annonciation et des trous laissés par les clous forgés d’un loquet correspondant, du côté de Marie (fig. 11.2c).26 La tige coulissante des loquets plongeait dans le nouveau socle. Les volets, comme avant, se fermaient en deux temps: d’abord par les loquets, puis par la feuillure. Le frottement des volets sur ce socle rendait l’ouverture difficile. Avant 1591-1592, on place une serrure. Celle-était destinée au retable bas : une serrure est incompatible avec une feuillure et, de surcroît, une serrure dans le retable haut aurait été difficile à atteindre. Les comptes de la fabrique d’église (1592-1597) mentionnent que la clé est confiée au trésorier. Le retable n’est plus ouvert que quatre fois l’an.27 C’était pourtant l’ouverture du retable haut qui était difficile, pas celle du retable bas. Ceci semble indiquer qu’on considérait normal d’ouvrir les deux retables en même temps. Figure 11.3. Du XVIIe siècle à la Révolution française Encadrement baroque Autour de l’autel surmonté de ses retables, Boudewijn van Dickele plaça en 1662 un encadrement architectural baroque à colonnes torses et chapiteaux composites, placées à l’avant d’un muret en briques et posées sur de hauts socles. Un

la présentation de l’agneau mystique dans la chapelle vijd

entablement horizontal surmontait les colonnes. Le tout était en bois peint en imitation marbre. Cet encadrement a subsisté jusqu’en 1951. Les deux colonnes torses (fig. 11.3b), aujourd’hui dans la crypte de la cathédrale, ont été examinées par Elisabeth Bruyns qui n’y a vu aucune trace de fixation. Les retables n’étaient donc pas attachés aux colonnes. L’encadrement baroque était simplement disposé autour des retables. La console centrale sous l’entablement a dû servir à maintenir le retable haut, conformément à la pratique baroque, qui faisait suite à la vieille habitude de retenir les retables par le haut pour les empêcher de basculer. Le retable haut avait encore, du temps de Van Dickele, ses remplages sculptés. Quelques photographies montrent, incomplètement, l’encadrement baroque.28 Un relevé sommaire, en plan et élévation (fig. 11.3a), daté de 1920, montre l’entablement dont nous n’avons pas retrouvé de photographie.29 C’est probablement par erreur qu’on a parlé d’un fronton.30 Destruction des cadres centraux Gilberte Émile-Mâle explique que les œuvres acheminées vers la France par les commissaires de la Révolution, pour la grande part des toiles attribuées à Rubens, Van Dyck, Jordaens, de Crayer, Van Thulden… furent extraites de leur cadre avant le transport.31 Les cadres furent abandonnés sur place, les toiles roulées et remises sur de nouveaux châssis à Paris. Parfois, si la toile était cintréé dans le haut, elle était rectifiée. Les tableaux, à Paris, étaient soumis aux traitements jugés nécessaires, puis encadrés avant leur exposition dans le musée. Pour déposer les retables des frères Van Eyck en 1794, les commissaires de la Révolution travaillent deux jours. Ils enlèvent d’abord le retable inférieur, puis le retable supérieur. Ceci démontre bien que les deux retables n’étaient pas posés l’un sur l’autre. Les archives ne disent pas que les commissaires détruisirent les cadres, mais font état de la difficulté de leur travail. C’est l’enlèvement des cadres qui fut difficile. Évidemment, il était plus facile d’enlever

163

les cadres à battée du XVIIe siècle, que les cadres rainurés, à assemblages chevillés du XVe siècle, surtout lorsque ceux-ci étaient consolidés par de fortes ferrures comme c’était le cas à Gand. Après arrachage des ferrures, les commissaires ont, aux angles des cadres, scié le long de la coupe de l’assemblage. Cela permettait de scier le tenon. Mais, ce faisant, on entaille légèrement le bord non peint du panneau original. Pour éviter qu’une fissure s’ensuive, la pratique était de scier le coin du panneau. C’est pour cela qu’il manque aujourd’hui un petit angle aux coins inférieurs (dans le haut, le bord original n’existe plus) (fig. 11.3c). Ce petit angle manquant s’observe dans d’autres tableaux également. Gilberte Émile-Mâle signale qu’il y a des erreurs fréquentes dans les dimensions fournies par les commissaires. Pour les dimensions du panneau de l’Agneau, on constate qu’il y a une erreur, comme on peut le vérifier en comparant les mesures actuelles du panneau restées inchangées, avec les mesures des commissaires par un calcul des proportions hauteur sur largeur. Les dimensions des panneaux du retable haut sont les suivantes: « n° 5120 Eyck van Jean, Dieu le Père/H 2m08 L 0,82 m. Vient de la cathédrale de Gand /n° 5121 Eyck van Jean, La Vierge/H 1m66 L 0,74 sur bois vient…/n°5122 Eyck van Jean. S. Jean/H 1m66 L 0,74. Vient de la cathédrale de Gand ».32 L’examen des rapports entre hauteur et largeur mène à la conclusion qu’il manque environ 5 cm aux panneaux de la Vierge et saint Jean et plus ou moins 22 cm à la Divinité qui est d’ailleurs le plus large des trois panneaux.33 Un document ayant trait à la vente des volets en 1816 dit que ces trois panneaux ont fait partie d’un retable dont : « […] les pièces principales étoient scellées en platre et a cloux ; de manière que pour les enlever les commissaires francois en ont presque demoli le couronnement ».34

hélène verougstraete

b

a

c

Fig. 11.3 Du XVIIe siècle à la Révolution française

la présentation de l’agneau mystique dans la chapelle vijd

Ce couronnement, ce sont les remplages originaux surmontant les trois « pièces principales », chevillés (et cloués ?) au panneau et au cadre, le tout noyé dans l’enduit. Ce terme de couronnement ne désigne évidemment pas le baldaquin, détruit au XVIe siècle. Les remplages sont sauvés à ce stade (presque démolis !), mais ils seront détruits à Paris, comme on le verra ci-dessous. Dans la caisse n° 13 du quatrième envoi (il y en a eu sept) se trouve la partie centrale du registre haut de l’Agneau Mystique, celle du registre bas suivra quelques mois plus tard. Le conseil arrête que les tableaux sur fond d’or arrivés de la Belgique, indiqués dans les états de réception : « sous le nom de Vermeyen, mais crus de Jean Van Eick [sic], seront nettoyés par Reser et que l’excédent des planches qui surhaussent ces 3 tableaux sera scié comme totalement inutile et pour faciliter l’encadrement » (fig. 11.3c).35 À Paris, les écoinçons de part et d’autre des cintres dorés sont grossièrement surpeints en brun. Cette peinture sombre correspond à une pratique courante, à l’époque, pour les parties d’une peinture destinées à être cachées par le cadre. Figure 11.4. Le XIXe siècle La première moitié du XIXe siècle On peut penser qu’en 1816, les Français se contentèrent de restituer ce qu’ils avaient pris… quatre panneaux, et qu’ils ne firent pas cadeau des cadres néo.36 À Saint-Bavon, on fit alors, pour ces parties centrales, de nouveaux cadres permettant leur suspension après le 10 mai 1816 – date du retour en Belgique de ces parties centrales – au sein de l’encadrement baroque resté en place. On sait que sept mois plus tard, les volets originaux furent vendus en Allemagne – sauf ceux d’Adam et d’Ève, remisés probablement séparément. Waagen écrit en 1819 : « les tableaux du milieu étaient jadis dans des cadres très larges, très simples. Ils sont aujourd’hui dans un retable de construction moderne ».

165

Le mot « retable » évoque un triptyque fixe. Le cadre « moderne » était probablement noir et or, comme presque tous les cadres de la cathédrale. C’est peut-être le cadre qu’on voit sur la photo d’ensemble datée de 1912 (Gent, Stadsarchief, scms-fo-5315) (fig. 11.4d).37 Waagen craignit qu’Adam et Ève étaient perdus, et se réjouit lorsqu’il apprit qu’on les avait retrouvés en 1823. La direction de l’église jugea inopportun d’exhiber nos ancêtres, et ces derniers furent remisés. En 1859, la fabrique d’église fait placer à l’autel de la chapelle Vijd une grille de fer pour empêcher les curieux d’escalader l’autel. Des rideaux protègent l’œuvre. Ceci fait penser que les retables étaient suspendus assez haut, et peut-être regroupés. La reconstitution de 1861/1865 En 1861, l’acquisition des volets de la copie de Coxcie permet de reconstituer un ensemble.38 Une convention entre le gouvernement et la fabrique d’église de Saint-Bavon précise la cession d’Adam et d’Ève au Musée des Beaux-Arts à Bruxelles. Une copie « décente » sera exécutée par le peintre Victor Lagye. Le gouvernement offre à la fabrique d’église les volets de Coxcie et paie le montage de l’ensemble, plus quelques autres frais. Pour cette reconstruction on s’inspire du schéma réalisé par Waagen (fig. 11.4c), qui replace les panneaux dans une juste séquence, mais qui esquisse un retable unique. Waagen lui-même a probablement vu les parties centrales rapprochées, telles qu’on les avait suspendues au retour de Paris. La photo de 1912 (Gent, Stadsarchief, scms-fo-5315) (fig. 11.4d) montre le retable tel qu’il fut réassemblé en 1865. Sur des photos plus récentes (1951), on peut voir comment on a procédé (fig. 11.4e). Une caisse sommaire est ancrée dans le mur ; elle réunit les deux retables, les rapproche du spectateur et facilite un déploiement plan des volets devant les colonnes baroques. La moitié basse de la caisse est renforcée d’étais métalliques destinés à supporter le poids des panneaux du haut. À chaque registre, la caisse comporte des battées contre lesquelles reposent les panneaux, et les cadres sont

hélène verougstraete

b

7

10

5

2

9

1

8

3

4

6

11

12

a

c

d

e

Fig. 11.4 Le XIXe siècle

f

la présentation de l’agneau mystique dans la chapelle vijd

appliqués par devant, vissés à la caisse selon un système un peu différent dans le haut et le bas, comme l’explique le chanoine Van den Gheyn en 1920 (voir plus loin). Entre les deux retables subsiste un petit espace sur lequel, après l’exposition de 1920, on apposera le fameux quatrain. On confectionne en 1861 des cadres neufs noir et or pour les volets de Coxcie, cadres qui existent encore aujourd’hui (fig. 11.4f). Les volets Victor Lagye y sont accrochés, les seuls dont les cadres sont dorés, fidèles aux originaux. La photo de 1912 montre le retable reconstitué avec les volets de Coxcie et, pour Adam et Ève, ceux de Lagye. Les volets sont tenus dans un cerclage métallique et pivotent, par paires solidaires, mais indépendamment pour le haut et pour le bas, autour d’une tige verticale ancrée dans l’autel. Ainsi, la partie centrale était libérée du poids des volets. Le rapprochement au sein d’une caisse, à l’intérieur de l’encadrement baroque, des deux retables (dont l’un était amputé de ses remplages), a eu pour conséquence disgracieuse de mettre à nu tout un pan de mur sous l’entablement. On y portera remède en cachant le mur derrière une petite tenture. Le musée de Bruxelles exposa les volets d’Adam et d’Ève dans une armature qui permettait de les faire pivoter. Les cadres ont conservé les trous dans lesquels s’enchâssaient les pivots. On exposait les revers. Pour voir Adam et Ève, il fallait en faire la demande. Les protestations ont fini par convaincre les conservateurs d’exposer les volets sur un socle autour duquel on pouvait circuler. Les photos dont nous disposons de cette présentation des volets au musée montrent les cadres dorés. Figures 11.5 et 11.6. Le XXe siècle La solidarisation définitive des deux retables En 1920, les volets originaux reviennent de Berlin. À Gand, le retable est démonté et envoyé à Bruxelles pour l’exposition Van Eyck-Bouts aux Musées royaux des Beaux-Arts de Bruxelles (14 août-26 septembre 1920).39 Le chanoine Van den Gheyn explique que :

167

« le grand panneau de l’Adoration était fixé aux poutrelles de l’établi [la caisse décrite précédemment] par le cadre qui l’entoure. Les trois panneaux supérieurs sont attachés autrement : le cadre luimême est fixé à la caisse, et le tout est fermé au moyen des baguettes dorées qui sont vissées ». Un courrier entre Van den Gheyn et le conservateur en chef des Musées royaux, Fierens-Gevaert, montre qu’on a envisagé d’envoyer à Bruxelles tout le dispositif avec l’autel tombeau à couvercle et la caisse, mais qu’on a reculé devant la difficulté de l’opération. Sur un document photographique (fig. 11.5b) qui doit dater de cette époque on peut voir que tous les cadres sont dorés, à l’exception de ceux de nos ancêtres (dont on n’aperçoit que le bas) surpeints en sombre. Notons que les volets de Coxcie et de Lagye ont été envoyés eux aussi à l’exposition. Autour de l’exposition, les demandes d’autorisation de photographier affluent.40 Après l’exposition, les cadres dorés des parties centrales seront adoptés pour la présentation à Gand. L’ensemble des originaux est ramené à Gand le 29 septembre ou il sera placé provisoirement dans la chapelle de l’Évêque. Le chanoine Van den Gheyn a envisagé d’exposer simultanément le retable ouvert sur l’autel, et le revers des volets, en face du retable ouvert. Il dût renoncer à ce projet, les volets d’Adam et d’Ève n’étant pas sciés dans l’épaisseur comme les autres. Face et revers des volets sciés dans l’épaisseur sont alors remis dos à dos. Une nouvelle petite tenture remplace l’ancienne, et un store descend à environ 20 cm devant les peintures pour, aux dires du chanoine, « les protéger des excréments des chauves-souris ». L’épisode du vol des panneaux en 1934 a été souvent évoqué. Les cadres sont laissés sur place (fig. 11.5c). Sur une photographie postérieure au vol, on voit que de petites béquilles amovibles solidarisent les volets hauts et bas, probablement pour des raisons de bon maintien des volets, mais sans doute également pour solidariser l’ouverture des volets hauts et bas (fig. 11.5d). En 1950, l’autel et l’encadrement baroque, tombés en disgrâce, sont

hélène verougstraete

a

b

c

d

Fig. 11.5 Le XXe siècle

la présentation de l’agneau mystique dans la chapelle vijd

a

a

c

b

d

Fig. 11.6 Le XXe siècle (suite)

170

hélène verougstraete

cachés par des tentures (fig. 11.6a). Des cadres neufs en bois sont confectionnés pour la partie centrale, les cadres antérieurs étant jugés en mauvais état (fig. 11a). Tous les cadres – originaux et refaits – sont enchâssés dans un cerclage métallique qui existe encore aujourd’hui. Chaque panneau avec son cadre est cerclé d’une cellule rigide (figs 11.6c, 11.6d). Une grosse poignée sous les volets, opère la rotation des volets en plan rigide, haut et bas solidarisés. Tout espace entre les deux retables est définitivement supprimé. Curieusement, on décale vers le haut les volets bas, cherchant sans doute à créer une harmonie des surfaces ; les volets hauts, dont les dimensions ne correspondaient plus à la partie centrale sont également rehaussés. L’ensemble, devenu très lourd, nécessite le remplacement de l’ancien autel en bois par un autel en maçonnerie (fig. 11.6b). En 1986, le polyptyque est déménagé avec son châssis métallique vers la chapelle Villa, à l’extrémité occidentale de la cathédrale, où il se trouve encore en 2015. Figure 11.7. Ces quelques étapes dans la présentation de l’Agneau Mystique dans la chapelle Vijd sont réunies dans la figure 11.7. Questions spéciales Les ombres peintes des cadres (fig. 11.8) Une ombre du cadre, peinte en trompe-l’œil, simule l’ombre formée par la lumière incidente venant de la fenêtre de la chapelle Vijd, située à droite. L’ombre réelle se projetait de cette façon uniquement lorsque ces panneaux étaient perpendiculaires à la fenêtre. On peut y voir une instruction au déploiement correct des volets. Ceci signifie qu’Adam et Ève devaient, à l’ouverture, former un angle droit avec les volets des anges. L’Annonciation, par contre devait épouser le plan du mur. On se trouve devant un système conventionnel de niche dans la niche, de double retrait (le cadre et son ombre, et la niche avec voûte). Le symbole est subtil. À ces deux endroits, l’artiste veut

enfermer, délimiter ou hiérarchiser l’espace, et utiliser le cadre réel (au sens matériel : le bâti de la menuiserie) au même titre qu’une architecture peinte en trompe-l’œil, niche ou arcade. Dans le même ordre d’idées, on relèvera que certaines niches du polyptyque sont des espaces clos voûtés. D’autres figurent un espace ouvert et non voûté. Rien de tout cela n’est peint au hasard, mais l’interprétation reste encore largement à examiner par des spécialistes de l’iconographie. Le texte dans la banderole du ‘Pressbrokat’ des deux intercesseurs Marie et saint Jean (figs 11.9, 11.10) Il y a des textes sur les banderoles du brocard appliqué derrière les grands personnages centraux. Les feuillets du brocard derrière la Divinité sont de fabrication différente de celles derrière les Intercesseurs Marie et saint Jean, et ils sont aussi disposés autrement sur le panneau. Partout les feuillets – ou parties de feuillets – sont alignés en bandes verticales, (3 feuillets superposés côté cadre, à droite et à gauche de chaque personnage, deux feuillets partiels jouxtant les personnages. Derrière la Divinité, la bande qui jouxte le personnage comporte des feuillets disposés en quinconce par rapport à l’autre bande de feuillets. Derrière les deux Intercesseurs, les feuillets sont alignés horizontalement sur les deux bandes de feuillets. Les banderoles sont différentes également, soit semi-ciculaires (Divinité), soit, pour les Intercesseurs, en enroulement vers le bas (partie gauche de la banderole) remontant ensuite vers un enroulement haut (partie droite). Les textes derrière la Divinité ont été étudiés, mais ceux derrière les Intercesseurs ont été jugés en trop mauvais état, ou considérés comme composés de caractères fantaisistes. Tous les textes sont très surpeints et leur étude ne pourra pas faire l’économie d’un examen minutieux au microscope. L’examen des radiographies que j’ai pu faire en été 2012 m’a permis de constater que ce document révèle, dans les rares zones très bien conservées du Pressbrokat derrière saint Jean, un texte ou une partie de texte. En effet les textes ont été gravés dans les banderoles, et pour cette raison ils se voient, sur

la présentation de l’agneau mystique dans la chapelle vijd

15th century

before 1557

after 1588

1662

1794

1816

1859

1861 - 1865

1920

1934

before 1950

1950

1986 -

Fig. 11.7 Résumé des étapes dans la présentation des retables dans la chapelle Vijd

hélène verougstraete

Fig. 11.8 L’ombre des cadres est peinte en trompe-l’œil à deux endroits : sur les volets d’Adam et celui d’Ève, et dans l’Annonciation. Sur les panneaux d’Adam et d’Ève, l’ombre formée par les cadres correspond à une ouverture des volets « en angle », en plan plus ou moins perpendiculaire à la fenêtre de la chapelle Vijd qui se situait à droite. Dans l’Annonciation, l’ombre s’allonge en oblique de manière identique sur tous les volets fermés.

la présentation de l’agneau mystique dans la chapelle vijd

173

a

b

c

Fig. 11.9 Détails du brocard appliqué en radiographie (à gauche) et en lumière ordinaire (à droite). De haut en bas : banderole en-dessous et à droite de la main bénissante de saint Jean. Les radiographies montrent un « B » qui diffère d’une banderole à l’autre, ce qui implique qu’il est tracé à la main sur chaque banderole. Dans le bas : partie d’une banderole à droite du manteau de saint Jean. La radiographie permet de voir la seconde partie de l’inscription.

le document, en sombre par rapport à ce qui les entoure. Les caractères identifiés sur la radiographie permettent de recomposer l’inscription complète d’une banderole. Soulignons-le, l’image radiographique ne rend pas compte d’un texte peint, mais de densités des matériaux en présence,

ici d’un relief : le texte s’inscrivant en creux dans les banderoles. La radiographie oblige à raisonner en « densités des matières ». De manière générale, un texte peint ne se voit sur une radiographie que s’il est peint avec un matériau de forte densité. Les textes sont

174

hélène verougstraete

souvent écrits en couleur sombre, avec des matières de faible densité ; ils sont alors imperceptibles sur une radiographie. La radiographie dans le cas présent n’aide pas à lire les textes peints. Il y a tellement de surpeints qu’on ne peut pas non plus se baser sur la lecture des caractères visibles à l’œil nu. C’est l’examen au microscope qui devra déterminer si dans les textes creusés dans les banderoles, observés à la radiographie (ce creux est visible également à l’œil nu), l’artiste mit, dès l’origine, un trait au pinceau, si la couleur qui forme le texte visible en surface aujourd’hui est en partie originale. Ces quelques explications méthodologiques m’ont parues nécessaires parce que certains des médiévistes et paléographes que j’ai consultés étaient perplexes devant une radiographie et sa relation avec l’image de la surface de la peinture, perplexes également devant le lien à établir entre l’aspect des lettres et les matériaux du Pressbrokat. On n’incise pas un texte dans un Pressbrokat comme on en trace un avec une plume libre sur le parchemin. L’originalité de l’inscription n’a jamais été mise en doute. Elle est formée dans le brocard appliqué qui est lui-même par endroits recouvert par la peinture originale. Par exemple, des cheveux de Marie recouvrent une zone de brocard. On peut dire à partir de la radiographie que la même inscription se répétait sur toutes les banderoles derrière Marie et derrière saint Jean. Même sur les banderoles en mauvais état, on reconnaît l’une ou l’autre lettre. Derrière Marie toutes les banderoles sont en mauvais état : deux banderoles complètes, une autre banderole presque complète et deux demi-banderoles. Derrière saint Jean il y a par contre des banderoles en excellent état, sous son bras bénissant : deux demi-banderoles, une au dessus de son bras, avec des textes – chaque fois la première partie de l’inscription – bien lisibles (nous ne parlons pas de la peinture, comme nous l’avons déjà précisé, mais de l’état du relief du brocard avec son sillon gravé par l’instrument dans la banderole). A droite, au dessus de l’épaule de saint Jean, une demi-banderole présente un texte bien conservé également : la seconde partie de l’inscription. Au sujet de l’état de

conservation, les anciennes copies et transcriptions sont intéressantes (fig. 11.10) parce qu’elles restituent l’état de ces textes à une certaine époque, mais sans doute l’état des textes peints, éventuellement déjà avec des surpeints ou retouches. Ce qu’on peut préciser également, c’est que le texte est inscrit à la main, banderole par banderole. L’inscription n’est pas conçue dans la matrice du brocard. Deux arguments permettent de le dire. Les zones plus claires sur la radiographie résultent d’une densité plus forte du matériau radiographié. Ce sont les zones en relief sur le feuillet, qui se sont formées dans un creux de la matrice. Autrement dit, le motif creusé dans la matrice devient relief sur le feuillet qu’on colle sur le panneau. Si ces caractères de l’inscription apparaissent en sombre sur la radiographie, c’est qu’ils sont en creux dans les feuillets du Pressbrokat posés sur le panneau. Pour que le creux dans le feuillet du brocard provienne de la matrice, le texte aurait du être façonné en relief dans la matrice, un travail fastidieux, voire impossible à réaliser. En outre on voit bien sur la radiographie que le texte n’est pas « épargné » ; il est tracé librement. L’instrument qui a réalisé les traits dans le feuillet du Pressbrokat (et pas dans la matrice, comme on l’a dit) a formé de petits bour-

Fig. 11.10 Le texte copié par Schultz en 1824 et celui transcrit par De Bast la même année. Ces témoignages reflètent l’état de conservation de l’inscription peinte (avec des retouches ?) au premier quart du XIXe siècle.

la présentation de l’agneau mystique dans la chapelle vijd

relets de matière en creusant son sillon. Il poussait de la matière devant lui, empêchant donc de lier les traits sous peine de les engorger. Le matériau constituant le brocard rendait le tracé difficile, obligeait à simplifier les formes et empêchait éventuellement de lier les traits. Les matériaux du Pressbrokat n’étaient pas propices à un tracé de grande qualité et donc pas tout à fait comparables à des modèles soigneusement calligraphiés sur parchemin.41 Le second argument qui permet de dire que le texte est tracé à la main directement dans les feuillets collés sur le panneau, est le fait qu’on observe de nettes différences dans l’exécution d’une même lettre, d’une banderole à l’autre. Nous avons un exemple dans les deux premières lettres sur deux banderoles différentes (fig. 11.9: documents radiographiques à gauche, haut et centre) ; la première lettre, en haut à gauche, très maladroite, a suscité le scepticisme de certains médiévistes consultés. La seconde, au milieu à gauche, est acceptée comme étant un B. Mais d’une banderole à l’autre, c’est bien le même texte qui est répété. Nous pensons que la première partie du texte peut se lire comme Brur L [Lubbrecht] (frère L[ubrecht]).42 Après un B et un r, le troisième signe est l’abréviation (latine,  ou inspirée du système abréviatif latin ?) de « ur ». Cette abréviation est inconnue de la plupart des médiévistes que j’ai consultés à son sujet, seul l’un d’eux évoquait la possibilité d’une abréviation par superposition de deux caractères. Elle est pourtant recensée dans la littérature.43 J’ai trouvé à deux reprises la mention de cette abréviation par réunion de deux caractères (en l’occurrence le u et le r, pour « ur ») dans des ouvrages de paléographie. Ce système d’abréviation par réunion de lettres est utilisé entre le VIIIe et le XVIe siècle, surtout par les graveurs, pour gagner de la place. Mediavilla mentionne neuf exemples d’abréviations de ce type, parmi lesquels le « ur » ; la transcription de cet auteur est un peu plus aplatie que dans la banderole de Van Eyck, mais très semblable. Cet auteur souligne par ailleurs que ces modes abréviatifs ont adopté des formes variées et capricieuses. Parmi les autres abréviations citées

175

par Mediavilla il y a le « cum », « er », « us », « rum », abréviations latines. Le « ur » pourrait donc être une abréviation empruntée au latin, comme pourraient l’être également les deux derniers caractères de l’inscription sur la banderole à droite de saint Jean (fig. 11.9, en bas à gauche). En effet, sur la seconde partie de la banderole on peut lire sur la radiographie « 27 » suivi de deux caractères. Bernard Coulie44 interprète ces deux derniers signes de la manière suivante : « Si les deux premiers signes peuvent être lus comme les deux derniers chiffres d’une date ([14]27), les deux signes qui suivent indiquent sans doute l’ère de datation, comme il est d’usage. Dans ce cas, ils indiquent l’ère du Christ, le premier signe représentant le mot ‘Christus’, et le second ‘Iesus’.45 Une autre lecture consisterait à voir dans le second signe la finale grecque « -ou », ligaturée selon l’usage grec, permettant de lire ‘Xristou’ ». Il est acquis qu’il subsiste au moins un texte entier bien conservé, réparti sur deux demi banderoles, lisible non au stade de la peinture parce qu’il y a trop de surpeints, mais sur la radiographie qui reflète le sillon gravé par l’instrument dans la matière du brocard. La radiographie en rend une image on ne peut plus objective. Et ce même texte était répété sur toutes les banderoles derrière les Intercesseurs. Pour certains historiens, les caractères sont fantaisistes. Ces derniers historiens pensent donc que Marie et saint Jean interviennent auprès de la Divinité en termes fantaisistes (répétition à l’identique d’un même texte et fantaisie sont pourtant contradictoires  !). Nous avons l’intime conviction que le texte dans cet endroit solennel du retable a un sens, qu’il doit être lu, et nous proposons la lecture qui s’est imposée à nous : Frère L[ubrecht /†14]27 de l’ère du Christ. Le tracé laborieux dans le matériau du brocard, les emprunts éventuels au latin ou à un système d’abréviation latine et les surpeints contribuent à la difficulté de la lecture de ce court texte, lui donnant un caractère tout à la fois malhabile dans l’exécution

176

hélène verougstraete

matérielle et savant dans sa formulation, qui a conduit à son incompréhension. Nous concluons donc que Jan van Eyck fait intercéder Marie et saint Jean en faveur de son frère Hubert décédé en [14]27 de l’ère du Christ. Le texte du Pressbrokat précise la nature du lien entre les deux retables et devra mener à une meilleure compréhension de l’iconographie de l’ensemble, et par la même occasion éclairer le sens du célèbre quatrain que nous pensons être de la main de Jean, conçu dans un même élan de profonde piété fraternelle. NOTES * Pour nous avoir facilité l’accès aux archives et à la documentation, merci à Véronique Bücken et Michèle Van Kalck (KMSKBMRBAB, dossiers 5521 et exposition 1920), Christina Ceulemans, Christina Currie, Bart Fransen, Hélène Dubois, Lyvia Depuydt, Catherine Fondaire, Dominique Deneffe, Dominique Vanwijnsberghe, Elisabeth Van Eyck et Bernard Petit (KIK-IRPA). Merci à Ron Spronk et Hugo Van der Velden pour les discussions fructueuses. A l’Université catholique de Louvain, Paul Augustin Deproost et Baudouin Van den Abeele se sont penchés sur texte du Pressbrokat ; Bernard Coulie, spécialiste des civilisations grecques et byzantines, propose une lecture de deux dernières lettres de ce même texte ; Elisabeth Bruyns a aidé à la recherche de photographies dans les archives de la Commission des Monuments et Sites, dans la photothèque des Archives de la Ville de Gand ainsi que celle de l’IRPA. 1 Coremans 1953. 2 Dhanens 1965 ; Dhanens 1980. Le relevé architectural de la chapelle qui figure dans mes planches est emprunté à cet auteur. 3 La photo de 1912 conservée aux Archives de la Ville de Gand (Stadsarchief, SCMS-FO-5315) est à notre connaissance la plus ancienne qui représente un ensemble (fig. 4, d). 4 Au XIXe siècle, le sort des copies et des originaux est lié. Bettina von Roenne, qui a réalisé une thèse de doctorat sur les cadres néo-classiques de Schinkel dans les musées allemands, apporte un éclairage sur le sujet dans ce même volume. 5 Verougstraete-Marcq, Van Schoute 1989, p. 275-284. 6 Coremans 1953, p. 38. Comptes de la fabrique d’église, 15841585. Roger Marijnissen et Antoine De Schryver s’interrogeaint sur le sens de ces mots. 7 Les chiffres “27” sont inscrits sur les banderoles du drap d’honneur de St Jean et de Marie. Voir dans cet article: III. Questions spéciales. Le texte dans les banderoles du Pressbrokat, et fig. 9, en bas à gauche. 8 Récemment, Hugo Van der Velden a proposé une répartition convaincante du travail entre les frères : Van der Velden 2011, p. 5-39. 9 Voir à la fin de cet article ce qui concerne les ombres peintes des cadres. 10 Dhanens 1980, p. 86, p. 115. Dans l’interprétation des trois personnages principaux de l’Agneau Mystique, attribué à Jean Gossaert, début XVIe (Madrid, Museo del Prado) l’auteur voit une évocation du baldaquin originel du retable. 11 Michel Coxcie, d’après l’Agneau Mystique, six volets à Bruxelles, Musées royaux des Beaux-Arts de Belgique, inv. nos. 6696 6701, l’Adoration de l’Agneau et la Divinité, Berlin, Bode-Museum, inv. nos. 524-525, la Vierge et saint Jean Baptiste, Munich, Alte Pinakothek, inv. nos. 653-654; ces différentes œuvres ont été réunies pour l’exposi-

tion Michiel Coxcie. The Flemish Raphael (M-Museum, Leuven, 31 October 2013 – 23 February 2014). Jan Gossaert, Le Christ entre la Vierge et saint Jean Baptiste, Madrid, Museo del Prado, inv. no. P01510 (fig. 1, d). D’après Van Eyck, La fontaine de vie, Madrid, Museo del Prado, inv. no. P01511 (fig. 1, a). 12 J’ignore si les panneaux de Marie et saint Jean de Coxcie sont cintrés. 13 Coremans 1953, p. 100. 14 Communication de Michael Lomax. 15 Système de fermeture par évidemment à mi-bois alterné des montants, voir fig. 11.2d. 16 Glatigny et al. 2010. 17 Den Spieghel, 1568, fol. 119 cité dans Coremans 1953, p. 36. 18 Coremans 1953, p. 36. 19 « … te repareren ende hulpen het tafereel van Adam ende Eva ». 20 « oem dat se min verarghert soude wesen duer het openstaan ». 21 Coremans 1953, p. 38. 22 La Haye, RKD, Archives Friedländer. 23 Cité dans Coremans 1953, p. 47. « Les anciennes ferrures, qui unissaient les volets entre eux et à la partie centrale, s’étaient abîmées au cours des temps, de sorte qu’on a jugé nécessaire de les remplacer par des charnières très lourdes et rudes, qu’on peut encore voir aujourd’hui […] ». 24 Comptes de la fabrique d’église 1584-1585, fol. 30 et 30v. Dhanens 1965, p. 116. 25 Dhanens 1965, p. 39. 26 Archive Friedländer. RKD, La Haye. 27 Coremans 1953, p. 42, 116. Comptes de la fabrique d’église (1592-1597, fol. 33v. Dhanens 1965, p. 116-117). 28 Une photographie de 1912 (Ghent, Stadsarchief, SCMSFO-5315), fig. 4, une autre, après le vol de 1934, fig. 5, et une de 1951, fig. 4. 29 Bruxelles, Archives des Musées royaux des Beaux-Arts de Belgique Agneau Mystique, exposition 1920. 30 Coremans 1953, p. 39 31 Émile-Mâle 1994. 32 Émile-Mâle 1994. 1 DD 17, Inventaire Napoléon: Tableaux flamands, p. 190. 33 Environ 22 cm correspond à ce qu’on avait dû enlever au XVIe siècle aux volets correspondants avec Adam et Ève. 34 Coremans 1953, p. 41, n° 29  . NB. Ce texte est écrit à un moment où les parties centrales sont revenues de Paris avec un « couronnement » détruit. 35 Émile-Mâle 1994, p. 66. Paris, Musée du Louvre, Archives 1 BB 4, séance n° 169 du 13 fructidor an 6 (30 août 1798). 36 Je me souviens avoir vu, il y a une quarantaine d’années dans les caves du Musée du Louvre un grand entrepôt de cadres néo-classiques, probablement ceux qui avaient été fabriqués pour le Muséum de Napoléon. 37 Il serait utile de faire rechercher les anciennes photos de l’Agneau Mystique. Dans les archives des Musées royaux des Beaux-Arts de Bruxelles, dossier exposition 1920, il est fait état de diverses campagnes de photographie. 38 La copie de Coxcie date de 1555 (date inscrite sur la fontaine). Van Vaernewyck mentionne également les dates de 1558 et de 1559. Déjà en 1824, Waagen dit la copie de Coxcie démembrée. 39 Le Bulletin de l’Art ancien et moderne des 10 septembre et 26 octobre 1920, fascicules 650 et 653 comporte des articles sur l’exposition illustrés par des photos. Voir le dossier d’archives 5521, exposition de 1920 aux Musées royaux des Beaux-Arts de Bruxelles. 40 Parmi les photographies citées : les établissements Jean Malvaux (« typogravure, photochromogravure, photogravure, photolithographie ») demandent l’autorisation de réaliser une reproduction en couleurs ; une maquette (H. 116 ≈ L. 78 cm) en noir et blanc est montée dans un cadre doré ; elle est conservée aux Musées royaux des Beaux-Arts de Bruxelles. Elle était à l’époque prêtée aux conférenciers.

la présentation de l’agneau mystique dans la chapelle vijd

41

Kirchner 1955, p. 57, table, no. 8. Hussmann 1977, p. 154,

158. 42 Dhanens 1980, p. 28 : Meester Luberecht, Ubrecht, Hubrechte den scildere, Lubrect Van Heycke. 43 Mediavilla 1993, 162-163. Album palaeographicum XVII Provinciarum 1992, p. XL-XLI.

177

44 Professeur à l’Université catholique de Louvain, philologie et histoire orientale; ancien recteur, président de l’institut de recherche en civilisation, arts et lettres INCAL. 45 Cappelli 1990, p. 401-402 et 403.

Fig. 12.1 Groundplan of the Gemäldegalerie Berlin (Waagen 1830)

12

The Frames by Schinkel for the Wings of the Ghent Altarpiece and the Copies in Berlin Bettina von Roenne

ABSTRACT: Six wings of the Ghent Altarpiece constituted one of the highlights of the picture gallery in Berlin for almost a century. To be exhibited in the newly-opened museum, the wings were given picture frames designed by the architect Karl Friedrich Schinkel. The frames are lost today or have not survived in their original format. Since Schinkel’s significant work as a framer had been long forgotten, his frame designs for the wings of the Ghent Altarpiece had attracted no attention by historians up until now. However, the surviving frames of the altarpiece copies in Berlin and a drawing by the architect give an impression of how the Schinkel frames for the wings might originally have looked.

—o— In 1821, Friedrich Wilhelm III of Prussia bought six wings of the Ghent Altarpiece from the collection of the Englishman Edward Solly.1 These comprised the Singing Angels and Musician Angels from the upper register, with the Angel Annunciate and the Virgin Annunciate respectively on their reverse sides, and the Just Judges, Knights of Christ, Hermits and Pilgrims from the lower register, with Joos Vijd, John the Baptist, John the Evangelist and Elisabeth Borluut respectively on their reverse sides. From 1830 to 1920 the panels were on permanent display in the newly opened Gemäldegalerie in Berlin, where they were one of the highlights of the museum.2 The Gemäldegalerie, designed by Karl Friedrich Schinkel (1781-1841), was located on the first floor of the Königliches Museum,

known today as the Altes Museum. Schinkel was also asked to design new frames for the six altarpiece wings. Schinkel was not only an architect; he was an all-round designer. He designed each and every exterior and interior detail of his buildings, often right up to the picture frames.3 This approach is exemplified in the Königliches Museum, one of his architectural masterpieces. For several reasons, more than half of the 1198 exhibited pictures had been frameless before the gallery’s opening.4 To remedy this, over 600 museum frames after Schinkel’s design were produced between 1827 and 1830. However, lack of time and cost restrictions meant that most of the pictures were given different but standardized frames with simple neoclassical patterns.5 For a few exceptional works by famous painters the architect was granted permission to design extravagant and therefore expensive frames. The ‘Van Eyckian’ pictures were clearly considered part of this category. This can be seen in a letter from Wilhelm von Humboldt to Gustav Friedrich Waagen written in 1829.6 Von Humboldt writes that concerning the picture framing he will only permit extravagant frames in the future ‘...for very special pictures such as the remaining Raphaels and the van Eyckian...’7 There are no pictures showing the gallery interior in 1830, nor is there a sketch of the precise hanging at that time. However, the spatial distribu-

180

bettina von roenne

tion of the pictures and the hanging can roughly be inferred from the museum catalogue of 1830 and the museum groundplan (fig. 12.1). According to the ground plan the wings had been exhibited left from the central rotunda in the first room of the Netherlandish section on the backside of the first room to the Italian section. In the catalogue, the panels of the Ghent Altarpiece are listed in order of their arrangement on the wall.8 The interiors of the wings were placed next to each other: the Just Judges, Knights of Christ, Singing Angels, Musician Angels, Hermits and Pilgrims.9 Contrary to the original concept of the altarpiece, the two wings of the upper register were placed in the middle of the four wings of the lower register.10 An eyewitness account from 1868 states that the altarpiece wings were turned twice a week, Saturdays and Wednesdays, so that the insides and outsides of the wings could always be seen for half a week each.11 Therefore the panels had to be mounted on the wall in such a way that they could be turned individually. This can also be inferred from the catalogue of 1830, where the outside paintings of the wings are listed in an order that is only possible if the panels are singly mounted.12 If the lower two wings had been attached in Berlin as they originally had been, the portrait of the donor Joos Vijd would have been first on the left in the row. In the picture gallery in Berlin, however, John the Baptist appeared on the left, followed by Joos Vijd, the Angel Annunciate, the Virgin Annunciate, Elisabeth Borluut and John the Evangelist. It is worth noting that reading the phrases of the quatrain on the outside panels in the right order was made impossible by this arrangement. In the picture gallery of 1830 two copies of the central panels of the Ghent Altarpiece were placed next to the six original wings: the Deity Enthroned (fig. 12.2) and the Adoration of the Lamb, both by Michiel Coxcie, who copied the whole altarpiece for Phillip II of Spain from 1557 to 1559.13 In addition to these older copies, Carl Friedrich Schulz copied the Virgin Enthroned (fig. 12.3) and John the Baptist in 1826.14 Although the Schulz copies

were exhibited much later (1883), they were given the same type of Schinkel frame as were the six wings of the altarpiece and the Coxcie copies. After 1878, the interior of the picture gallery in Berlin was redesigned. The Schinkel frames were removed and the wings of the Ghent Altarpiece were mounted in a wall so that both sides could be viewed from different rooms.15 This arrangement can be seen in a painting of the gallery’s interior by Carl Bennewitz von Löfen from the early 1880s (1881-1884) (fig. 12.4),16 where in place of the Schinkel frames there is one dark frame bordering the six wings and covering the original frames.17 That Coxcie’s copy of the Adoration of the Lamb still had its Schinkel frame at that time is shown in von Löfen’s painting; his version of the Adoration hung above the newly framed original wings. At what date Schinkel’s frame of the Adoration of the Lamb was lost is not documented, but it was probably during the Second World War.18 The only remaining Schinkel frames of the ‘Van Eyckian’ paintings today are for Coxcie’s Deity Enthroned (fig. 12.2) and the two copies by Schulz, the Virgin Enthroned (fig. 12.3) and John the Baptist.19 The three frames are very similar, and indeed, the Schinkel frames for Schulz’s copies are identical. On the top and the sides of the frames the profile consists of meandering leaf tendrils on the hollow between two beads. Rosettes are placed on the middle of the round arched upper frame end and above the plinths of the engaged columns. These are ornamented by Gothic style windows on the frames of Virgin Enthroned and John the Baptist. The bottom profile of the frames is characterized by a sill ornamented with trefoils (fig. 12.5). The moulding of ‘Van Eyckian’ Schinkel frames was always the same. Only the design of their upper frame end differed, and this depended on the painted image area. Contrary to the original panel, that of Coxcie’s copy of Deity Enthroned is rounded at the upper end. The panels of the copies of Virgin Enthroned and John the Baptist by Schulz are rectangular at the top, as in the original panels.20 All

the frames by schinkel for the wings of the ghent altarpiece and the copies

Fig. 12.2 Michiel Coxcie, copy of the Deity Enthroned, 1558, with historicizing Schinkel frame

181

182

bettina von roenne

three copies have Schinkel frames that extend to a rectangle with the help of spandrels, which is typical for classical frames and especially for Schinkel (fig. 12.6).21 The Schinkel frames for the copy of the Adoration of the Lamb panel and the original wings are lost today. Nonetheless, we know the design of their upper frame ends from a full-scale sketch by Schinkel. In this drawing we can see that the frames for rectangular pictures each had a rosette in the upper corner (fig. 12.7).22 There is one Schinkel frame left today, identical to the frames for the rectangular pictures of the Ghent Altarpiece. This is the frame for Goswin van der Weyden’s Madonna and Child with Donors (fig. 12.8).23 In 1830, the work was attributed to the ‘Van Eyckian school’ (van Eyck-Schule) and exhibited in the same room as the altarpiece wings, which is probably the reason why it was given the same type of Schinkel frame. Schinkel had no model for the composition of the original architectural frame of the Ghent Altarpiece. But obviously he took inspiration for his frame from the pictures themselves. The trefoil ornament in the frames’ spandrels refers to the spandrel design of the painted stone alcove in the four lower panels of the closed altar. The window ornament on the frames’ plinths of Schulz’s copies of Virgin Enthroned and John the Baptist refers to the windows in the background of the painting of the Archangel and the Virgin on the outside of the altarpiece as well as to the wooden architecture in the Singing Angels and Musician Angels. The leaf tendrils on the frames’ mouldings pose a riddle at first, since they did not appear in the Gothic period. However, it is likely that the vegetation of the Adoration of the Lamb inspired Schinkel’s leaf tendrils. This could have been the reason why the architect chose the meandering variation of the ornament rather than the static one, both of which appear on his sketch (fig. 12.7). In the design of the frames the architect did not aim for a complete copy of a Gothic frame. He used Gothic elements in his own, specific and some-

times idiosyncratic way. One example is the quadratic rosettes that are modelled on Gothic finials (fig. 12.8). They look like the top view on a finial. The shape of the rosettes may also be inspired by the golden fleurs-de-lis ornaments with a sapphire in the middle of the crown of the Deity Enthroned. Another curious feature is the sill of the frames of the Virgin Enthroned and John the Baptist. Two little knobs suggest that this could be opened by a hatch. Indeed, there is an empty hollow space in the sill, which is tightly closed by nails today. In the Gothic period such hollow spaces in altarpiece frames had been used to hide relics.24 Schinkel probably used this element to make the frames appear more authentic. All the detailed and elaborate ornaments on the wooden frame are made of lead and fixed with tiny nails. Only a few very small filigree elements are in chalk ground pastose painted on the frame, as at the base of the engaged column. The delicate designs in lead can be seen in the meandering leaf tendrils on the hollow on the frame mouldings. Ornaments made of lead are the most characteristic feature of the Schinkel frames. Lead was easy and fast to produce for a reasonable price but its main advantage was that it could be easily adapted to the shape of the moulding. In addition, Schinkel’s lead ornaments appear much more sculptural than the stucco cast frame elements that were typical in France and in other German regions as well as Berlin at the same time. There has not yet been any analysis of the gilding of the Schinkel frames for the ‘Van Eyckian’ paintings. But it is probable that these were oil gilded like most of the Schinkel frames for the picture gallery.25 That the architect preferred flat oil gilding for the museum had been often criticized.26 But by choosing a flat gold colour instead of a bright gold tone it seems that he consciously opted for a harmonic adaption of the new mouldings to the historic pictures on exhibition in the gallery. Due to the restrained colour of the gilding the Schinkel frames did not come to the fore but rather receded from the pictures.

the frames by schinkel for the wings of the ghent altarpiece and the copies

Fig. 12.3 Karl Friedrich Schulz, Virgin Enthroned, 1836, with historicizing Schinkel frame

183

184

bettina von roenne

Fig. 12.4 Karl Bennewitz von Löfen, interior of the Gemäldegalerie Berlin, 1881-1884

As is well known, all the wings of the Ghent Altarpiece still have their original frames today.27 The six wings were sawn into two pieces in Berlin in 1894 so that all the paintings could be viewed next to each other. Of the former six wings, there were now twelve panels with twelve frames.28 However, during Schinkel’s time the panels still retained their original format. Gustav Friedrich Waagen described their frames in a publication about the altarpiece in 1824.29 As is the case now, the frames were gilded on the inside and painted with a dark colour on the outside. In 1823, a green colour on the outer side of the wing frames was removed, which revealed the famous quatrain on the bottom of the frames.30 In his text Waagen also mentioned some of the other inscriptions on the frames, documenting the fact that they could also be seen at that time.31 In his 1830 museum cata-

logue Waagen mentioned the original frames with the quatrain on the outside wings, but not again the gilded back of the frames nor the original bordering of the upper wings.32 Schinkel was certainly allowed to design special frames for the ‘Van Eyckian’ paintings.33 Of course this permission was not only given for the copies, which still have their Schinkel frames. The aforementioned sheet of sketches by Schinkel shows a drawing of a round-arched frame in the centre, roughly drafted, with a lunette separated from the rectangular picture below (fig. 12.7). This composition corresponds to the Angel Annunciate and the Virgin Annunciate on the closed side of the panels. Two undated photographs from the Friedländer archive show the upper part of these two wings with their Schinkel frames.34 These photos and the remaining frames on the copies prove

the frames by schinkel for the wings of the ghent altarpiece and the copies

185

Fig. 12.5 Detail of fig. 12.3, the lower frame edge

that the inside and the outside of the altarpiece wings had the same style of Schinkel frame. That means that they always had the same pattern with varying upper edges. Accordingly, the interior Just Judges, Knights of Christ, Hermits and Pilgrims panels must have had the same type of frame as Goswin van der Weyden’s Madonna and Child with Donors has today (fig. 12.8). On the closed side this frame type was present on the panels depicting Joos Vijd, John the Baptist, John the Evangelist and Elisabeth Borluut. Whereas the Singing Angels and Musician Angels on the inside and the Angel Annunciate and the Virgin Annunciate on the outside had the same frame ends as the copies by Schulz (fig. 12.3).

According to the photos in the Friedländer archive, Schinkel’s picture frames covered the original frames of the altarpiece wings.35 Whether the quatrain on the original outside frames had also been covered or still could be seen at that time cannot yet be proven. But it is certainly important to note that the quatrain was mentioned in the museums catalogue of 1830.36 Since the altarpiece wings could be turned individually in the gallery in 1830, the six panels did not have an overall frame, but single frames, which had been either mounted on both sides or completely enclosed each wing.37 There are several indications that two of the Schinkel frames for the altarpiece wings were not

186

bettina von roenne

Fig. 12.6 Detail of fig. 12.3, the upper frame edge

destroyed at the time of the new framing but rather reused for the copies of the Virgin Enthroned and John the Baptist by Schulz. One can see on the frames today that they were adapted to the pictures since the rectangular panels fit into a former roundarched rebate.38 In addition, the altarpiece wings had already received new frames by the time the two copies by Schulz were exhibited in the gallery in 1883, which makes the reuse of the Schinkel frames probable. To which of the six wings the Schinkel frames of the two copies once belonged is an open question. Due to the former round arched

rebate it could either have been the frames for the musicians on the inside or the Annunciation pictures on the outside of the altarpiece.39 It is more likely that the hatch in the sill to hide relics belonged to the Annunciation. The same applies to the little window ornament on the base of the engaged columns, which appears to refer to the windows in the background of the pictures with the Annunciation. Schinkel’s museum frames were often criticized in the past for the reason that neoclassical nineteenth century patterns were not appropriate for

the frames by schinkel for the wings of the ghent altarpiece and the copies

187

older pictures. It is true that due to the great quantity of paintings that had to be framed before the opening of the Gemäldegalerie as well as for reasons of cost, most of the architect’s museum frames had been standardized. But the critics disregarded or were simply unaware of the museum frames for the paintings of famous artists that Schinkel tailored to the pictures by choosing ornaments of the period of origin and in harmony with the subject. That was totally new in Schinkel’s time and set a new precedent for the framing of historic pictures. The comparison of Schinkel’s frame for Coxcie’s copy of the Deity Enthroned with the nineteenthcentury frame of Coxcie’s Virgin Enthroned in Munich reveals this. Karl Albert von Lespiliez designed the Munich frame in 1825 after a typical French pattern.40 Neither the moulding nor the ornaments make reference to the period of origin or the subject. The stucco cast ornaments appear less sculptural than Schinkel’s lead ornaments.

Fig. 12.7 Karl Friedrich Schinkel, full-scale sketch for the picture frames of the Ghent Altarpiece wings

To sum up, Schinkel’s frames for the six altarpiece wings were specially designed for the presentation of the ‘Van Eyckian’ pictures in the newly opened picture gallery in Berlin in 1830. Most of these frames were then destroyed after the new framing of the panels in the early 1880s. Since the Schinkel frames for three copies of the Ghent Altarpiece survived and probably two of them once belonged to the original wings, we know the type of the mouldings of the lost frames in detail. The profile on the top and the sides of the frames and the sill at the bottom of the frame is always the same. Only the design of the upper frame end varies significantly, depending on the pictorial design. The frames of the pictures that are rounded at the upper edge take the form of a rectangle with the help of a spandrel and Schinkel’s frames of rectangular pictures each had a rosette in the upper corner. Due to the fact that each wing was mounted singly on the wall, they each were given individual frames. Although Schinkel’s frames clearly covered the original frames it is not sure whether the quatrain on the outside wings could still be seen.

188

bettina von roenne

Fig. 12.8 Detail of the Schinkel frame of Goswin van der Weyden’s Madonna and Child with Donors

the frames by schinkel for the wings of the ghent altarpiece and the copies

There is no doubt that Schinkel’s frames are classical but they are historicized by the application of Gothic elements. Schinkel designed the frames in a way that was in his opinion Gothic in spirit. Due to the fact that it was totally new at the beginning of the nineteenth century to consider the period of origin of paintings in designing frames for them, Karl Friedrich Schinkel was breaking the mould in the field of picture framing. N OTES * From the Gemäldegalerie, Berlin, I would like to thank Stefan Kemperdick, Marion Kolanoski, Rainer Michaelis, Rainer Wendler and especially Christine Exler for their assistance and provision of detailed information. My special thanks go to Hélène Verougstraete for her constant support and useful exchanges in the field of framing. 1 In 1816 the wings were sold to the art dealer Nieuwenhuys. Edward Solly bought the wings in 1818 in Aachen and sold them in 1821 to the Prussian king as a part of a whole collection. In 1920 the wings were given to Belgium in compliance with the Treaty of Versailles, § 247. 2 Inventory numbers of the picture gallery in Berlin in 1830: cat. no. II., 1-12 (inner and outer sides of the panels each had own numbers): Waagen 1830, pp. 129-130. 3 For further research on Schinkel’s frames, see Schottmüller 1936, Sievers 1950, von Roenne 2007, von Roenne 2011 and von Roenne 2012. 4 The constantly repeated thesis that all the pictures exhibited in the gallery received new Schinkel frames before 1830, with the purpose of unifying the gallery’s interior, is inaccurate. The assertion goes back to von Bode 1929, pp. 7-45. Historical sources mention that due to the desire for diversity older frames were to be kept and restored if possible: Draft of a lost memorandum by Gustav Friedrich Waagen, published in Schottmüller 1936, f. 69; Sievers 1950, p. 76 and von Roenne 2011, pp. 21-22. 5 On Schinkel’s museum frames, see von Roenne 2007 and von Roenne 2011, pp. 13-67. 6 At that time von Humboldt was the head of the commission for the establishment of the museum and Waagen became the first director of the picture gallery in 1830. 7 Letter by Wilhelm von Humboldt to Gustav Friedrich Waagen, 25 July 1829: Geheimes Staatsarchiv Preußischer Kulturbesitz (GStA PK), I. HA, Rep. 137 II, Museumsbehörden und Kommissionen, D No. 3, p. 8 r. 8 The consecutive numbering of the pictures in each room, and in there at each wall, always started at the window, as Waagen explains in his catalogue preface (Waagen 1830, Preface XV, NB 3). 9 Waagen 1830, pp. 130-131. A copperplate on the wall described the original arrangement of the whole altarpiece (Waagen 1830, p. 130). 10 This arrangement can be seen in a painting of the gallery’s interior by Carl Bennewitz von Löfen of 1880-1884 (fig. 12.4). Although the wings were differently mounted on the wall and had new frames, the wings were still placed in the same order as in 1830. 11 Which side of the panels could be seen in which half of the week is not mentioned: Eduard Magnus, Premoria zu einem Gutachten über den geplanten Umbau der Galerie im Museum am Lustgarten, Berlin 1868, 23, cited in Michaelis-Vogtherr 2011, p. 241. 12 Waagen 1830, pp. 131-132.

189

13 Michiel Coxcie (1497-1592), Adoration of the Lamb and Deity Enthroned, copies of the Ghent Altarpiece by the Van Eycks, 1557-1559, Staatliche Museen zu Berlin (SMB). Gemäldegalerie, cat. no. 524 and 525, Waagen 1830, cat. no. II.13 and II.14. The whole altarpiece copy was originally exhibited in the chapel of the Royal Palace in Madrid in 1563. In 1808, the French General Belliard sent the altarpiece from Spain to Brussels. In 1820, Maximilian I Joseph, King of Bavaria, bought the copies of the Virgin Enthroned and John the Baptist Enthroned (since 1836 Munich, Alte Pinakothek), and the Adoration of the Lamb and the Deity Enthroned were sold to the Prussian king in 1823. The wings, except Adam and Eve, were owned by the Prince of Orange, and are now in the KMSKB-MRBAB, Brussels. The whereabouts of Adam and Eve has been unknown for a long time, and today they are probably in Saragossa. For further information on Coxcie’s copy of the altarpiece see Roberts-Jones 1992 and De Antonio 1998. 14 Carl Friedrich Schulz (1796-1866), Virgin Enthroned and John the Baptist, copies of the Ghent Altarpiece by the Van Eycks, 1826, oil on oak wood, 167 ≈ 74 cm and 170 ≈ 75 cm, SMB, Gemäldegalerie cat. no. 525 D and 525 E. Picture frames after Schlinkel’s design: wood, ornaments with lead and chalk grounds, gilded, inner measurements of the frame: 161 ≈ 69 cm, outer measurements: 183 ≈ 90.5 cm, width: 10.5 cm, R.I.No. (frame number) 93.1455 and 93.1454. For a long time the catalogue of the picture gallery in Berlin only included exhibited pictures, not works in storage. At what point the copies by Schulz were acquired by the gallery is not known. However, they had been exhibited for the first time in 1883 (Berlin 1883, p. 147). I thank Rainer Michaels (Berlin, Gemäldegalerie, 2012) for providing this reference. 15 The Schinkel frames were removed at that time, as mentioned in Schottmüller 1936, p. 82, NB 26. 16 Carl Bennewitz von Löfen, Interior of the Gemäldegalerie Berlin, 1881/1884, SMB. Nationalgalerie, cat. no. A III 759. 17 How the original frames looked after they were covered for a certain time is shown in an archival photograph of the Annunciation panels: Friedländer Archive, RKD, The Hague, also published in Anne van Grevenstein 2011. 18 Since the picture was almost totally destroyed by water, it is very likely that the same happened to the frame that was not restored, unlike the picture. 19 The Schinkel frames of the two Schulz copies are stored in the picture reserve in the Gemäldegalerie. The Schinkel frame of Coxcie’s copy of the Deity Enthroned is in the frame reserve (R.I.No. 93.1353), since the picture was given a new frame for its presentation in the reopened Bode-Museum in 2006. Coxcie’s copy of the Adoration of the Lamb received the same bordering at that time, a simple and patinated golden moulding, which is meant to be more or less neutral. 20 The edges of Schulz’s Virgin Enthroned and John the Baptist Enthroned are painted in a monochrome brown colour. 21 There are several examples of rectangular Schinkel frames on arched pictures or tondi in the Gemäldegalerie, see von Roenne 2007, pp. 41, 54-55, 59-61, 64-65, 69, 89, 97. 22 Karl Friedrich Schinkel, True to scale sketch for the picture frames of the Ghent Altarpiece wings, 1827-1830, pencil, 174.5 ≈ 51.7 cm, SMB. Kupferstichkabinett, SM 46.30. Only the ornaments on the plinths and the sill on the drawing differ from the left frames. 23 Goswijn van der Weyden, Madonna and Child with Donors, 1511-1515, SMB. Gemäldegalerie cat. no. 526 (exhibited in the BodeMuseum), Waagen 1830, p. 133, cat. no. II.15. 24 One example is the Golgotha with reliquary, Altarpiece, Erfurter Tafelmalerei, 1350, Predigerkirche zu Erfurt. Information kindly provided by Ute Stehr, Gemäldegalerie Berlin, 2007. 25 There is every indication that the museum frames by Schinkel were oil gilded. In terms of the value of the raw materials, oil gilding does not differ much from traditional leaf gilding, but the process is simpler and therefore cheaper and faster. These were two very important criteria for the museum frames of the Gemäldegalerie in 1830. Finally, lead ornaments are normally oil gilded.

190

bettina von roenne

26 Schottmüller 1936, f. 72, Sievers 1950, S. 84. 27 In the case of the Just Judges, the original frame survives, but the panel itself was stolen in 1934. The frames of the centre panels are completely lost. 28 Only the frames of Adam and Eve on the inside and the interior on the outside are still in their original shape. 29 Waagen 1824, p. 89. 30 Waagen 1824, pp. 99-100. Van der Velde describes the discovery of the quatrain in detail, see Van der Velden 2012, p. 9, N. B. 4. 31 Waagen mentions – besides the quatrain – the following inscriptions on the frames: On the lower frame of the Musician Angels Laudate cum in cordis organo (p. 93), the text on the frame of the Singing Angels, which he could hardly read (p. 94) and the inscriptions on the crossbars that divide the Virgin Annunciate and the Angel Annunciate from the prophets Micha und Zacharias (p. 98) Waagen 1824. 32 Waagen 1830, p. 132. 33 See n. 7. 34 Two photographs show the upper edge of the Angel Annunciate and the Virgin Annunciate wings: Accession numbers 0000215046 and 0000215047, Friedländer photoarchive, Rijksbureau voor Kunsthistorische Documentatie RKD / Netherlands Institute for Art History,

The Hague. Many thanks to Suzanne Laemers for calling my attention to these. 35 It is certain that the original frames have not been dismantled for a long time in the Gemäldegalerie in Berlin, otherwise they would not have survived. The conservation, documentation and careful storage of frames only began a century ago. 36 Waagen 1830, p. 123. 37 At the edge of the frames of the copies of John the Baptist and the Virgin Enthroned in the Gemäldegalerie there are nail holes at regular intervals. It is possible that a wooden panel once covered the junction of the frames on the two sides of the pictures. 38 The frames’ mouldings are always cut twice on both sides; above the base at the bottom and before the round arched part begins at the upper frame end. These cuts refer to the original frame construction and it is likely that the mouldings were cut at the same point when they were reused later on. 39 It cannot be seen from the frames today whether or not a crossbar had been previously attached, which would suggest the reuse of the Annunciation frames. 40 Siefert 2010, p. 250, cat. no. 90.

PART II STUDIES IN VAN EYCK PAINTINGS

Fig. 13.1 Jan van Eyck, Giovanni(?) Arnolfini and his Wife, detail, 1434, oil on panel, 82.2 x 60 cm London, National Gallery (NG186)

13

Van Eyck’s Technique and Materials: Historical Perspectives and Contemporary Context Marika Spring and Rachel Morrison

ABSTRACT: Research on painting technique at the time of Jan van Eyck has traditionally relied on documentary sources and practical experimentation to reproduce effects seen on paintings. Scientific analysis now contributes to this research, but a review of past literature demonstrates that interpretation of results can be challenging, particularly for paint binders, and is often strongly guided by prevailing beliefs derived from historical studies. The influence of lead pigments and ultramarine on chemical changes in paint on ageing has consequences for medium analysis, as do red lakes made from dyed wool. Van Eyck’s paint shows evidence of pre-polymerization of the oil, either by sun thickening or boiling. This process could include a lead siccative, and two additives perhaps also used as driers, or to modify other properties, have been confirmed – white copperas (zinc sulphate) and colourless powdered glass, both referred to by Eastlake in his remarkably accurate deductions about oil media in Van Eyck’s time.

—o— Introduction Admiration for the creativity and beauty of Early Netherlandish paintings, in particular those by Jan van Eyck, as well as the very evident skill that the artists displayed in handling paint in order to achieve the effects they desired, led almost inevitably from a very early date to interest in their materials and techniques. This continuous debate has usually centred around the nature and qualities of

the binding medium that they used, even before Vasari erroneously placed Van Eyck in the position of the inventor of oil painting. It has traditionally relied on study of historical documentary sources, sometimes combined with practical experimentation involving various mixtures of materials in attempts to reproduce the effects seen on the works themselves. This has been supplemented from the mid-twentieth century onwards by scientific analysis of samples from paintings, often with the most advanced methods available at a particular time, giving another more direct source of knowledge of the materials used by these artists that might be considered to have a more reliable basis. Interpretation of the results can still be challenging, however, particularly for identification of the paint binders. The conclusions are always to some extent subjective, since they are made within the bounds of knowledge at the time of the analyses, often strongly guided by prevailing beliefs derived from historical studies. In recent years, a greater awareness has developed of the influence of certain pigments on the chemical changes that occur on ageing and therefore on these analyses. This effect can often explain conflicting results obtained in the past and inform our reading of the published identifications of binding media in paintings by Van Eyck and his contemporaries, which

196

marika spring and rachel morrison

otherwise seem changeable and inconsistent. At the same time the more sophisticated analytical techniques that have developed since the pioneering study of the Ghent Altarpiece in the 1950s have improved our ability to investigate modifications and additions to the oil that would have influenced its working and drying properties. This article considers current understanding of each of these issues, based mainly on analyses and re-analyses of Early Netherlandish paintings in the National Gallery, including those by Jan van Eyck. This is preceded by a review of historical and earlier scientific studies, as background and context for the new reassessments and observations on the materials used by Van Eyck and his contemporaries. Van Eyck’s Medium: the Debate based on Documentary Sources and Reconstructions The earliest remarks that we have about Van Eyck as an oil painter are from the Italians Bartolomeo Fazio and Antonio Averlino Filarete,1 writing in the 1450s and 1460s, in praise of the skill of northern European artists in a technique not yet being generally used in Italy at that time, where egg tempera was still the predominant medium for panel paintings. Filarete names Jan van Eyck and Rogier van der Weyden as the greatest of the masters painting in oil. In the ‘question and answer’ format of his Trattato di Architettura he asks what oil is used, how one works with it, and whether it might be rather dark, with the answer that it is linseed oil, which can be clarified by leaving it to stand for a long time in a small dish. There is no strong sense that for him this medium is out of the ordinary, but he does seem to be concerned with how to work well with it, key considerations being that it should be light in colour, and how it could be prepared in order to achieve this.2 Giorgio Vasari’s well-known account crediting Van Eyck with the discovery of oil painting also focuses on preparation of the oil so that it dried well and was durable.3 This continued to be the main theme addressed by subsequent writers, in parallel with debates about Van Eyck’s place in the

history of oil painting, the latter turning to the idea of some special improvements that he might have made once it was appreciated that the oil medium was already being used well before his time. By the nineteenth century, writers and also artists were in addition motivated by a desire to understand why Early Netherlandish paintings had survived in such good condition, partly in the hope that their research might be put into practice by their own contemporaries, whose paintings were not proving so durable. There are several thorough modern reviews of the literature on this subject, starting with that by Ziloty in 1941, who divided the prevailing opinions into two groups.4 These form a convenient structure for the summary in Table 1, which also draws on the three more recent reviews by Pim Brinkman, Ashok Roy and Elise Effmann.5 The first group takes as its premise a medium based on oil, and centres around the ways in which oil can be modified to alter its working and drying properties, and its clarity or colour, either through the method of preparation or through different additives. This especially addressed the luminous translucent glazy passages in Van Eyck’s paintings, evident mainly in the deep greens and reds (figs 13.1, 13.2). The second group consists of those writers who discuss the idea of a mixture of media, either in separate layers or as emulsions of oil combined with proteinaceous materials of various kinds. These hypotheses began to emerge during the nineteenth century and came to predominate in the first half of the twentieth century and beyond. They were brought into the discussion because it was felt that oil alone, or an oil-resin mixture, did not have the necessary handling properties for the precision of detail that can be seen when looking closely at Jan van Eyck’s paintings, especially under magnification, particularly the individually delineated hairs in furs and the tiny yellow highlights on cloth of gold or depictions of metal objects, where a creamy bodied paint seems to have been used, applied in fine, delicate and crisp brushstrokes (fig. 13.3).

van eyck’s technique and materials

197

TABLE 1: Summary of hypotheses in the literature on Van Eyck’s medium Group 1: Oil and oleo-resinous media r r r r r

Group 2: mixed media

Sun-thickened oil Boiled (heat-bodied) oil, with or without siccatives Clarified oil Oil with the addition of resins (hard or soft) Oil (prepared in various ways, and with or without resin), modified with diluents

Charles Lock Eastlake, in the long discussion he published in 1847 in his book on the materials and history of oil painting, encompasses most of the ideas and variations expressed either before or after by the proponents of the first category – oil, but with various modifications and additions.6 His ideas were based on an extensive consideration of the documentary sources, an approach which earlier writers such as Mérimée, Lessing and Raspe had

Emulsions r Oil (and resin) + egg (yolk or white) r Oil (and resin) + gum arabic r Oil (and resin) + glue ‘Mixed technique’: aqueous binders r Egg tempera underpaint beneath oil glazes r Proteinaceous binder for ultramarine

also taken,7 but while their discussions were based on the manuscripts of Theophilus and Eraclius, Eastlake was the first to refer to two fifteenth-century manuscripts: the Strasburg manuscript and the so-called De Ketham manuscript.8 Eastlake became Keeper at the National Gallery in 1843, when Van Eyck’s Portrait of Giovanni(?) Arnolfini and his Wife was already part of the collection. His close study of this work, and his inferences about properties of

Fig. 13.2 Arnolfini Portrait, photomicrograph of the woman’s green dress, showing blotted translucent green glazes applied on more opaque green underpaint

198

marika spring and rachel morrison

Fig. 13.3 Arnolfini Portrait, photomicrograph of the fur at the bottom of the woman’s gown, showing the crisp brushstrokes of individual hairs

the paint from its appearance, informed his efforts to rationalize his observations on Early Netherlandish painting technique with what could be deduced from the recipes and instructions in the documentary sources.9 He believed that the essential properties of Van Eyck’s medium must be that it was drying, nearly colourless and ‘of a consistence (though no doubt varied in this respect as occasion required) which allowed of the most delicate execution’. He concluded that the components were linseed or nut oil and ‘resinous ingredients of a durable kind’, by which he meant hard resins such as amber or copal.10 Mérimée, among others, also considered that the oil was combined with a hard resin,11 and it was only later that some writers started to suggest that a soft resin might have been used.12 Eastlake devoted a considerable part of his book to the methods for preparing or clarifying oil

described in the documentary sources, including recipes for an oil with greater body and that would dry more quickly, many of which involved boiling the oil with various driers and other substances. He concluded that the pigments were first ground with oil that had been clarified but not bodied or thickened, and that thickened oil, which might also contain resin, was then incorporated to a greater or lesser extent depending on the colour of the paint. Additional drier could be added on the grinding slab if required. Eastlake considered Van Eyck’s primary aim to be to improve the siccative properties of the oil, and stated that the ‘chief drier which he used in preparing the oils or varnishes appears to have been white copperas’, on the basis that it is included in recipes in both the Strasburg and De Ketham manuscripts.13 By ‘white copperas’ Eastlake meant zinc sulphate (ZnSO4·7H2O), also called white vitriol or zinc vitriol, his transcription of the

van eyck’s technique and materials

unambiguous term galicen stein used in the manuscripts.14 Gotthold Ephraim Lessing had already mentioned vitriol in 1774, as an example of a material that Van Eyck could have used to accelerate drying, although this seems to have been based simply on knowledge of the siccatives that were in use at the end of the eighteenth century, and not on observations made from historic documentary sources.15 According to Eastlake, the recipe entitled oleum preciosum in the Strasburg manuscript ‘comprehends the chief improvements’ of the ‘Flemish system of oil painting’. The manuscript states that ‘all painters are not acquainted with it’, which Eastlake took to infer that it was ‘not older than Van Eyck’s time’.16 The recipe instructed that the oil should first be boiled with calcined bones and an equal quantity of pumice stone, removing the scum. White copperas was to be added once the oil had cooled, which the recipe states will make the oil ‘limpid and clear’. The oil is then strained and placed in the sun for four days, giving it a thick consistency and making it as ‘transparent as a fine crystal’. The manuscript states that ‘this oil dries very fast, and makes all colours beautifully clear, and glossy bedsides’. The colours should be well ground in the oil and a few drops of varnish should be added. A small quantity of calcined bone or a little white copperas ‘about the size of a bean’ could be mixed ‘in order to make the colour dry readily and well’.17 Driers could therefore be introduced on the grinding slab, and not only during preparation of the oil. Eastlake appreciated that to achieve some effects a thickened oil might not be suitable, and mentioned that it could be thinned by adding some of the non-thickened oil, but some other writers such as Mérimée considered instead that essential oils must have been used as a diluent.18 Others did not believe that this practice would have been known by Van Eyck and his contemporaries: Arthur P. Laurie, for example, argued that although essential oils were mentioned in treatises, it was never in conjunction with recipes relating to painting practice and they were therefore probably not used for this purpose at this time.19

199

Early in the nineteenth century a few writers had suggested that Van Eyck applied thin oil-based glazes on an underpaint bound in egg tempera or glue,20 but it was Ernst Berger in his book of 1897 who fully developed the hypothesis of a mixed technique. He argued that Early Netherlandish painters used a medium consisting of an emulsion, based mainly on his practical experiments with a mixture of egg yolk and oil or of gum and oil, but also referring to certain documentary sources which he considered supported his argument.21 Laurie was in agreement in his book published in 1910, The Materials of the Painter’s Craft, citing the good results in terms of handling properties that he had observed in his experiments with an emulsion of Canada balsam and egg white.22 These conclusions were modified in his 1926 book, where he stated that a little egg yolk added to a sticky oil or varnish gave a very crisp medium, but his ideas differed from Berger’s in that the main component of this emulsion was oil and it could not be diluted with water. Laurie also reviewed the interpretation of what he considered to be the only three references in historical sources that might be interpreted as emulsions, and concluded that Berger’s theories rested on ‘slender evidence’, set against the many more that existed for preparation of oils that quite clearly did not involve emulsions. In an article in 1934 he described yet more experiments, which convinced him that Van Eyck’s medium could have been stand oil alone, with no additions.23 Despite the doubts of Laurie and others, the theory of a ‘mixed technique’ was taken up by Max Doerner in the 1920s and 1930s, and the ideas he expressed in his publications proved to be very enduring.24 Many of the documentary sources pay particular attention to the medium for blue pigments, with the concern that they might be affected to a greater extent than other colours by yellowing of the binder over time, acknowledged by Eastlake with the suggestion that the oil-resin medium, or the varnish added on the palette, should be a clear or so-called ‘white’ varnish.25 This problem is also behind the

200

marika spring and rachel morrison

idea that a completely different binding medium might have been used for blue paint, and there is some evidence that this was indeed the practice in other branches of the arts such as polychrome sculpture, and occasionally in paintings.26 In the context of Netherlandish painting, the conclusion that an aqueous binding medium had been used for ultramarine reported in the publication of the examination of the Ghent Altarpiece in the early 1950s has had considerable influence on later studies.27 Van Eyck’s Medium: the Debate based on Scientific Analyses The Ghent Altarpiece With the study of the Ghent Altarpiece during its conservation treatment in the 1950s, scientific analyses of paint samples were introduced to the debate. Using a sequence of microchemical and staining tests which had been developed for analysis of the Altarpiece of the Last Supper by Dieric Bouts,28 Coremans and Thissen reported that the medium of the Van Eycks was generally based on drying oil, plus variable quantities of an unidentified component that they called ‘x’, which they considered had the character of a natural resin.29 The exception was the uppermost layers in the Virgin’s cloak, as well as a few other blue areas, all of which contained ultramarine. These seemed to behave differently when subjected to the tests, and were therefore thought to have an aqueous binder. In particular, it was noted that the fluorescence of the upper ultramarine-containing layers under ultraviolet light was different to that of the azuritecontaining underlayers, taken as an indication that they had a different medium.30 The authors considered this choice by the artist to be associated specifically with this pigment, and made two apparently contradictory speculations about the reasoning behind it: they suggest it would have given a paint with greater covering power, allowing thinner applications of this expensive pigment, but also propose that in addition to remaining lighter as it aged than oil paint, it would be more transparent. As was generally believed at that time, the

dark green glazy translucent paint was considered to consist of ‘copper resinate’, with or without oil. Coremans and Thissen are careful to lay out the degree of certainty of their methodology. They state that they felt that with these tests they were able to differentiate between the two main groups of tempera and oil satisfactorily. Distinction between a result of oil alone, or oil plus ‘x’, was based mainly on a difference in behaviour towards potassium hydroxide. Although it is not clear whether they used it on the blue paint for which they postulated an aqueous (i.e. proteinaceous) binder, they also had available to them the Emich test for nitrogen, which was their principal means of confirming a protein identification, giving negative results for oil and positive results for tempera and animal glue, to which they added the qualification ‘under good circumstances’.31 It is worth noting, however, that this uncertainty in interpretation was considered by others at that time to make the Emich test too unreliable to be useful.32 Coremans and Thissen did not ignore the possibility of emulsions, but after discussing their properties, they stated that so far as they could conclude from the analytical methods at their disposal, the characteristics of the paint layers did not show any evidence for this type of medium. Over the years further tests have been carried out, and this view has evolved as samples from the Ghent Altarpiece have been re-examined with new techniques (Table 2). In his publication of 1973/74, Kockaert was the first to make the important observation that translucent white globules are nearly always visible under a microscope in the small raised strokes of paint based on lead-tin yellow found on Early Netherlandish painting highlighting cloth of gold or gold objects, hypothesizing that these were the proteinaceous component of a mixed medium. He investigated samples from around ten Early Netherlandish paintings, including the central panel of the Ghent Altarpiece, concluding mainly on the evidence of a positive result in the white globules with a stain specific for proteins that the

201

van eyck’s technique and materials

TABLE 2: Summary of binding medium analyses of the GHENT ALTARPIECE Coremans and Thissen 1953

Kockaert 1973/4

Kockaert and Verrier 1978/9

Kockaert, Gausset and Dubi-Rucquoy 1989 Masschelein-Kleiner 2005

Drying oil and variable quantities of ‘x’, which ‘seems to have the quality of a natural resin’ r Final deep blue ultramarine ‘glaze’ on Virgin’s drapery in an aqueous medium r Lead-tin yellow highlights; gelatin/oil emulsion

Microchemical tests and staining tests

r Light and opaque colours in an emulsion, probably oil/egg except in lead-tin yellow highlights where oil/gelatin r Probably egg white binder for blues r Oil or oil-resin media elsewhere r Final deep blue ultramarine ‘glaze’ on Virgin’s drapery: egg white binder

Staining tests specific for proteins, occasional microchemical tests

r Re-interpretation: lead-tin yellow highlights do not have an emulsion as binder

 

r

paint was composed of an emulsion of gelatin and oil.33 Further re-examination in the late 1970s studied a broader range of samples held at the kik-irpa – thirty-eight in total – from the Ghent Altarpiece, the Virgin with Canon van der Paele and the Portrait of Margaret van Eyck (both Bruges, Groeningemuseum). Kockaert and Verrier concluded that while certain brown and ochre colours seemed to be oil alone, and the green ‘copper resinate’ glazes did not stain for protein, the medium of the light opaque areas was indeed an emulsion. They stated that the protein component generally seemed to be finely dispersed, which suggested to them that it was more likely to be egg than glue; in the lead-tin yellow highlights, however, since there were larger translucent globules that stained for protein, they again reported an oil-gelatin mixture as in the 1973/74 study. They also concluded that certain ultramarine blues on both the Ghent Altarpiece and on the Virgin with Canon van der Paele had an aqueous binder that was probably egg white. In the late 1980s, more specific immunofluorescent staining directed at the detection of ovalbumin seemed to support this conclusion. At around the same time, another refinement was made in the

Staining tests specific for proteins, occasional microchemical tests

Immunofluorescent staining for ovalbumin protein (egg white)

analysis of Van Eyck’s medium, which was that colophony (pine resin) was identified as an addition to oil in the green glazes, reported by Brinkman et al.34 Kockaert’s scientific methodology was sound, but the white inclusions in lead-tin yellow proved to be prone to false positive results with the staining tests that he used. New research in the early 2000s was able to show, with ftir microscopy, that the white inclusions in lead-tin yellow-containing oil paint are in fact agglomerates of lead soaps formed by reaction of the pigment with fatty acids in the oil binder.35 Lillian MasscheleinKleiner’s short note published in 2005 did not present new analyses but reassessed the earlier work on this basis, stating that it could be assumed that the translucent globules described by Kockaert in Van Eyck’s paint must also be lead soaps and that the medium was not therefore an emulsion.36 Binding Medium Analysis of Netherlandish Paintings in the National Gallery, London With the introduction of gas-chromatography– mass-spectrometry (gc-ms) to the National Gallery laboratory at the very end of the 1970s it

202

marika spring and rachel morrison

became possible to gain some information on the preparation of the oil: before this, analyses with gc only could only specify the type.37 gc-ms and Fourier transform infrared micro-spectroscopy (ftir) are now the methods used most routinely for analysis of paint binders in the National Gallery. The medium of forty-five of the fifty-two paintings in Lorne Campbell’s catalogue of the fifteenth-century Netherlandish School paintings in the National Gallery was analysed by these methods in the 1990s, giving a good overview of artists’ practice in general at the time, even if this included only a few analyses of samples from works by Van Eyck.38 Linseed oil was identified in all of the forty-four oil paintings (the only other painting analysed is a Tüchlein). In nine of these, walnut oil was also found, generally in lighter colours or for cool colours such as pale blues.39 In many cases there was some evidence that the oil had been pre-polymerized or heat-bodied, in keeping with the instructions for preparation of the medium in documentary sources, and in twelve of the paintings a little pine resin was reported as an addition to the oil in translucent red and green glazes.40 It was only possible for a few paintings to take a range of samples for medium analysis from a variety of colours, and these numbers cannot therefore be taken to be statistically significant, but nevertheless this body of results does demonstrate that artists in this period were choosing to use different types of oil prepared in various ways for different areas of the same painting. Only two samples were taken for medium analysis from Van Eyck’s Arnolfini Portrait, from the red tester of the bed and the woman’s green dress, both of which proved to contain linseed oil; the more translucent glazy paint of the latter seems to have been heat-bodied and to include a little pine resin. Linseed oil was also identified in the one sample from his Portrait of a Man (Self Portrait?) (ng222), from the red glaze on the turban, again with a little pine resin, while it was only possible to sample the paint of the marbling on the back of Portrait of a Man (‘Léal Souvenir’) (ng290), where heatbodied linseed oil had been used. No walnut oil was

found, but no light areas where it might be expected were sampled, nor was it possible to analyse the binder in blue paint, which would have been of interest considering the results reported for the Ghent Altarpiece. The results from the two paintings ascribed to followers were similar; in Marco Barberigo (ng696) the translucent purple paint of the cornette contains linseed oil and a little pine resin, and in A Young Man holding a Ring (ng2602) heat-bodied linseed oil was used for his black cloak, while the linseed oil in the lighter area of the background was not heat-bodied. For the marbling on the back of this small portrait a combination of egg tempera and oil, in separate layers, was reported. This last work is among the twelve paintings in Campbell’s 1998 catalogue that are described as having a medium of oil but with some egg tempera. These include three paintings in the Campin group, three in the Van der Weyden group and four in the Bouts group,41 in which egg is thought to have been used in underlayers in certain areas. These results seem, however, to depart from the practice of the majority of the paintings and it is worth reconsidering their reliability, taking into account current knowledge of the limitations of the analytical techniques that were used, in particular for paint containing certain pigments. gc-ms analysis, for example, relies on identification of a characteristic pattern of derivatives of fatty acids, but this is influenced by the chemical changes that take place in the paint film on ageing and drying. The interpretations in Campbell’s 1998 catalogue were made in the context of understanding at the time, but in recent years a greater awareness of the extent to which some pigments affect this process has developed. This can sometimes disturb the expected pattern of fatty acids, giving gc-ms results that are ambiguous or that can be misinterpreted. Other methods such as staining tests and ftir can also be affected by the pigment content of the paint. The reassessment of these results, presented below, has included not only new interpretation of some of the earlier analyses, but reanalysis of the

van eyck’s technique and materials

existing samples. An important aspect of this new work has been the use of attenuated total reflectance Fourier transform infrared micro-spectroscopic imaging (atr-ftir imaging),42 introduced to the laboratory in 2011, to analyse organic materials directly on some of the cross sections; this was of considerable benefit in making progress towards understanding inconsistencies in the earlier gc-ms results. Where suitable samples were available, this technique was used for the paintings mentioned above where the earlier analyses suggested a mixture of media, and also on samples from the National Gallery paintings by Van Eyck. Nevertheless, this reanalysis has built on the earlier work by Ashok Roy, Raymond White and Jennie Pilc published in the postprints of the Van Eyck symposium at the National Gallery in 1998, and in Lorne Campbell’s catalogue published in the same year.43 At the same time, energy dispersive X-ray analysis in the scanning electron microscope (sem-edx) has led to the discovery of certain paint additives that had not been found during the earlier examinations, and which could have had a role in modifying the properties of Van Eyck’s medium. In addition, it was possible to draw on analyses of samples from Van Eyck’s Portrait of Margaret van Eyck, examined while it was undergoing conservation at the National Gallery, and on analyses from a collaboration with the National Gallery of Art, Washington, on re-examination of cross sections from their Annunciation. These works are each the subject of other papers in this volume, so the scientific examination will not be described in detail here, but certain results are particularly relevant to the discussion below and will be mentioned. The Influence of Pigments on Binding Medium Analysis Lead-based Pigments The precise strokes and raised texture of the paint based on lead-tin yellow used to depict the delicate highlights on golden objects or gold thread in cloth of gold brocades and decorated textiles in Early

203

Netherlandish painting give the impression of a thick buttery bodied paint. This can be seen in the Arnolfini Portrait in the fine criss-crossing pattern of the gold embroidery on the woman’s belt (fig. 13.4), the tiny beads of her necklace (fig. 13.5) and in the chandelier hanging in the background behind the figures. As already outlined above, the idea that these effects were only made possible by modifications to the binding medium, in particular the use of an emulsion, had arisen from the experiments of Berger, Laurie, Doerner and others, and was given further credence by Kockaert’s scientific analysis in the 1970s. His significant observation that white inclusions are commonly seen in leadtin yellow paint has proved to hold true – they are just about visible with a microscope in the beads of the woman’s necklace in the Arnolfini Portrait – but the staining tests on which he based his theory that they are globules of protein in an emulsion have now been shown to have been unreliable. With the advent of ftir microscopy they were identified instead as agglomerations of lead soaps, formed by reaction of the pigment with fatty acids in the oil binding medium. The possibility of false positive results with staining tests is now widely understood, but an even more fundamental problem is that it seems that for oil paints based on lead-tin yellow or red lead, the fatty acid ratios determined by gc-ms can often differ from that normally expected for an oil that has aged. The palmitate to stearate ratio indicates the type of oil – for example, linseed or walnut. The azelate to palmitate ratio has been used to distinguish between oil and egg tempera. Azelaic acid forms in oils on ageing due to degradation of unsaturated fatty acids, so if the paint contains a drying oil it is present at relatively high levels. Egg does not contain a high proportion of unsaturated fatty acids, and rather little azelate forms on ageing. For oil, an azelate to palmitate ratio greater than 1 is expected, while for egg tempera this ratio is typically around 0.1 or 0.2. In paint containing lead-tin yellow or red lead, even where all other observations indicate an oil

204

marika spring and rachel morrison

Fig. 13.4 Arnolfini Portrait, photomicrograph of the woman’s belt

Fig. 13.5 Arnolfini Portrait, photomicrograph showing the woman’s necklace

van eyck’s technique and materials

binding medium, low azelate levels have often been found, an effect that may well relate to the influence that formation of lead soaps has on the chemistry of the oil.44 To illustrate this point with an example distant from Early Netherlandish paintings, in a large altarpiece by Moretto da Brescia from around 1540-1545 (London, National Gallery, ng625), an azelate to palmitate ratio of 0.1 was found in a sample from the expanse of pale yellow paint in the mandorla behind the Madonna, even though it is clear from the paint texture of the broad brushstrokes that it is in oil, and with ftir analysis there was no indication of the protein content that would be expected with egg. gc-ms is not, however, directly identifying a component that is unique to egg, and instead conclusions are based on what is considered to be a typical pattern of fatty acids. While in the painting by Moretto the low azelate level was simply considered to be an anomaly, in the context of Early Netherlandish painting and the firmly established mixed medium theories, this phenomenon has led to misinterpretation of the binder as egg. A similar effect occurs for paint containing red lead. Raymond White’s paper in the postprints of the 1998 Van Eyck symposium, includes analyses of samples from the wall paintings from St Stephen’s Chapel, Westminster Palace, as an example of oil painting in England in an earlier era – indeed these paintings are discussed by Eastlake in relation to the entries in the Westminster Abbey accounts of the purchase of oil for this purpose.45 Although the upper paint layers were confirmed as having an oil binder, the medium of the first priming layer, consisting of red lead, was reported as egg tempera, based on the low azelate levels seen by gc-ms.46 Again, large translucent white lead soap inclusions are visible in this layer in a paint cross section, and when re-analysed by ftir there was no indication of protein, making it clear that the earlier result needs to be reinterpreted and that the binder is in fact oil.47 An example where at first sight it was less obvious that this phenomenon had affected the gc-ms analyses is in Rogier van der Weyden’s Lamentation

205

(ng6265), one of the twelve paintings listed as oil with some egg tempera in the National Gallery Early Netherlandish school catalogue. An egg binder was reported in the underpaint beneath St Jerome’s drapery, again on the basis of a low azelate to palmitate ratio. This orange-red underpaint appeared to be based on vermilion (mercury sulphide), but further investigation with sem-edx has shown that although this pigment is present, there is also a significant amount of lead. atr-ftir imaging indicates that this is now present in the form of lead soaps, and it seems that a lead pigment in the original mixture has almost completely reacted with fatty acids in the medium. This is most likely to have been red lead, quite often found in combination with vermilion not only in paintings of this period but also later. Red lead might have been included as a drier mixed into the paint on the grinding slab – as recommended in the sixteenthcentury south Netherlandish recipe book now in the Plantin-Moretus Museum48 – or might have been introduced without the artist’s knowledge, through an adulterated grade of vermilion.49 Despite the earlier gc-ms results, there was no real evidence from the ftir results of a proteinaceous binder, as would be expected if it was egg. Again our greater knowledge of the influence of lead pigments on the analyses provides an explanation of the earlier misidentification and allows us to reinterpret the results. Natural Ultramarine Another pigment that has been found to influence the chemistry of ageing of oil media is ultramarine, whether natural or synthetic, and again great care needs to be taken in interpreting the results of binding medium analysis where it is present in the paint. The situation is made more complex by the fact that painters were well aware that blue paint could be especially affected by yellowing of the oil, and historical documentary sources therefore often discuss the use of a less yellowing oil such as walnut, or even a different medium such as gum or glue.50 As with lead pigments, gc-ms analyses

206

marika spring and rachel morrison

over many years at the National Gallery have demonstrated that ultramarine-containing oil paint often shows depressed azelate levels. This is well illustrated by The Adoration of the Shepherds by Guido Reni (ng6270), the largest work in the National Gallery at 480 centimetres tall, where it is evident from the broad brushstrokes and the thick application of the ultramarine blue paint in the sky that there is no question that it is in oil. gc-ms analysis, however, gave an azelate to palmitate ratio of only 0.5, considerably lower than the ratio of about 1 or more that would be expected for a drying oil. It is also likely to be significant that the sky paint is blanched and broken up. This must be related to the pigment since it has not occurred in other colours. The mechanisms involved in the degradation of ultramarine-containing paint are not yet fully understood.51 It has long been known that the lazurite mineral that gives the pigment its colour is sensitive to acids, which cause loss of colour through release of the chromophore (S3-) and disruption of the aluminosilicate framework.52 When deteriorated ultramarine-containing oil paint has been examined in paint cross sections, however, it often seems that much of the pigment is still blue, but the binding medium appears cloudy and broken up, suggesting that more complicated chemical processes are at work than simply the action of acids on the pigment.53 As with the lead pigments, there is some indication of interaction between the pigment and the oil, but there is still much to be discovered about the chemical processes that are occurring. Another indication that breakdown of the binder might be involved in the degradation is that it has sometimes been reported that ultramarine glazes seem to be sensitive to cleaning agents,54 as described in the paper on the Washington Annunciation by Gifford et al. in this volume.55 It seems very probable that this deterioration could have an influence on binding medium analysis, both gc-ms and also the solubility and staining tests that were used to examine the medium of the blue paint in the Ghent Altarpiece, and caution is

therefore needed when assessing analytical results that suggest that ultramarine paint might be in a different medium to the rest of the painting. Jan van Eyck’s paintings seem to have been particularly susceptible to this type of degradation. One striking example is in the blue velvet pattern on the cloth of gold cope in his Virgin and Child with Canon van der Paele. The same can be seen in the Arnolfini Portrait, in the sleeves of the woman’s blue underdress, so that what appears to be a pattern in the cloth, perhaps a damask, is now barely visible (fig. 13.1). In the man’s purple robe, painted with ultramarine and red lake, this degradation has had the effect of reversing the tonal values of the modelling, as the half-shadows appear lighter than they would have been originally (fig. 13.6). In the photomicrograph of the unmounted paint sample, from a half-shadow, the uppermost layer appears broken up and deteriorated (fig. 13.7). This has happened wherever ultramarine has been used on this painting, including in small details such as the blue lozenges on the frame around the mirror on the back wall, the blue cloak of the man seen in reflection in the mirror and the small ultramarine dots on the man’s hat that indicate the texture of the material. The deterioration has not only affected the blue and purple draperies, but also the red draperies, due to Van Eyck’s consistent practice of mixing a small amount of ultramarine with red lake and some black in the deepest shadows. Blanching confined to the darkest areas of the red draperies can be seen, for example, in the bed hangings on the Arnolfini Portrait, the red headdress in the Portrait of a Man (Self Portrait?) (fig. 13.8), the red coat in the Portrait of a Man (‘Léal Souvenir’) and the red dress in the Portrait of Margaret van Eyck.56 The ultramarine used by Van Eyck in the Arnolfini Portrait is of notably small particle size. An estimate made from cross sections under a microscope with a calibrated graticule eyepiece, although not a rigorous method of measurement, indicated that the largest ultramarine particles in this painting are only 8 μm across. This can be compared to Rogier van der Weyden’s Exhumation

van eyck’s technique and materials

Fig. 13.6 Arnolfini Portrait, detail showing the man’s robe

207

208

marika spring and rachel morrison

Fig. 13.7 Arnolfini Portrait, unmounted paint sample from the man’s robe

of St Hubert (ng783), where the ultramarine is coarser and the particles measure 10-20 μm. The evidence gained from the National Gallery cataloguing programme suggested a preference for finely ground ultramarine not only in the Arnolfini Portrait but also more generally in works by Van Eyck. This would have given his paint better handling properties, as coarse ultramarine does not give a very manageable paint and would be difficult to apply as thin glazes, but might be implicated in the poor state of the blues in quite a number of his paintings, since the larger surface area of more finely-ground pigment would be expected to make it more susceptible to interaction with either the binder or other environmental factors.57 The possibility that this aspect of technique was the conscious choice of an artist, or part of the practice of certain workshops, was explored already in 1974 by Van Asperen de Boer, in his investigation of the particle size distribution of ultramarine in eight Early Netherlandish paintings. Although not comprehensive due to the small number of works included in the study, the results do seem to indicate interesting differences between artists. The ultramarine in Van Eyck’s Ghent Altarpiece was the most finely ground of all those measured, with a mean particle size of only 4.9 μm, supporting the observations made on the National Gallery paintings.58 The Virgin and Child by Dieric Bouts is another case where the gc-ms results of paint from the

Fig. 13.8 Jan van Eyck, Portrait of a Man (Self Portrait?), 1433, oil on panel, 26 x 19 cm, London, National Gallery (NG222), detail of the man’s headdress, showing blanching in the deepest shadows

ultramarine blue of the Virgin’s cloak showed a relatively low azelate level, intermediate to that expected for egg and for oil. When first published in the National Gallery Technical Bulletin in 1985,59 this was taken as an indication of a paint binder of oil but with the addition of some egg, a conclusion that was altered slightly in a more extensive article in the same journal in 1986,60 where the suggestion was made that the egg component was in the underlayers. This was on the basis of staining tests on a cross section, which seemed to give a weak indication of protein, although the authors did include the proviso that this result should be treated with caution. This interpretation was published more or less unchanged in Lorne Campbell’s 1998 catalogue.61 Re-examination of another of the cross sections from the cloak gave more reasons to be cautious about this result. Overall the cloak initially gives the impression that it is well preserved because of its intense deep blue colour, which is due to the high proportion of the blue mineral lazurite in the ultramarine pigment relative to the colourless associated minerals that are always present. It is, however, rather flat in appearance, with little distinction between light and shade in the modelling, suggesting that there has in fact been some degradation, particularly in the shadows. In the cross

van eyck’s technique and materials

section the thick ultramarine blue paint appears quite saturated at the far left of the sample, but patchy and slightly lighter in the middle (fig. 13.9a). The uniformly grey appearance of this ultramarine layer in the backscattered sem image (fig. 13.9d) shows that this lighter appearance under normal light is not due to the presence of lead white and must instead be an indication of deterioration that is causing scattering in the paint. The sem image also shows holes in this layer, making the tentative interpretation of the staining tests for protein in 1986 seem prudent, since the stain could have been drawn into this porous surface. An additional factor to be borne in mind is that the minerals that make up ultramarine pigment have active and hygroscopic surfaces that could very well result in adsorption of a stain. Either of these might be responsible for a false positive result. Coremans stated that his conclusion that the binding medium of the top layer of ultramarine in some areas of blue in the Ghent Altarpiece was different to that of the lower layers was supported by a difference in fluorescence under ultraviolet light. The upper blue layers in the cross section from the painting by Bouts are also fluorescent (fig. 13.9), but this has now been observed many times in National Gallery paintings, and is common for ultramarine-containing paint. It may well be a consequence of some kind of interaction between the pigment and binder, but does not necessarily indicate that a different binding medium was used. Another observation that has been made over many years of analyses of ultramarine-containing paint at the National Gallery is that calcium oxalate is very regularly detected in it.62 This is not likely to be an original component of the paint, so again can be taken as evidence of some kind of degradation or alteration. It can be detected in cross sections using atr-ftir imaging. Figure 13.9 includes an image prepared from the data collected from an area 64 ≈ 128μm in size on the cross section from the painting by Bouts, showing the distribution of an absorption band characteristic of calcium oxalate. It is evident that it is concentrated in the

209

matrix between the ultramarine particles, and that there is more of it towards the top of the layer, near the paint surface. The reason that this is associated with ultramarine is not yet fully understood, but it can be postulated that the physical breakdown of the medium in the paint that has been observed in the form of blanching may correlate with chemical breakdown that could generate small molecules such as oxalic acid that then react with sources of calcium in the film (including in blue lazurite and many of the associated minerals found in ultramarine), and it may turn out to be significant in understanding the deterioration process of this pigment. Returning to the Arnolfini Portrait and the sample from Arnolfini’s robe, the cross section shows that it was painted with an initial deep plumcoloured paint containing large red lake particles mixed with only a little ultramarine (fig. 13.10). The very thin layer at the surface forms the halfshadow of the robe, and as in the painting by Bouts, despite its light appearance it does not contain any lead white but consists only of ultramarine, as is evident from its uniformly grey appearance in the backscattered sem image: instead it is scattering light because it is deteriorated. atr-ftir imaging again showed that calcium oxalate was present – in fact rather more than was found in the Bouts sample (fig. 13.10). A spectrum extracted from the data from one of the areas rich in calcium oxalate confirms this identification and shows the characteristic infrared absorption bands of this compound at 1620 cm-1 and 1320 cm-1. It is also worth noting that the ftir spectra in both cases showed at least some evidence in the ftir spectrum of an oil binding medium for this upper layer. Rather similar results were obtained from the blue paint of the Virgin’s robe in Van Eyck’s Annunciation, described in detail in the paper in this volume by Gifford et al. Red Lake Pigments The red lake pigment of large particle size in the thickest layer of paint in the cross section from Arnolfini’s purple cloak can be more easily seen

210

marika spring and rachel morrison

c

a

100 μm

d

b

Figs 13.9a-d Dirk Bouts, The Virgin and Child, c.1465, oil on panel, 37.1 x 27.6 cm, London, National Gallery (NG2595), cross section from the Virgin’s blue drapery: (a) image under visible light, with the area analysed by ATR-FTIR imaging in a white box and SEM-EDX in a red box; (b) image under ultraviolet light; (c) ATR-FTIR image produced by integration of the band between 1328-1309 cm-1 showing the distribution of calcium oxalate; (d) back-scattered electron SEM image

under ultraviolet light, as it shows an orange-pink fluorescence that suggests it contains madder dyestuff.63 The back-scattered sem image reveals the irregular shape and soft edges of the particles, and the edx spectrum shows that they have a significant sulphur content, with rather little of the aluminium that would be expected if the pigment had a conventional alumina substrate (fig. 13.11). These observations, and the large particle size, are characteristic of red lake pigments prepared with a recipe that uses wool shearings as the source of the dyestuff, and where the alkali used to extract the dyestuff into solution is so strong that it has dissolved the wool, so that it has been incorporated into the pigment. Numerous recipes of this type exist: a typical one appears in a late fifteenth-century southern Netherlandish manuscript, which specifies ‘red shearings that are from good cloth of the best wool’. Wool of the highest quality could have been dyed

with kermes, but it would more often be a source of madder dyestuff, as seems to be the case in the red lake used in the purple mixture here. The recipe – entitled ‘to make synoper’ – begins with preparation of the alkaline solution by soaking ashes and lime in water. The strength of the alkali is tested with a feather and is strong enough if the feather on the quill becomes soft and sticky. This would have been capable of dissolving the wool fibres, as becomes clear in the next stage, where the wool shearings are boiled in the alkali until, as it says, ‘no wool nor hair can be seen’. The cooled solution was then filtered through a pointed bag, keeping the colour that remained in the bag, which would have been a kind of red ‘jelly’ of dissolved wool. At the last stage a solution of alum was added to precipitate the pigment.64 atr-ftir imaging of the cross section from the purple cloak shows that the red lake particles have

211

van eyck’s technique and materials

a

20 μm

b

c

1621

0.25

1318

Absorbance Units

0.20 0.15 0.10 0.05 0.00 2000 d

1800

1600

1400

Wavenumber cm-1

1200

1000 e

Figs 13.10a-e Arnolfini Portrait, cross section from Arnolfini’s robe: (a) ATR-FTIR image produced by integration of the band between 1045-899 cm-1 showing the silicate distribution; (b) ATR-FTIR image produced by integration of the band between 1335-1307 cm-1 showing the distribution of calcium oxalate; (c) visible light image; (d) sodium EDX map showing the distribution of ultramarine; (e) extracted FTIR spectrum from an area high in calcium oxalate (in red) with a calcium oxalate reference spectrum (in black)

a considerable protein content (fig. 13.11), confirming that the pigment was prepared in this way, with wool (a proteinaceous fibre) as the source of the dyestuff. This type of lake is very common, and although it had already been reported in works by his contemporaries, it is now clear that it can also be found in Van Eyck’s paintings. Recent analysis has identified it (again by atr-ftir imaging) in samples from the Portrait of Margaret van Eyck and The Annunciation. It has been found in several National Gallery paintings by Van der Weyden, and broader studies have indicated that generally all of the madder lakes in Netherlandish and German paintings of this period that have been analysed showed evidence of having been prepared from wool shearings, so that this was probably the usual method of manufacture. There are also one

or two examples of kermes lakes of this type, although they are rather more rare, probably because the more expensive kermes dyestuff was more commonly used to dye silk rather than wool.65 Interestingly, references to red lake pigments in the accounts of the Burgundian court include a record of the purchase by the painter Jean Malouel (d. 1415) of bourre de fine esclarte vermeille de Bruxelles pout fiare cynopple. Bourre was a term used for shearings from woollen cloth, while the term esclarte indicates that it would have been of high enough quality to be dyed with kermes, although madder is also possible. This constitutes the only reference that is known to date of purchase of shearings by a painter specifically for making red lake pigment – cynopple.66 It seems that by Van Eyck’s time this was a standard source of dyestuff for

212

marika spring and rachel morrison

20 μm

b

c

1646

a

1537

S

Absorbance Units

Al

Si K Na

1

Ca K

Cl Cl 2

3

0.15

0.10

0.05 Ca 4

5

1800

e

1600

1400 Wavenumber cm-1

1200

1000

f

d

Figs 13.11a-f Arnolfini Portrait, cross section from Arnolfini’s robe: (a) ATR-FTIR image produced by integration of the amide II band between 1576-1503 cm-1; (b) normal light image; (c) ultraviolet light image; (d) back-scattered electron SEM image; (e) EDX spectrum from a large lake particle; (f) extracted FTIR spectrum from a large lake particle

lake pigment making: perhaps the earliest example that has been reported of a ‘shearings lake’ is in The Crucifixion with Memorial Portrait of Hendrik van Rijn, ascribed to the Northern Netherlandish School, now in the Antwerp Museum of Fine Arts and dated 1363.67 Establishing the composition of this type of red lake is important not only for understanding pigment manufacture in this period, but also because its presence could affect interpretation of medium analysis. Although it does not influence gc-ms, in some cases where only a small amount of sample is available it might be that only ftir is used for analysis, and the protein content might be interpreted as a component of the binder rather than the pigment, and it would of course affect any protein analyses carried out using other techniques.

Paint Additives White (zinc) Vitriol A small but significant amount of zinc was detected by sem-edx analysis in the cross section of the paint sample from Arnolfini’s plum-coloured cloak.68 This result is especially interesting considering Eastlake’s assertion, from his research on documentary sources, that white copperas (zinc sulphate, ZnSO4.7H2O), was the most common drier used by Netherlandish and German painters at the time of Van Eyck.69 In fact, copperas is mentioned for this purpose in quite a number of documentary sources of a wide range of dates, either in recipes for preparation of the oil or as an addition on the grinding slab. Rather little attention had been paid to it, since not much evidence for its use had arisen during examination of paintings, but more recently, more and more examples of its

van eyck’s technique and materials

presence in works from different periods and geographical origin are being discovered, either through sem-edx on samples or X-ray fluorescence analysis.70 Eastlake’s conclusion came mainly from his study of the Strasburg manuscript, so it is perhaps significant that so far in fifteenth century paintings it has been found most commonly in works from Cologne. It was at first postulated to be present on the basis of detection of zinc by edx analysis in Stefan Lochner’s Saints Matthew, Catherine of Alexandria and John the Evangelist (ng705, about 1450), in the red paint of St Catherine’s cloak. Later, atr-ftir imaging confirmed that it was present in the form of zinc sulphate, and in addition showed that some zinc soap had formed, giving firm evidence that the copperas had indeed reacted with the oil medium and could have acted as a drier.71 Zinc sulphate (galicen stein) is also mentioned, however, as an ingredient in a red lake recipe in the Liber Illuministarum, a German source from around 1500,72 and since zinc was found in all the paint containing red lake in the Lochner, but not in other colours, it seems likely that it was introduced as a component of the pigment rather than the binding medium. This also needs to be borne in mind for the plum-coloured paint in the Arnolfini Portrait since there too the zinc is in a paint layer rich in red lake. A much broader study of Lochner carried out at the Doerner Institute in collaboration with the Wallraf-Richartz Museum, Cologne, found zinc in nearly every painting by him in areas of red lake, but importantly also in black paint,73 where it is possible to be more sure that zinc sulphate was included an additive with the intention of making the paint dry faster. Similarly, in Margaret van Eyck zinc has been detected not only in the red lake paint of the sitter’s robe but also in the black background, so it seems possible to say that Van Eyck did indeed use white copperas as a drier. Colourless Powdered Glass Two occurrences have also been confirmed in paintings by Van Eyck of another paint additive,

213

colourless powdered glass ground to a fine powder (particle size 5–20 μm). It is present in the Arnolfini Portrait, in the plum-coloured paint layer in Arnolfini’s robe (fig. 13.12), and in the red paint of the bed. In The Annunciation it was found in the translucent red paint of the brocade pattern on the Angel Gabriel’s cloak, again associated with red lake, and in the oil mordant for the gilded rays.74 The historic documentary sources refer to both adding powdered glass on the palette and in recipes for boiled oil.75 One of the earliest, in the Marciana manuscript from the second half of the sixteenth century, is for a mordant for gilding in which an ingredient is ‘glass first ground in water and then dried’, together with lead white, ochre, lead-tin yellow and boiled oil.76 The same list cited above from a south Netherlandish recipe book that mentions that copperas should be mixed with black advises that red lake is mixed with glass, and since all the other mixtures include a material that must be acting as a drier, it seems likely that this was also the intended function of the glass.77 The earliest reference that explicitly suggests this dates from the very end of the sixteenth century; Lomazzo notes that glass should be mixed with orpiment in his treatise on painting from 1584, with no indication of the purpose, but in his 1598 translation Richard Haydocke added the note ‘as a drier’ at this point in the text.78 Although Eastlake does not mention colourless powdered glass in connection with early northern painting, he does review references to it in seventeenth-century documentary sources, commenting that if it was a lead glass it might indeed be capable of acting as a drier.79 His contemporary Mary Philadelphia Merrifield, in her 1849 book on treatises on the art of painting, discusses this subject in much more detail and mentions that although a lead glass might have been used, colourless glass can also contain manganese, which could also catalyse the drying of oil.80 The glass added to paint in the Arnolfini Portrait and The Annunciation is not a lead glass but instead quantitative sem-edx analysis has shown that it is ordinary local colourless glass

214

marika spring and rachel morrison

50 μm

a

b

c

Figs 13.12a-c Arnolfini Portrait, cross section from Arnolfini’s robe: (a) visible light image; (b) back-scattered SEM image; (c) silicon EDX map. The red circles indicate the glass particles

made with wood or plant ash.81 It does contain manganese, however, which would have acted as a decolourizer, counteracting the slightly green tint caused by iron impurities, although in northern European glass this is likely to have been introduced as part of the wood ash rather than as a separate component. Colourless powdered glass has now been found as a paint additive in fifteenth- and sixteenth-century paintings from all over Europe. It has been so far been confirmed in seventy paintings in the National Gallery, and quantitative analysis of the composition has shown that it is usually the locally available glass.82 It is most commonly present in these works in combination with red lake, although in Italian paintings it is also regularly found in other colours. In the northern European paintings, however, its use was limited to mixtures with red lakes and occasionally to oil mordants for gilding, as in the Washington Annunciation. It is also always associated with an oil medium, and with both red lakes and mordants it might be desirable to add a siccative to enhance the drying properties of the paint. There is however still some uncertainty as to whether the manganese in glass is mobile enough to act on the oil. Powdered glass might also have been capable of modifying the working properties of the paint, perhaps giving it more body, and it may be that it was primarily used for this reason rather than as a siccative.83 The extensive study of colourless powdered glass as an additive in paintings in the National Gallery attempted to trace its use backwards in time. The paintings by Van Eyck are the earliest occurrences so far reported, although the signifi-

cance of this remains to be seen since not many earlier oil paintings have yet been analysed with this in mind. For Van Eyck and his contemporaries the reasons for adding powdered glass must have been empirical – either it was a practice that was already well established, or if they were the first generation to use it as an additive to oil they may have carried out their own experiments and felt that it in some way enhanced the properties of the paint. Interpretation of the historical recipes is hampered by the problem in interpretation of the word ‘glassa’. In the seventeenth-century recipes in English and French it is sometimes unambiguously glass, but some of these recipes are thought to be derivations, and perhaps mistranscriptions, of ‘glassa’, which is sometimes used to signify a hard resin. Despite these difficulties, since scientific analysis shows that colourless powdered glass certainly was used as a paint additive, some of the documents probably do refer to glass rather than resin. Resin as an Additive and the ‘Copper Resinate’ Problem Conversely, some of the documentary sources on preparation of oils do indeed include resins as an addition to oil, and they have quite often been detected through analysis in the most saturated and translucent red and green glazes on Netherlandish paintings. The darkest green of the woman’s dress in the Arnolfini Portrait, for example, has been found to contain some pine resin together with pre-polymerized linseed oil.84 This was identified by gc-ms carried out on a micro-sample in the form

van eyck’s technique and materials

of a paint scraping, and since spatial information about the location of the resin in the layer structure is therefore lost, there remains a slight doubt as to whether it is actually in the paint layer or whether there is some contamination of the sample by old varnish on the surface. atr-ftir imaging is useful for investigating this type of problem and a cross section from the green paint of the woman’s dress in the Arnolfini Portrait was re-analysed with this technique for this study (fig. 13.13). The deep greens in the Ghent Altarpiece were described in the 1950s study as ‘copper resinate’, by which it was meant that they were made by heating a copper salt with resin, with or without oil, and then used directly as a glaze paint. At that time and well afterwards it was believed that translucent greens in early Netherlandish paintings were generally prepared in this way, based on a few recipes in documentary sources (even though these did not relate to painting but to other branches of the arts), and on the observations that the glazes usually appeared very homogeneous in samples, without obvious pigment particles. The current prevailing opinion, however, is that the green pigment verdigris is used, and that it is only on ageing that this reacts with the oil in the medium to form copper oleate and, where resin is present as well, copper resinate, so that often the pigment appears to have ‘dissolved’ in the medium and the particles are no longer well defined.85 In the cross section from the Arnolfini Portrait the deep green upper layer of paint is not in fact perfectly translucent and there is some sense of the presence of particles. In spectra from atr-ftir imaging there is evidence of unreacted verdigris (fig. 13.13), showing that the paint was indeed made with this pigment, and there is also clear indication that it has reacted with the oil to form copper soaps (copper oleate). But what of the resin that was detected by gc-ms? The ftir spectrum of copper resinate has a very distinctive sharp band at around 1605 cm-1. The atr-ftir image of the distribution of this band across the area that was analysed on the cross section confirms that it is present, and it can be seen

215

that there is a concentration along the top of the sample in the zone which appears slightly brownish. This seems most likely to result from reaction of the paint with an old oil-resin varnish that has been applied to the surface in the past. Nonetheless, this does not exclude the possibility that some resin might also have been incorporated as an addition made by Van Eyck to the original paint, but to prove this it would be necessary to find copper resinate lower down in the paint layer. It is not possible to draw firm conclusions about this, however, even with the new analyses, since verdigris, copper oleate and copper resinate have many overlapping bands in their ftir spectra. Conclusion This article begins with a review of the literature on Van Eyck’s materials, and especially his paint medium, which provides a context that makes evident the strong interdependence between the interpretation of scientific analyses and the prevailing beliefs derived from historical sources, or from reconstructions and experiments, apparent in the evolution of the published analyses from the Ghent Altarpiece and from Netherlandish paintings in the National Gallery. In particular, where analyses have given results that cannot be interpreted in such a straightforward way as oil, there has been a tendency to turn to this literature for possible answers, or potential alternative media, in a way that has underestimated the influence of the pigment content of the paint on the chemistry of ageing of the oil. Armed with more recent knowledge of the latter, it is now possible to identify certain pigments as problematic in this respect, and to gain more understanding of the sometimes conflicting and confusing results from analyses of the binder. Thick impasto lead-tin yellow highlights in Early Netherlandish paintings are not in an emulsion and there is now greater awareness that the interaction between the oil medium and this pigment, as well as other lead pigments, can affect not only staining tests but also gc-ms analysis. Care is also needed when considering binding medium analyses

216

marika spring and rachel morrison

50 μm

a

c

1606 0.12 Absorbance Units

1732

Absorbance Units

0.14 0.12 0.10 0.08

d

1714

1411

1583 1554

b

0.06 0.04

0.10 0.08 0.06 0.04 0.02

0.02 0.00

0.00 2200

2000

1800 1600 Wavenumber cm-1

1400

2200

1200 e

2000

1800 1600 Wavenumber cm-1

1400

1200 f

Figs 13.13a-f Arnolfini Portrait, cross section from the woman’s green drapery: (a) visible light image showing the area analysed by ATR-FTIR imaging; (b) ATR-FTIR image produced by integration of the band between 1593-1572 cm-1 showing the copper carboxylate distribution; (c) detail of the visible light image showing the area analysed; (d) ATR-FTIR image produced by integration of the band between 1623-1594 cm-1 showing the area of copper resin interaction; (e) spectrum from a carboxylate-rich region (in red), with a reference spectrum of a copper-fatty acid soap (in black); (f) spectrum (in red) from the region in (d) showing the copper resin interaction, with a reference spectrum for a copper-resin acid soap (in black)

of paint containing ultramarine, as again there is evidence that it interacts with the medium on ageing, also implicated in the deterioration seen especially in paintings by Van Eyck. Early Netherlandish painters clearly did use oils that had been modified in various ways, and although the many different recipes for preparing oils have not been dealt with in depth here, evidence from analyses of Van Eyck’s paintings for pre-polymerization, either by sun thickening or by boiling, has been mentioned and this must have

had an important influence on the handling properties. These processes could include lead as a siccative, and two more additives perhaps also used to improve drying of the oil have been confirmed – white copperas (zinc sulphate) and colourless powdered glass, both referred to in Eastlake’s remarkably accurate deductions about oil media at the time of Van Eyck from documentary sources. Even though some of the historic recipes do refer to these additives as driers, in others the purpose is less obvious, hinting at the role of copperas as a

van eyck’s technique and materials

clarifier, or the use of glass to give the paint body, and there is still more to learn about how these act on the oil and how they would have modified its properties. Coremans and Thissen adopted an admirably cautious approach to interpretation of analyses in their exemplary examination of the Ghent Altarpiece in the 1950s. Even with increasingly sophisticated analytical techniques analysis of the medium in particular is still challenging, and there is much that is still not fully understood about the ageing and associated chemical changes in paint. Progress has nevertheless been made, giving a greater body of knowledge than was available in the 1950s, and no doubt much more will be discovered during the forthcoming research in association with conservation treatment of the Ghent Altarpiece. NOTES 1 See Baxendall 1964 on Bartolomeo Fazio and von Oettingen 1890 on Filarete’s treatise, which dates from between 1461 and 1464. 2 See Filarete in von Oettingen 1890, pp. 640-641. 3 Vasari 1568 (Bettarini 1968), vol. 1 (1966), pp. 132-133 (as summarized in Chapter XXI of Della Pittura) and vol. III, 1971, pp. 301-310 ( for the Vita di Antonello da Messina). 4 Ziloty 1941. 5 Brinkman 1993, Roy 2000, Effmann 2006. 6 Eastlake 1847, esp. vol. 2, ch. II, but also vol. 1, ch. X (pp. 348352) and ch. VIII. 7 Mérimée 1830, Raspe 1781. For Lessing see Soehnée 1822, pp. 47-71. 8 The Strasburg manuscript is thought to be from southern Germany and to date from around 1400. The De Ketham manuscript, Sloane MS 345 in the British Library, is partly in Middle Dutch and is probably from the end of the fifteenth century (Eastlake 1847, vol. 1, p. 282). See Campbell, Foister, Roy 1997, pp. 14-16, for documentary sources relevant to fifteenth-century northern European painting technique, their interrelationship and date. 9 Eastlake 1847, vol. 1, p. 289. 10 Eastlake 1847, vol. I, p. 265. 11 Mérimée 1830. 12 See Effmann 2006. 13 Eastlake 1847, vol. 1, pp. 130-131, for the Strasburg manuscript recipe that includes white copperas; vol. 2, pp. 284-285 for the recipe in the De Ketham manuscript ‘to temper all colours’, which includes this ingredient. Eastlake notes that the latter is for printing on cloth but also states that it ‘throws some light on the general Flemish practice of the fifteenth century’. 14 The term galicen stein or galitzenstein in the Strasburg manuscript refers specifically to zinc sulphate. In the De Ketham manuscript the term coperoet is used, which applies to copperas or vitriol more generally, the main three varieties often being distinguished by their colour – blue (copper sulphate), green (iron sulphate) and white (zinc sulphate), or by their place of origin. 15 Lessing 1774, p. 37. 16 Eastlake, vol. I, p. 277.

217

17 Eastlake, vol. I, pp. 130-131. 18 Mérimée 1830, p. 14. 19 Laurie 1933, p. 12, and Laurie 1934, p. 124. 20 See Effmann 2006, p. 20. 21 For a summary of Berger’s theories see Effmann 2006, pp. 20-21. 22 Laurie 1910, p. 375. 23 Laurie 1926, p. 44, pp. 189-190. See also Laurie 1933. 24 Doerner 1921 (1935), pp. 327-337. 25 Eastlake 1847, pp. 38-39. 26 See Spring 2000, which discusses documentary evidence for this practice and cites several examples where an aqueous binding medium has been reported for blue paint in works of art. 27 Coremans 1953, p. 70. 28 Coremans, Gettens, Thissen 1952, pp. 17-21. 29 Coremans, Thissen 1953. The results were also published in Coremans 1954, but there the component ‘x’ is not referred to and the medium is simply described as à base d’huile siccative, with the exception of the green glazes where ‘copper resinate’ with or without oil was proposed; see p. 155. 30 Coremans 1953, p. 70. An aqueous medium was reported for the two upper layers of the Virgin’s blue cloak, the final glaze on the blue jewels on her crown and the clasp of God’s cloak, and the blue tiles of the paved floor on two of the panels. 31 Coremans, Gettens, Thissen 1952. 32 See Plesters 1956, p. 129. Plesters mentions Margaret Hey’s experiments evaluating micro-tests for nitrogen such as the Lassaigne (Emich) test, where she observed that samples which were certainly oil paint quite often gave a false positive result. The analysis of the blues in the Ghent Altarpiece are not described in enough detail in the 1950s publications to know whether this test was actually used, or whether the conclusion that these blues were in an aqueous binder was based principally on the fluorescence and on a negative result in the tests that they were using to confirm oil, such as saponification tests. 33 Kockaert 1973-1974. The conclusion that the protein component was gelatin was based on staining with acid fuchsine, a stain specific for protein, and on Kockaert’s opinion that the physical properties were only compatible with egg white or gelatin of the fish glue type; he excluded egg white on the basis of a Raspail test. In a few cases samples that had not been mounted as cross sections were available for microchemical tests such as saponification and the Emich test. 34 Brinkmann et al. 1988/89. It is suggested that some oil may be dispersed within the layer of ‘resinate’. 35 Higgitt, Spring, Saunders 2003; Van der Weerd et al. 2002. 36 Masschelein-Kleiner 2005. 37 This is an issue to be aware of when considering older published results from before about the mid-1980s. 38 See Campbell 1998, pp. 30-31; Campbell, Foister, Roy 1997, pp. 40-43. 39 For a list of the paintings in which walnut oil was identified, see Campbell 1998, Introduction, n. 97. 40 For a list of the twelve works in which resin was found as an addition to the oil in red or green translucent paints see Campbell 1998, Introduction, n. 100. 41 In the Campin group: the portraits of a man and of a woman (NG653.1, NG653.2), and The Virgin and Child before a Firescreen (NG2609), ascribed to a follower. In the Van der Weyden group: the Exhumation of St Hubert (NG783), the Portrait of a Lady (NG1433) and the Pietà (NG6265). In the Bouts group: Portrait of a Man (NG943), Christ crowned with Thorns (NG1083), Virgin and Child (NG2595). An egg tempera medium was also reported in marbling on the reverse of A Young Man holding a Ring by a follower of Van Eyck (NG2602). Two other paintings reported as oil with some egg tempera are rather later in date: the Virgin and Child with Saints and Donor by a follower of Lieven van Lathem (NG1939) and the Triptych by the Master of Delft (NG2922). 42 See Spring et al. 2008.

218

marika spring and rachel morrison

43 Roy 2000, White 2000, Campbell 1998. 44 This conclusion is based on empirical observations made from a large number of GC-MS analyses; the exact chemical mechanisms are yet to be fully understood. See Higgitt, Spring, Saunders 2003. 45 Eastlake 1847, vol. 1, pp. 55-57. 46 White 2000. 47 Spring, Higgitt 2006. 48 Vandamme 1974, p. 119. Under the heading ‘Om te allieren alle manieren van verwen in dolie’ are a series of mixtures combining a pigment and the appropriate drier – among which is Root met menie, with menie referring to minium, or red lead. 49 Kroustallis, Bruquetas, Gómez 2011 discusses adulteration of vermilion with red lead in Spain in the eighteenth century, for which there is substantial documentary evidence, but it is clear from the results of analyses of paintings that this was already common practice in far earlier periods. 50 Van Eikema Hommes 2004 reviews the documentary sources in relation to methods of ameliorating discoloration of blue paint; see pp. 20-22. 51 Possible causes are reviewed in Plesters 1993, pp. 45, 58. See also Klaas 2011 and Spring 2012a. 52 In Del Federico et al. 2006 NMR was used to demonstrate the structural changes that can occur in ultramarine; one set of experiments explored the effect of an alkaline environment by subjecting test panels of ultramarine in a fresco technique to accelerated aging at 85% relative humidity. This seems a reasonable parallel to real fresco paintings, but when investigating the effect of an acidic environment (more relevant to oil paintings) the pigment was exposed to 3M HCl, and it is more difficult to assess the results in terms of processes that might actually occur in easel paintings. The effect of oxalic acid, an agent that can be found in paint layers through biological attack, or degradation of organic materials, has been studied by Zoppi et al. 2010: They found that oxalic acid did cause loss of colour in the pigment, with concurrent formation of elemental sulphur on release of the S3- chromophore. Calcium oxalate is often detected during FTIR analysis of ultramarinecontaining layers from oil paintings. See ‘Appendix: Formation of metal oxalates’ in Higgitt, White 2005, esp. p. 93. 53 Plesters 1980. See also Keith et al. 2011, esp. n. 25. The fact that oxalates are often detected in ultramarine-containing paint also suggests that it is not only the pigment that is affected but also the binding medium. See Higgitt, White 2005. 54 Van Loon 2008, p. 82. 55 The low azelate level detected by GC-MS analysis of an ultramarine glaze on the Washington Annunciation is an indication that the ultramarine has indeed affected the oil medium. See Gifford et al. in this volume for this result and for re-analysis of the medium in this painting. 56 The blanched shadows of Margaret’s red dress can be seen at http://www.nationalgallery.org.uk/paintings/research/the-restorationof-margaret-the-artists-wife/margarets-red-dress. 57 Paintings by Claude also typically suffer from this type of degradation. Interestingly, Plesters 1980 mentions he used ultramarine of small particle size, and speculates that it perhaps rendered it more susceptible to attack by acids in the environment. 58 Van Asperen de Boer 1974. 59 Mills, White 1985. 60 Bomford, Roy, Smith 1986, esp. pp. 53-54. 61 Campbell 1998. 62 See n. 52 for calcium oxalate. Spring, Higgitt 2006, p. 226, discusses its impact in ultramarine-containing layers on interpretation of FTIR spectra; if both lead soaps and calcium oxalate are present the spectrum can appear to contain protein, due to the overlap of bands from these components with amide I and II bands. 63 Not enough sample was available to confirm this by analysis with high performance liquid chromatography (HPLC).

64 Braekman 1986. I am grateful to Jo Kirby for bringing this recipe to my attention and for translating it. She has prepared a pigment of this type in the laboratory from a similar recipe in the Nuremberg Kunstbuch: see Ploss 1962. In the German recipe the jelly is simply ground with alum. See also Spring 2012a where the different stages are illustrated. 65 Kirby, Spring, Higgitt 2005; Kirby, Saunders, Spring 2006. 66 Nash 2010. 67 Wallert et al. 2012. 68 The spot EDX spectrum was from the middle of the paint layer so is certainly not associated with retouching. 69 Eastlake 1847, vol. 1, p. 349, vol. 2, pp. 34-35, discusses the merits of white copperas in some detail, including why it might be preferred over the more usual lead drier. He also points out that it should be well dried, or calcined, as mentioned in quite a number of treatises. 70 Merrifield 1847 gives references to white copperas in treatises by Pacheco, Palomino, and Roger de Piles (pp. ccxli-ccxliii). Reviews of documentary sources on this subject can also be found in Keller 1973 and in Dunkerton, Spring 2013, p. 15, pp. 24-25 and nn. 64-71. In the latter, occurrences are given of zinc sulphate in several Italian paintings, including those by Titian, as well as some additional northern European examples. Others identified since 2013 are: François Quesnel, Portrait of a Lady (NG2617), two paintings by Cranach, Vermeer’s Young Woman standing at a Virginal (NG1383), and a Dutch flower painting of the eighteenth century. 71 These ATR-FTIR imaging results are published in Spring et al. 2012. Zinc ions are a so-called secondary drier. They increase the effectiveness of a primary drier (one which accelerates the drying of the paint surface) and encourage through-drying of the paint film below the surface; see Tumosa and Mecklenburg 2005, and Erich 2006, pp. 10-12, for a summary. 72 Bartl et al. 2005. 73 The presence of zinc (but not confirmed as zinc sulphate) was reported in a number of paintings by Lochner and other Cologne painters; see Stege et al. 2012. 74 The occurrences in the National Gallery works were first published in Spring 2007. They were republished in Spring 2012b, together with the results from the Washington painting. This more extensive article brings together around seventy occurrences confirmed in paintings. 75 Spring 2012b includes an extensive review of documentary sources on colourless powdered glass as a paint additive. 76 Frezzato, Seccaroni 2010, p. 155. 77 Vandamme 1974: sinober met glas. Although sinober sometimes refers to vermilion, elsewhere in this book vermilloen is used, and sinober is most likely to mean red lake since it is close to sinople which appears in this document in the title for a red lake recipe. 78 Haydocke 1598. 79 Eastlake 1847, vol. 1, p. 351. 80 Merrifield 1849. The references to powdered glass are discussed in the introduction, vol. 1, pp. ccxl-ccxlii. 81 See Spring 2012b for the quantitative results. 82 Spring 2012b includes a full discussion of the quantitative analysis of the glass in around seventy occurrences confirmed in paintings in the National Gallery. 83 A full discussion of this point can be found in Spring 2012b, including references to publications describing practical reconstruction experiments attempting to investigate whether or not powdered glass can act as a drier or whether it changed the working properties in some other way. 84 See White 2000. 85 Van Eikema Hommes 2004 discusses documentary sources for the preparation of green paint and puts forward the argument that they are not in fact ‘copper resinate’ but verdigris in oil (with or without resin as an additive).

Fig. 14.1 Jan van Eyck and workshop, the Crucifixion and the Last Judgment, c.1435-1440, oil on panel transferred to canvas, each 56.5 x 19.7 cm The Metropolitan Museum of Art (no. 33.92ab), Fletcher Fund, 1933

14

Revelations Regarding the Crucifixion and Last Judgement by Jan van Eyck and Workshop Maryan W. Ainsworth

ABSTRACT: Among the intriguing questions about the Crucifixion and Last Judgement by Jan van Eyck and workshop at the Metropolitan Museum of Art in New York is whether these two paintings originally formed a diptych or the wings of a triptych. Recent investigations of the original frames through X-radiography and dendrochronology and of the paintings themselves with infrared reflectography add new information to help solve this question. Also pertinent to this discussion is the relationship of the recently resurfaced Crucifixion drawing and the underdrawing of the Metropolitan painting of the same theme. * This paper has now been superseded by much new information and conclusions presented in the Metropolitan Museum of Art exhibition A New Look at a Van Eyck Masterpiece (January 25 – April 24, 2016), publication forthcoming.

—o— Despite the considerable literature to date on the Metropolitan Museum’s Crucifixion and Last Judgement by Jan van Eyck and a workshop assistant (fig. 14.1), there are a number of issues that remain unresolved.1 One of these is the fundamental question of the relationship of these paintings to each other – whether they originally formed a diptych or were the wings of a triptych. The reconsideration of this matter here is prompted by new technical examinations of the frames by X-radiography and dendrochronology, and new documentation of the underdrawings in the paintings, the latter updating

the preliminary results that were first published by Stephanie Buck in 1995.2 As is widely known, the Crucifixion and Last Judgement panels – like The Annunciation in the National Gallery of Art, Washington – were transferred from wood to canvas when they all belonged to the Imperial Hermitage Museum, formerly Leningrad and today Saint Petersburg. In the case of the Metropolitan panels, we know exactly when this was done, as the restorer who undertook the transfer documented his work on the reverse of each canvas. Translated from the Russian, it reads, ‘Transferred from wood 1867 A. Siderov.’ Fortunately, Alexander Sidorowitsch Siderov (18351906) retained the original frames of the panels, although they were both cut through and thinned during the transfer process. Even though the paintings’ primary supports are lost to us, it is clear from the vertical grain of the wooden frames and a continuous crack running from the frame into the paint film at the bottom of the Crucifixion that the painting supports and frames were each carved out of one piece of oak.3 Therefore, Ian Tyers’s dendrochronology of the frames, likewise, could provide information about the dating of the missing original supports.4 The wood comes from the eastern Baltic area and the two frames originate from the same tree, apparently

222

maryan w. ainsworth

from the same board, linking head to head.5 The tree was certainly still growing in 1402, and considering minimal amounts of missing sapwood, was likely felled after around 1410. Tyers’s further calculations indicated that the boards could have been felled before around 1440. Although no definitive matches could be found so far between these boards and earlier dated Van Eyck paintings such as the Ghent Altarpiece, there are connections with the board used for the copy in a private collection of Van Eyck’s 1439 Virgin by a Fountain (Antwerp, kmska).6 Tyers noted that the tree-ring evidence suggests that the private collection copy and the Metropolitan Museum frames are probably not from the same tree, but from boards derived from the same general area of origin. Although not absolutely definitive, these results lend support for the supposition that the Metropolitan Museum paintings and frames date from the later part of Van Eyck’s career, that is, around 1435-1440. Close examination of the frames with the help of conservators George Bisacca and Alan Miller at the Metropolitan Museum provided a number of fresh observations that help to clarify the question of the original relationship of the paintings to each other. There is evidence found half-way up on the sides of the frames, namely a large hole on the right side of the Last Judgement that is certainly a very early artefact (fig. 14.2). The rectangular section hole, seen in the X-radiograph (marked in fig. 14.3b with the arrow), extends deep into the frame profile, tapering toward the tip of the cavity. Both of these aspects are characteristic of period handwrought nails. There also appear to be remnants of iron staining into the oak, as Bisacca discovered, a phenomenon that occurs from repeated humidity cycling causing the leaching of iron oxide into the wood and imparting a black stain. This is typical of nails that have been embedded for extensive periods of time. The left side of the Crucifixion frame (fig. 14.2) exhibits two holes, a larger one below a smaller one, that appear in the X-radiograph to go deep into the profile of the frame and to taper, as one would expect of old nails (fig. 14.3a,

Fig. 14.2 View of sides of frames showing nail holes

marked with arrows). These holes were probably made to accommodate the hardware of a latch, or perhaps two different types of latches during the history of the piece: one, a standard horizontalstyle latch, and another, a rather long pivoting L-shaped hook which would have swung downward (when the diptych was lying flat and was closed) into the fixed loop in the centre of the Last Judgement panel. All the holes are slightly cut off, indicating that they pre-existed the transfer of the panels to canvas in 1867 when the frames were cut through and thinned during the transfer process. In fact, the holes were already quite old by that date, and seem thus to support the idea that these two paintings once formed a diptych. Corroborating this idea has been the close connection of the pastiglia texts from the Bible that circumscribe each scene on the inner profile of the

revelations regarding the crucifixion and last judgement

a

223

b

Figs 14.3a-b (a) X-radiograph of left frame, detail, centre of left stile; (b) X-radiograph of right frame, detail, centre of right stile. Arrows indicate tapering cavities of previous metal attachments inserted from the sides. Circles indicate square section nail cavities (nails outside circles are modern). Closing latch hardware inserted from the sides is typical for diptychs, whereas latch hardware inserted from the reverse is common in triptychs

frames. Surrounding the Crucifixion is a text from Isaiah (53: 6-9, 12), while those around the Last Judgement are texts from Revelation (20: 13; 21: 3, 4; 20: 13); and Deuteronomy (32: 23, 24). As Hans Belting and Dagmar Eichberger noted, the inscriptions are specifically placed in order to correspond to the scenes depicted.7 This would explain, as Jacques Paviot continued, why – as a matter of symmetry – the text of the Crucifixion matches that of the Last Judgement by beginning at the inside centre-right of the frame, thus reinforcing the notion, as Belting and Eichberger had argued, that the panels were meant to form a diptych.8 But this organizational pattern is unique for Van Eyck paintings in which the texts on frames characteristically start at the upper left corner, as in

the Virgin and Child with Canon van der Paele, the Portrait of Jan de Leeuw, and the Dresden Triptych. Furthermore, Van Eyck’s inscriptions are invariably painted in trompe-l’oeil to imitate either deeply carved or raised letters. In this way, as Maurits Smeyers rightly observed, ‘[Van Eyck’s] frames are part of the work of art, of the painted reality.’9 The Metropolitan Museum frames are unique within Van Eyck’s oeuvre, for texts are rendered in actual relief in pastiglia. Also unusual is the crude insertion of text on right and left at mid-height of the Last Judgement frame that has been inelegantly reduced in size and squeezed into place among the larger-scale letters of the inscription. Furthermore, the style of the script that is used here varies from Jan’s habitual manner. As Smeyers noted, Van

224

maryan w. ainsworth

Eyck’s inscriptions are always a mixture of Roman capitals and uncials, derived from an epigraphic script from the last quarter of the twelfth century.10 Although the epigraphic style of the pastiglia inscriptions on the Metropolitan Museum’s frames is broadly in line with that of Van Eyck’s trompel’oeil texts, there are significant differences in some of the majuscules, such as the square C characters and the A with the forked cross-stroke so common in Van Eyck’s inscriptions, both of which are missing.11 These observations all raise the question of whether the pastiglia inscriptions were indeed planned from the outset by Van Eyck himself or were added later to the frames. Perhaps an answer to these troubling observations concerning the pastiglia script may be found in a recent discovery resulting from new X-radiography of the frames. Here, to our great surprise, there appeared beneath the admittedly quite crude applications of gilding on the flat part of the frame a fragmentary text in Gothic miniscule script that circumscribes the two frames. A detail of the best-

preserved portion that appears at the lower edge of the Last Judgement is turned here 180 degrees so that it is possible to read the letters in proper orientation. It shows that the damaged text is limited to the lower half of the letters (figs 14.4a-b). It has not yet been possible to decipher the text, and contact has been made with computer experts in probabilistic graphical models (‘pgm’) to see whether this problem can be solved with existing programs. Further study and comparison is necessary in an attempt to decipher the text and to answer the question of originality regarding this inscription in relation to the Crucifixion and Last Judgement images. Nevertheless, the placement of inscriptions on the flat portion of the frames is in fact typical of Van Eyck examples, while their position on the interior profile of the frame is not, even if the Gothic miniscule is not Jan’s preferred script. How might one explain the presence of the two texts on the frames? One possible scenario is that the frames of the Crucifixion and Last Judgement became so damaged at some point that the original

Figs 14.4a-b Hidden inscription at lower edge of the Last Judgement frame, turned 180°; (a) normal light, (b) X-radiograph

revelations regarding the crucifixion and last judgement

inscriptions could no longer be read in order to serve their key function for devotional practice. It was perhaps then that the pastiglia texts were added and the remnants of the original lettering on the flat portion of the frames were covered over by the rather crudely-applied gilding. When might this have occurred? As the cracks that developed in the frame also go through the pastiglia, the latter is very old. Johann David Passavant reported that the Russian Count Dmitry Tatistcheff, the first-known owner of the paintings, acquired them from a Spanish convent outside Burgos.12 Is it possible that the inscriptions in pastiglia (an unusual form of text in northern European paintings of the fifteenth century) were added during the time that the paintings were owned by a religious order where inscriptions in pastiglia were a typical accompaniment to frames? These new findings now throw into question what has long been assumed about the biblical texts surrounding the paintings, that is, that they were specifically incorporated by Jan van Eyck to create a kind of symbiotic relationship with the paintings. Furthermore, it also raises the question

225

of whether these two panels were simply adjusted later on to form a diptych after a central painting or sculptural group was destroyed, if at all completed.13 This line of questioning led to closer scrutiny of the new X-radiographs of the frames in order to search for additional helpful evidence. As previously mentioned, while excising the painting for transfer the frame’s thickness was reduced by half. This means that the back half of the current frame (marked in fig. 14.5 by the red arrows, when viewed from the side) is a later addition and not part of the original framing elements. Therefore, any evidence of a latch closing for a triptych, which would necessarily have appeared on the reverse of the frames when closed, is not readily visible because that wood has been covered by replacement wood. However, close examination of the X-radiographs revealed the remnants of four filled holes in the original wood of the frames.14 Two of these are matching, very old square-cut nail holes (circled in figs 14.3a, 14.3b), and two are the round-cut holes of modern screws, all found at the interior centre

Fig. 14.5 Inside edges of frames showing hinges and added wood

226

maryan w. ainsworth

back of the frames when they are arranged in the closed position of wings of a triptych. The two square-cut nail holes might well have been related to an original latch for two wings of a triptych. There is no getting around the fact that the Metropolitan paintings have the very tall, narrow dimensions of triptych-like wings, such as those of the Dresden Triptych, as opposed to the more squat dimensions of panels that formed a Van Eyck diptych, like the Annunciation in the Thyssen-Bornemisza Collection, Madrid. In the first description of the Metropolitan paintings, Passavant stated in 1841 that the Crucifixion and Last Judgement were wings of a triptych that formerly had as its centre an Adoration of the Magi, since stolen, and that on the reverse of the wings were remnants of paintings in grisaille.15 But an Adoration of the Magi has always seemed a curious theme to unite with a Crucifixion and Last Judgement. There is no precedent for this configuration or even certainty that the Adoration was original to the ensemble, and in the literature of the last decades there has been more acceptance of the idea that these two paintings originally formed a diptych.16 Augmenting this line of reasoning is another consideration, which has to do with the exceptional drawing of the Crucifixion that has recently resurfaced and has been studied by Arie Wallert (see fig. 37.1 in Wallaert’s paper in this volume). Although I have had an excellent photograph of the drawing in my files since 1993 from the rkd, The Hague, often pondering its relationship with the Metropolitan’s Crucifixion, it has only been possible to study it first-hand since it surfaced at the Rotterdam exhibition The Road to Van Eyck.17 Coincidently, recently the Metropolitan’s Crucifixion and the Last Judgement paintings were restudied with infrared reflectography, this time with the updated Indigo Systems Merlin Near Infrared (nir) camera.18 The enhanced results not only supported the previous conclusions reached at the time of the initial examination of the paintings but also provided such enhanced clarity of the details of handling and execution in the underdrawing that

it has become possible to make other comparisons, namely between the underdrawing of the Metropolitan Museum’s Crucifixion and the newly resurfaced drawing. Let us briefly consider the possible relationship of this drawing to the Metropolitan’s Crucifixion, as this also may have bearing on the question of whether the Metropolitan paintings were originally intended as a diptych or the wings of a triptych. At once apparent is that the drawing has a direct relationship to the Crucifixion painting, but neither is an exact copy of the other in any respect (compare fig. 14.1 with fig. 37.1). The most notable difference is that the drawing is a vertical rectangle, while the painting stretches the composition into the long narrow form familiar to us from triptych wings. Both scenes – specifically representing the moment that Longinus thrusts his lance into Christ’s side – take place in the foreground rocky landscape of Golgotha, before a cityscape representing Jerusalem, but the precise buildings in the background do not replicate each other. Beyond is a mountainous landscape that is rendered in breathtaking detail only in the Metropolitan painting. The various types of figures and the grouping of them are also distinctly similar in type, although none mimic each other exactly. Above all, the drawing and the painting closely parallel each other in the mood set by the range of expressions of the figures, despite their extremely small scale. In particular, there is the tumultuous crowd of jeering, laughing, and mocking men who support the action of the blind Longinus, assisted by another who guides the lance thrust into Christ’s side at the left of the Cross (figs 14.6a-b), balanced against those figures at the lower right of the Cross who – all except for a few – consider the gravity of the moment with far greater concern, a sense of wonder, or even pathos. Also in both drawing and painting is the group of figures surrounding the collapsed figure of the Virgin, supported in this tragic moment by John, who comforts her (figs 14.7a-c). In these as well as in other examples one could choose, the painting appears to fine-tune and

revelations regarding the crucifixion and last judgement

Figs 14.6a-b Comparison details of figures at the left of the cross in: (a) the Crucifixion drawing; (b) the Metropolitan Museum’s Crucifixion painting

228

maryan w. ainsworth

Figs 14.7a-c Comparison details of mourning figures in: (a) the Metropolitan Museum’s Crucifixion painting; (b) the Crucifixion drawing; (c) the underdrawing of the Crucifixion painting

adjust the preliminary ideas found in the drawing in order to enhance the desired emotional impact. The design of the painting is also adjusted from that of the drawing so as to fill the empty spaces that have developed from elongating the rectangular form of the composition from an independent study for the Crucifixion to accommodate the shape of the wing of a triptych. For example, to fill the space between the mourning figures at the bottom of the painting and those beneath the Cross the two onlookers at the left are brought out from behind the rocks in the drawing and fully defined in pose and costume, thus acquiring a greater role and leading our eye up to the Cross. The gypsy woman holding a child and seen from the back in the drawing is changed into the Cumaean Sibyl and turned to face the mourning group in the painting. Mary Magdalene is separated from the mourners in the drawing where she blends into the figure of John, and in the painting is featured not only by the green robe she wears but by the silhouetted pose of anguished supplication toward the crucified Christ. Greater emotional impact is gained in the painting vis-à-vis the drawing by hiding or only half-revealing the faces of some of the mourners. Especially important to note is the way in which the drawing and the painting’s underdrawing are closer to each other in the motif of the anguished expression of the Virgin, which is mollified in the painted version where her furrowed brow is smoothed, and her hand is raised to her cheek to convey a sense of intense sorrow and acquiescence (figs 14.7a-c). Further comparisons between the drawing and the underdrawing show startling similarities in the handling and execution of both – and here, in addition to the facial expression of the figure of John, we may note the specific treatment of the curls of the hair and so forth. In spite of the fact that the underdrawing is made with a very fine brush and the drawing is made with metalpoint, there are distinct similarities. The description that Stephanie Buck made of the underdrawing in the Metropolitan painting, closely comparing it to the underdrawing in the

revelations regarding the crucifixion and last judgement

unfinished 1437 St Barbara (Antwerp, kmska) in her publication of 1995, also suits the drawing: ‘The [foreground] figures are similarly outlined with continuous strokes; narrow parallel hatchings and cross-hatchings are placed rectangularly around these contours to bring the figures into relief and to indicate shadow. Long strokes with a thicker, roundish end which clearly divide the folds, and fine parallel hatchings that follow the direction of the ridge, characterize the modeling of elaborate draperies…’.19 This last statement in particular describes well the costumes of the two horsemen in the Crucifixion drawing and the underdrawing of the Metropolitan Museum’s painting, where both show comparable stiff, angular folds (figs 14.8a-b). In both the drawing and the painting’s underdrawing, the facial features of subsidiary figures are marked with simple dark strokes of the metalpoint or dots and singular lines made with the point of the brush, comparable to what also may be found in the 1437 St Barbara – a work that, like the

229

underdrawing of the Crucifixion panel, is just as detailed in the preliminary description of the foreground as that of the background. Even the handling of the description of the rocks and the bare ground between figures (as between the Magdalene and the Sibyl) shows a similar complex arrangement of even parallel hatching and cross-hatching in varied directions to indicate not only light and shade but also the distinctive characteristics of the facets of the rocks.

Figs 14.8a-b Comparison details of figures in: (a) the Crucifixion drawing; (b) the underdrawing of the Metropolitan Museum Crucifixion painting

230

maryan w. ainsworth

The Crucifixion drawing and the underdrawing of the Crucifixion painting have so many traits of handling and execution in common, and are so similar in their subtle expression of pathos, that they appear to be by the same hand. The drawing serves a preliminary function by working out a complicated composition from which the painting is further developed, clarifying as it does the distinctive features of the enhanced emotional appeal and suggested realism of the event at hand. The main differences from the drawing to the painting are due to the necessity of making a taller, thinner composition out of the shorter, wider workshop design. The most logical reason for this alteration would be the conversion of the workshop drawing’s composition into the elongated form necessary for the wing of a triptych. This provides further support for the hypothesis that the Metropolitan paintings originally were conceived and painted as the wings of a triptych and only somewhat later were reconfigured to be used in diptych form. This conclusion would explain Petrus Christus’s logical reuse of the composition of the Last Judgement for the wing of a triptych in the Berlin Gemäldegalerie Last Judgement and Annunciation and Nativity. There are still questions to be answered about the Metropolitan paintings– not the least of which is the transcription of the partially remaining texts buried beneath the gilding on the frame. However, what does seem clear is that the two extant Metropolitan Museum paintings, which have in modern times been regarded as a diptych, more likely originated as the wings of a triptych. This conclusion is supported by new findings regarding old nail holes not only on the sides but also on the flat interior portions of the frames when placed in the closed position of a triptych. Such a conversion from the wings of a triptych to a diptych helps to explain the addition of the pastiglia text on the frames, a unique form and location for texts in Van Eyck’s oeuvre. Further research may provide a reading of the newly discovered and very damaged text that

appears on the flat portion of the frames, the more usual position for Eyckian inscriptions. Finally, the newly resurfaced Eyckian Crucifixion drawing that so closely matches the style and execution of the underdrawing in the Metropolitan’s painting, appears to be adapted in its composition in the latter in order to conform to the requirements of a more vertically-oriented form, that of a wing of a triptych. Although this new technical investigation does not solve all of the perplexing mysteries of the Crucifixion and Last Judgement by Van Eyck and his workshop, it hopefully gets us closer to rediscovering its original format. Afterword: Since the paper published here, continuing resarch on the Crucifixion and Last Judgment and their frames has led to significant new findings: technical examination of the Latin pastiglia texts indicates that they are original to the frames and the paintings, the newly uncovered fragmentary text on the flat part of the frames is a Gothic minuscule Flemish translation of the Latin pastiglia, and continuing investigations point to the possibility that these paintings once decorated the interior surfaces of doors of a tabernacle or reliquary shrine. These and other discoveries were presented in an exhition at The Metropolitan Museum of Art, A New Look at a Van Eyck Masterpiece (25 January – 24 April 2016), and the related publication is forthcoming. Further observations concerning the relationship between the Metropolitan’s paintings and the recently rediscovered Crucifixion drawing now in the Museum Boijmans Van Beuningen, Rotterdam, have been published by the author in ‘The Crucifixion: The Relationship Between the Rotterdam Drawing and New York Painting,’ in Albert J. Elen and Friso Lammertse, eds., An Eyckian Crucifixion explored: ten essays on a drawing, Musem Boijmans van Beuningen, Rotterdam, 2016, pp. 116-33.

revelations regarding the crucifixion and last judgement

NOTES * I am particularly grateful to George Bisacca and Alan Miller, conservators in the Paintings Conservation Department, Metropolitan Museum of Art, for carrying out new X-radiography of the frames of the Crucifixion and Last Judgement and discussing the findings with me. I also thank Shawn Digney-Peer for undertaking new infrared reflectography on both paintings for my recent research. Ian Tyers investigated the frames and provided new dendrochronological results. 1 For the complete references on the two paintings, see the Metropolitan Museum of Art website, under ‘collections’. 2 Buck 1995, pp. 65-69, 71-72, 74, 77-78, 79-83 nn., figs 1-4 (details). 3 For comparisons with other Van Eyck frames, see Verougstraete, Van Schoute 2000, pp. 107-15, esp. p. 108. In order to excise the painted supports from the Museum’s frames during the transfer process, Siderov cut the top edges of the frames at the right and the left and planed off the lower halves of the frames along with the support. 4 Ian Tyers, Dendrochronological Consultancy Limited, Report 565, November 2012. It is rare that a transferred panel retains enough sufficient original timber to allow for tree-ring analysis. This raises the question about other deeply carved panels transferred by the Siderov brothers, which might well be a rewarding area of study. For other studies about Siderov transfers see Marconi 1965, pp. 246-254; Hartwieg 2005, pp. 38-46, esp. p. 45. 5 Email 4 October 2012 from Ian Tyers to Maryan Ainsworth (European Paintings Curatorial Files, MMA). 6 The dendrochronology of the Antwerp painting and its copy was carried out by Joseph Vynckier and is reported by Depuydt-Elbaum 2002, p. 19. The Baltic oak trees for the panels of the Antwerp painting and the copy were felled in 1411 and 1419 (at the earliest) respectively.

231

7 See Belting, Eichberger 1983, pp. 103-107; Eichberger 1987, pp. 85-90; Paviot 2010, pp. 150, 161. 8 Ibid. 9 Smeyers 1983, p. 404. 10 Smeyers 1995, pp. 403-414, esp. p. 406-407. 11 I am grateful to Douglas Brine for a continuing discussion about Jan van Eyck’s texts (email of 18 November 2012 to the author; European Paintings Curatorial Files, MMA). These anomalies have led some, such as Smeyers (1983, p. 406), even to question whether the Metropolitan paintings are by Van Eyck. 12 Passavant 1841, p. 9. 13 Recently Stephan Kemperdick and Friso Lammertse have once again favoured the notion that the Metropolitan paintings originally formed the wings of a triptych whose centrepiece was a sculptural group of the Lamentation. Rotterdam 2012, pp. 105-106, fig. 20. 14 These were first noticed by George Bisacca who discussed the implications of the finding with me. 15 Passavant 1841, pp. 9-10. 16 Among others, see Belting, Eichberger 1983; Paviot 1996; and Borchert 2006, pp. 175-177. 17 See Messling in Rotterdam 2012, exh. cat., no. 86, pp. 303304. 18 The painting was examined in July 2012 using an Indigo Systems Merlin Near Infrared (NIR) camera. The Merlin has a solidstate InGaAs (Indium Gallium Arsenide) detector sensitive to wavelengths from 900 to 1700 nanometres. The array format is 320 ≈ 256 pixels. The camera is used in conjunction with a National Instruments IMAQ PCI-1422 frame grabber card and IRvista 2.51 software. The lens is a macro custom-made by StingRay Optics, LLC, optimized for wavelengths from 900 to 2500 nanometers. Images are acquired as 16-bit .BIN files, exported as 8-bit .TIFF files and assembled in Adobe Photoshop (usually CS3). 19 Buck 1995, p. 67.

Fig. 15.1 Jan van Eyck, Portrait of Margaret van Eyck, the artist’s wife, 1439, oil on panel, 41.2 x 34.6 cm Bruges, Groeningenmuseum (inv. no. 0000.GRO0162.I)

15

Remarks on Character and Functions in Jan van Eyck’s Underdrawing of Portraits: the Case of Margaret van Eyck part 1 Rachel Billinge

ABSTRACT: In 2009 the Portrait of Margaret van Eyck from the Groeningemuseum in Bruges was the subject of technical examinations conducted by members of the Conservation and Scientific Departments of the National Gallery in London and similar to those already carried out on the Van Eyck portraits in the National Gallery. Infrared reflectography played an important part in the examinations, taking advantage of the opportunity to use OSIRIS, one of the new InGaAs digital cameras. The new infrared reflectograms are discussed and compared with those of the National Gallery portraits

—o— The Portrait of Margaret van Eyck (fig. 15.1) from the Groeningemuseum, in Bruges, was very generously lent to the exhibition Renaissance Faces: Van Eyck to Titian at the National Gallery in London (15 October 2008 – 18 January 2009). At the close of the exhibition the curator of the Groeningemuseum asked the Conservation and Scientific Departments of the National Gallery to conduct technical examinations similar to those already carried out on the Van Eycks in London for the Gallery’s fifteenth-century Netherlandish catalogue.1 The portrait was also cleaned by Jill Dunkerton (see the paper by Dunkerton, Morrison and Roy in this volume).

Infrared reflectography was an important part of the examinations, taking advantage of the opportunity to use osiris, one of the new InGaAs digital cameras.2 Of course, this was not the first time that infrared imaging had been used to look at the Portrait of Margaret van Eyck – most comprehensively, in 1992, Professor J.R.J van Asperen de Boer conducted a full study with infrared reflectography using an infrared vidicon.3 His examination of the painting had been very thorough and it was not expected that any new, exciting discoveries would be found, but it was hoped that the digital camera would be able to produce a clearer, less patchy image to illustrate what Van Asperen de Boer reported as having seen, which was not always easy to see in the published mosaic. The new reflectogram, recorded after the removal of discoloured varnish and repaints, did, as hoped, produce a much clearer image (fig. 15.2). As Van Asperen de Boer reported, infrared imaging shows ‘numerous shifts in the outer contours of the head scarf and red cloak [...]’.4 He does not say this explicitly but these changes, such as the repositioning of Margaret’s right sleeve and the extending of the horns of her headdress, were made during

234

rachel billinge

Fig. 15.2 Portrait of Margaret van Eyck, IRR

remarks on character and functions in jan van eyck’s underdrawing of portraits

Fig. 15.3 Portrait of Margaret van Eyck, detail, the veil and horn, IRR

235

236

rachel billinge

Fig. 15.4 Portrait of Margaret van Eyck, detail, the eyes, IRR

painting and show because both positions of the paint can be seen. The underdrawing in the headdress is quite bold and free (fig. 15.3) – basic outlines include a broad black wiggly line indicating the bottom edge of the crimped part, and hatching for the areas of deep shadow on the horn. Probably the most exciting of Van Asperen de Boer’s findings was the shift in the eyes, which now shows very clearly: the painted eyes are slightly lower and to the right of their underdrawn positions (fig. 15.4). Van Asperen de Boer went on to say: ‘There is little hatching visible in the face – the age crack pattern interfering – except for the shading of the temple, cheek and neck. This would seem to be in agreement with the hypothesis that Van Eyck portraits of persons available for sitting show paucity of underdrawing’.5 In fact, although the images are still dominated by the crack pattern, the new reflectograms show that in the face there is extensive, delicate and careful hatching, creating a fully modelled and almost three-dimensional image of

Margaret van Eyck’s face (fig. 15.5). This drawing is different in character to the more formulaic, less detailed drawing for the headdress and clothes. Many of the portraits by Jan van Eyck have now been examined with infrared reflectography but comparisons will be limited to the three signed works in the National Gallery. In comparing the infrared images it is worth bearing in mind that the 1433 probable Self Portrait (ng222) and Léal Souvenir (ng290) are similar in scale to the Portrait of Margaret van Eyck, while the full-length Portrait of Giovanni(?) Arnolfini and his Wife (ng186) has proportionately much smaller figures. Having achieved such good results with osiris on the Portrait of Margaret van Eyck the two headand-shoulders portraits were re-examined. These new reflectograms were recorded in the National Gallery one morning prior to the arrival of the public, which limited what could be done with the lighting. Even so, the results are improvements on the previous Hamamatsu mosaics recorded for Lorne Campbell’s 1998 catalogue.6 For the Self

remarks on character and functions in jan van eyck’s underdrawing of portraits

Fig. 15.5 Portrait of Margaret van Eyck, detail, the face, IRR

237

238

rachel billinge

Fig. 15.6 Jan van Eyck, Portrait of a Man (Self Portrait?), detail, the face, IRR London, National Gallery (NG222)

remarks on character and functions in jan van eyck’s underdrawing of portraits

Fig. 15.7 Jan van Eyck, Portrait of a Man (‘Leal Souvenir’), detail, the head, IRR London, National Gallery (NG290)

239

240

rachel billinge

Fig. 15.8 Jan van Eyck, Portrait of Giovanni (?) Arnolfini and his Wife, detail, the face of Arnofini’s wife, IRR London National Gallery (NG186)

Portrait, Campbell reported that ‘In the painting itself, a very small amount of underdrawing can be discerned under the microscope and in infrared reflectograms. […] The contour of the nose appears to be underdrawn slightly to our right of the painted contour […] and there may be some exceedingly fine hatching in the shadow of the nose’.7 The new reflectogram is not much clearer but there is hatching for the shadow in the nose and possibly also in the temple and cheek (fig. 15.6). Hatching was easier to make out in the face of Léal Souvenir (fig. 15.7), although as Campbell says in the catalogue ‘The infrared reflectograms […] are difficult to interpret because of the extensive retouching’.8 Digital infrared imaging does not help with that particular problem, which will only be solved if and when the painting is cleaned. Nonetheless, the new reflectogram is clearer and it is possible to see that Van Eyck has modelled the head with extensive hatching, less delicate than

the hatching used for the Portrait of Margaret van Eyck but similarly detailed. Perhaps unexpectedly, given the very different scale, the portrait whose underdrawing is most like that in the Portrait of Margaret van Eyck is that of Arnolfini’s wife (fig. 15.8). Her face is only 6.5 centimetres high (compared to about 11 centimetres for Margaret) but the underdrawing is very detailed, fully modelled and similarly three-dimensional. Also comparable are the somewhat formulaic, freer, broader lines for the headdress, the broad wiggly outline for the edge of the headdress, and hatching, in two directions, on the horns. Writing in 1991, Van Asperen de Boer reported that ‘it had been argued recently’ that detailed underdrawing like that seen in the Portrait of Cardinal Niccolò Albergati in Vienna or the face of Joos Vijdt in the Ghent Altarpiece was made using a preparatory drawing, while other, signed and dated portraits such as Jan de Leeuw (Vienna) and the Self Portrait (London) ‘show hardly any underdrawing and would have been painted directly from the sitter’.9 This hypothesis has since often been repeated but it would appear that as ever-clearer images of underdrawing become available, there is a need to revisit the evidence and think again about how Jan van Eyck might have been using underdrawing in his paintings. NOTES 1 Campbell 1998. The results were first made available to the public on research pages of websites connected with the two museums: http://www.nationalgallery.org.uk/paintings/research/the-restoration-of-margaret-the-artists-wife/introduction and http://vlaamseprimitieven.vlaamsekunstcollectie.be/en/research/webpublications/ the-restoration-and-technical-examination-of-jan-van-eycks-margaret-the-art. 2 The digital infrared scanning camera OSIRIS contains an indium gallium arsenide (InGaAs) sensor. For further details about the camera see www.opusinstruments. com/index.php. 3 For infrared photography, see Janssens de Bisthoven 1981. The account of the Van Asperen de Boer study was presented in 1993, see Van Asperen de Boer 1995. 4 Van Asperen de Boer 1995, p. 81. 5 Van Asperen de Boer 1995, p. 82. 6 Campbell 1998, pp. 215, 221. 7 Campbell 1998, p. 214. 8 Campbell 1998, p. 218. 9 Van Asperen de Boer, Ridderbos, Zeldenrust 1991, pp. 9, 10.

remarks on character and functions in jan van eyck’s underdrawing of portraits

241

part 1i Till-Holger Borchert

ABSTRACT: Jan van Eyck’s Portrait of Margaret van Eyck in the collection of the Groeningemuseum in Bruges underwent conservation treatment at the National Gallery in London in 2008-2009. During this treatment the painting was investigated by scientific methods and new infrared reflectograms and X-radiographs were taken. The results of IRR examination revealed a hitherto unnoticed richness of the underdrawing, which is compared with other underdrawings in Van Eyck’s portraits as well as with the metalpoint portrait drawing in the Kupferstich-Kabinett in Dresden. This article analyses the form and function of the underdrawing in Van Eyck’s portraits and raises questions regarding the production process of portraiture in his workshop.

—o— Any observations regarding the character and functions of Jan van Eyck’s underdrawing of the Portrait of Margaret van Eyck cannot be separated from enquiries about the painting’s original destination and context. These two aspects are of vital importance to our understanding of the process by which the painting was produced and are therefore relevant to the study of its underdrawing. Nothing is known about the motives and ideas that inspired Jan van Eyck to paint a likeness of his wife, however, and we are similarly ignorant of the portrait’s original use as well as the place for which it was intended. Nevertheless, it should be stressed that Jan van Eyck’s Portrait of Margaret van Eyck (fig. 15.1) is in more ways than one an extraordinary and exceptional painting. It is one of only a few Netherlandish portraits of the early fifteenth century to represent a female sitter that is not part of devotional imagery, and it is therefore often considered – perhaps somewhat misleadingly – as

an independent or autonomous portrait.1 While it may not be the earliest example of its kind, north of the Alps it is the first surviving non-dynastical portrait of a woman whose sitter can be identified.2 The identification is based on the illusionistic inscription that Jan van Eyck provided on the original frame. The meticulously-painted inscription gives the appearance of having been engraved into metal plates attached to the upper and the lower edges of the panel’s marbled frame. The inscription reveals the identity of the sitter as the painter’s wife, known from documents as damoiselle Marguerite.3 The portrait seems to be speaking since the sitter is addressing the viewer directly by way of the inscription: CO(n)IV(n)X M(eu)S IOH(ann)ES ME C(om)PLEVIT AN(n)O . 1439. 15º.IVNIJ. / ETAS MEA TRINGINTA TRIV(m) AN(n) ORV(m). (‘My husband Johannes completed me on the 15th of June 1439, my age is 33 years’). To this statement is appended Jan van Eyck’s personal motto, ALC.IXH.XAN (Als ich kan).4 The panel’s superb condition re-emerged gloriously during Jill Dunkerton’s recent conservation treatment.5 Its pristine appearance strongly suggests that the portrait was kept in favourable circumstances from the outset and that it has been treated with great care through the centuries. Most likely the careful treatment of the painting was primarily linked to the prestige of its creator, Jan van Eyck. In addition, circumstantial evidence points to the fact that the portrait must also have been considered to be of special significance and value precisely because of the painter’s prominent relationship with the sitter, his wife. That Margaret’s

242

till-holger borchert

Fig. 15.9 Jan van Eyck, Portrait of Nicolò Albergati, 1435, silver- and goldpoint on prepared paper, 21.3 x 18 cm Dresden, Staatliche Kunstsammlungen, Kupferstich-Kabinett (inv. no. C775)

remarks on character and functions in jan van eyck’s underdrawing of portraits

portrait was held in great esteem is attested by the fact that it was venerated in a manner that to some extent mirrors the devotional veneration of relics. When the Portrait of Margaret van Eyck was first recorded in the second half of the eighteenth century in the possession of the Bruges guild of painters it was already customary to exhibit the panel annually in the chapel of the painter’s guild on 21 October during the Feast of St Luke.6 Interestingly enough, at that moment in time the painting apparently had to be secured against theft since, according to local tradition, a ‘pendant’ that used to be shown alongside the Portrait of Margaret van Eyck in Bruges had been stolen in the past.7 This information is particularly revealing since at first glance it seems to shed light on the possible original context in which the portrait might once have been shown: on second inspection, however, things turn out to be more complicated. While it is not impossible that the Portrait of Margaret van Eyck was originally conceived as a pendant to a self-portrait by Jan van Eyck – in this case the likeness of the painter would have quite literally ‘completed’ the likeness of his wife as part of a conjugal representation8 – the eighteenth-century sources contain no evidence whatsoever that the pendant was indeed a self-portrait by Van Eyck. The stolen pendant portrait might have represented someone entirely different, and only later was it taken to be the likeness of the famous painter. It might even have been a representation of Van Eyck painted at a much later date by another artist altogether, incidentally becoming a ‘pendant’ of the Portrait of Margaret van Eyck in the context of the Bruges painter’s guild. At what precise point in time the portrait came into the ownership of the painters’ guild remains unknown. It was probably in the later sixteenth century, since Albrecht Dürer made no mention of the existence of a self-portrait by Van Eyck or the portrait of the painter’s wife when he visited the guild’s chapel on 8 April 1521.9 We can only speculate on what the original function of the Portrait of Margaret van Eyck might have been, as portraits of the time seems to have

243

served in a variety of contexts, such as fulfilling representational or commemorative purposes.10 The mere fact that in this marvellous portrait Jan van Eyck represented his own wife – the first recorded portrait of a painter’s wife to have survived – would seem to suggest that it was among the most personal paintings that he ever made.11 We can safely assume that the portrait was not commissioned but was produced on Van Eyck’s own initiative. Were it not for the conspicuous fact that the ‘speaking’ inscription on the frame is in Latin rather than Netherlandish one could assume that the portrait was originally destined for a private rather than a public context. The inscription and the portrait’s inherent interaction with the viewer, however, clearly suggest that its original context and function was not exclusively private.12 While the original function of the Portrait of Margaret van Eyck remains uncertain, we can safely assume that the circumstances of its production may have differed somewhat from those portraits that the artist made on demand. In contrast to portraits that were commissioned, for which Jan van Eyck presumably had only limited access to his sitters, he must have had ample opportunity to study his own wife’s likeness either in the privacy of their living quarters or in his workshop. Notwithstanding the possibility that Jan van Eyck knew the facial features of his wife by heart, it seems more than reasonable to assume that he could easily return to his model to confirm specific details of her appearance at any moment during the painting process. The exceptional circumstances of the portrait’s genesis were alluded to in 1995 by J.R.J. van Asperen de Boer in his brief discussion of its underdrawing following his own investigation of the panel by infrared reflectography.13 In addition to the discovery of shifts in the position of the eyes that had previously not been detected, he was struck by the fact that only a limited amount of hatching was detectable in Margaret’s face. This observation seemed to confirm an earlier hypothesis – based on a detailed study of Van Eyck’s

244

till-holger borchert

Portrait of a Man with a Blue Chaperon14 – that the ‘paucity of underdrawing’ was a typical feature of Van Eyck’s portraits of persons who were easily ‘available for sitting’. It was implied that those portrait commissions for which he only had limited access to the sitters were underdrawn in a much more detailed way. Nonetheless, in 2000 Rachel Billinge commented that this hypothesis ‘was less than clear cut’.15 In the light of the results of Rachel Billinge’s recent re-examination of the Bruges portrait,16 however, it is necessary to reconsider Van Asperen de Boer’s hypothesis, since the painting is far more extensively underdrawn than the earlier infrared images revealed. Indeed, the entire face of Van Eyck’s wife now appears to have been prepared by means of an extremely detailed and careful underdrawing, features not previously apparent. These consists of contour and placement lines as well as larger zones of hatching around the eyes, nose, chin, cheek and forehead (figs 15.2-15.5). Arguably, the underdrawing of the Portrait of Margaret van Eyck is the most detailed of the surviving portraits by Jan van Eyck that have been examined, including those that have been reinvestigated recently. How does the careful and detailed underdrawing of the Portrait of Margaret van Eyck relate to the production of the panel? In contrast to Van Asperen de Boer’s hypothesis, it could be argued that the underdrawing of the Bruges portrait is particularly detailed precisely because the painter’s wife was available to the artist as a model during the entire production process. Whereas usually Van Eyck would have had to make use of detailed preparatory portrait drawings as a basis for painted portraits, as in the case of the remarkable metalpoint drawing of Niccolò Albergati in the Kupferstich-Kabinett in Dresden (fig. 15.9),17 he could have avoided this intermediary step in the Bruges portrait and prepared the detailed drawing of his wife directly on the chalk ground of the panel using Margaret as a model. This scenario, however, is not very likely if one takes into account that underdrawing consisting of

contours and placement lines as well as various zones of hatching can be observed in most of the portraits by Jan van Eyck, including his donor portraits. The portraits of Joos Vijdt and his wife Elisabeth Borluut (figs 15.10a-b) are both carefully underdrawn in a liquid medium, presumably with a brush. Single or multiple lines indicate contours and define the placement and position of mouth, nose and eyes. Smaller zones of hatching are used to subtly enhance the plasticity of the face while larger areas of hatching are employed to define the distribution of light. All this took place during the preparatory stage of the painting process.18 Even if the donors of the Ghent Altarpiece were unavailable for extended portrait sessions – forcing the painter to make use of preparatory drawings for underdrawing and painting – Jan van Eyck seems to have regularly used detailed drawings such as the Albergati metalpoint drawing in the preparation of his portraits. Indeed, their use seems to be a consistent part of Van Eyck’s working method as portraitist. The amount of time that sitters were available to him as models does not alter the production process and consequently fails to explain variations in the extent of underdrawing that one encounters in his portrait paintings. One particularly intriguing example is the Portrait of a Man, 1432, in the National Gallery of Art in London. The panel, which carries the enigmatic inscription ‘leal souvenir’ on a painted parapet as well as the word tym.otheos written in pseudoGreek letters, is the earliest dated portrait by Jan van Eyck. The likeness of the unidentified man is underdrawn in a detailed manner, using a brush (fig. 15.7). Although some of the nineteenthcentury restoration occasionally interferes with the genuine underdrawing, the extent of the portrait’s underdrawing – especially the zones of hatching around the chin and cheeks – is noteworthy.19 Although we cannot be sure, it seems likely, as has been suggested, that Jan van Eyck’s Léal Souvenir of 1432 served as a commemorative portrait and may actually have been painted posthumously.20 If it were indeed a posthumous portrait

remarks on character and functions in jan van eyck’s underdrawing of portraits

with a commemorative function, then in all likelihood Jan van Eyck would have had no access to the sitter at all; instead, the portrait would either have had to be based on the artist’s memory or on visual and verbal records provided by third parties. How authentic would Jan van Eyck’s representation – the Léal Souvenir – of a deceased ‘sitter’ have been? And what precisely was the function of the underdrawing, given that its purpose could have not been a record of the sitter’s physiognomy based on preparatory studies? I would like to argue that if one disregards for the moment the contour and placement lines of the portrait that are constituting a generic face, it is clear that the primary purpose of the underdrawing was the careful modelling of the facial features by means of light and shade. The distribution of light helped define the volumes of facial features such as nose, eyes and mouth. Moreover, it greatly enhanced the portrait’s three-dimensional qualities. The underdrawn zones of hatching were therefore an aid to enhancing the illusionistic character and the vividness of the portrait that in turn served as surrogate for an actual sitter. The underdrawing in the Portrait of Margaret van Eyck serves an identical function in as much as its extensive hatching defines the light, dark and shaded parts of her physiognomy, the light clearly coming from the left. The pronounced representation of the effects of a strong light source boosts the plasticity of her face and enhances the vividness of her appearance. The same approach is seen in the underdrawing of Jan van Eyck’s presumed self-portrait of 1433 (fig. 15.6). Even though the underdrawn contours and zones of hatching register more faintly than in the case of the Bruges portrait, the primary function of the hatching is again the distribution of light and shade of the face and its plasticity.21 Finally, zones of hatching around the chin, nose and cheek can be seen in the recent infrared reflectograms of the Vienna Portrait of Jan de Leeuw, painted by Van Eyck in 1436, although both contours and hatching register only faintly (fig. 15.11).22

245

Of particular interest is the underdrawing of the Vienna Portrait of Cardinal Niccolò Albergati (fig. 15.12) since it can be directly compared in its form and function with the preparatory drawing in Dresden (fig. 15.9).23 In April 1435, Jan van Eyck was summoned by the Duke of Burgundy to travel to Arras, where the peace negotiations between Burgundy and France were taking place. In the capital, Artois, Jan van Eyck must have been able to draw the Cardinal’s portrait. This drawing ultimately served as a model for the painted portrait that he and his workshop produced three years later. It is generally assumed that Van Eyck made the Dresden drawing – the only surviving autograph sheet by him – while Albergati sat for him in Arras.24 The identification of the sitter as Niccolò Albergati has sometimes been doubted, but it is firmly based on an inscription on the original frame, now lost, or on a commemorative text that was once attached to the portrait.25 Peeter Stevens, a picture dealer from Antwerp who sold the Vienna portrait to Archduke Leopold, recorded the sitter’s identity in some detail in his own edition of Karel van Mander’s Schilderboeck by copying this information.26 According to Stevens, Van Eyck’s portrait was dated 1438 and represented the ‘Cardinal of Santa Croce’, who had been sent by the pope to negotiate the peace between Burgundy and France. The papal legate at the Congress of Arras was indeed the same Niccolò Albergati who had been made Cardinal Priest of Santa Croce in Gerusalemme in Rome in 1426:27 Noch bij Peeter Stevens een fraey conterfaysel van Jan van Eyck met dato 1438, wesende den Cardinael Santa Croce, die alsdoen tot Brugge was gesonden vanden Paus om de peys te maecken met Hertoch Philips over syn vaders doot met den dolphyn van Franckryck (‘Also at Peeter Stevens’s, a fine portrait by Jan van Eyck, dated 1438, being the Cardinal Santa Croce who at that time was sent to Bruges by the Pope in order to make peace between Duke Philip and the French Dauphin in the matter of Philip’s father’s death’).28 The underdrawing of the Portrait of Niccolò Albergati consists of placement lines and contours

246

till-holger borchert

Fig. 15.10a Ghent Altarpiece, detail, Joos Vijd, IRR

remarks on character and functions in jan van eyck’s underdrawing of portraits

Fig. 15.10b Ghent Altarpiece, detail, Elisabeth Borluut, IRR

247

248

till-holger borchert

Fig. 15.11 Jan van Eyck, Portrait of Jan de Leeuw, 1436, oil on panel, 33.2 x 27.6 cm, IRR Vienna, Kunsthistorisches Museum, Gemäldegalerie (inv. no. GG 946)

remarks on character and functions in jan van eyck’s underdrawing of portraits

Fig. 15.12 Jan van Eyck, Portrait of Nicolò Albergati, 1438, oil on panel, 34.1 x 27.3 cm, IRR Vienna, Kunsthistorisches Museum, Gemäldegalerie (inv. no. GG 975)

249

250

till-holger borchert

that were executed with a small brush. In addition, there is a considerable amount of parallel hatching, also applied with a brush, which covers larger areas of the right side of the Cardinal’s aged likeness. As in the other portraits by Van Eyck discussed above, hatching indicates shade and defines the distribution of light and shade across the sitter’s physiognomy. This kind of hatching allows the painter to model the face and to heighten the sense of the sitter’s presence and vividness of the likeness. The graphic vocabulary of the underdrawing of the Vienna portrait clearly corresponds to the form of the underdrawing of the Portrait of Margaret van Eyck, even though the hatching is more refined and detailed in the Bruges panel. The modelling of Margaret’s forehead, cheek and chin by means of carefully applied hatching, however, is closer in style to the drawing in Dresden than it is to the underdrawing of the Vienna panel.29 Executed in several working stages with goldand silverpoint, the sheet in Dresden served as a preparatory drawing for the portrait in Vienna. Jan van Eyck used various metalpoints to reproduce the likeness of his sitter in several stages, clarifying and correcting contours and carefully modelling his physiognomy by means of subtle hatching. It is most likely to be a first sketch made by Van Eyck in the presence of his sitter as there are several small pentimenti to be observed; nonetheless, it might possibly be a subsequent, detailed working drawing produced with the specific intention of aiding the production of the portrait painting within the workshop. In any case, the Dresden portrait was not drawn in one go, but was executed in several stages. In the first stage, Van Eyck used a silverpoint to draw the contours of face and bust. He then drew the curly hair and started to model the face using subtly hatched shadows. A horizontal line at the lower edge of the sheet delimits the bust of the sitter and already anticipates the frame of the painted portrait in this initial stage. In a second stage, he used a different silverpoint that contained a higher degree of copper. This was used to hatch the shadow

of the sitter’s head against an indistinct background. The drawing displays a distinctly pictorial approach in its tonal modelling and in the creation of a sense of space and depth through the suggestion of an indeterminate background. Finally, a goldpoint was used to correct and reinforce contours and placement lines and to document the appearance of the sitter by adding inscriptions. These explanatory notes on the left side of the off-white prepared paper are an extraordinary feature of the drawing. The artist meticulously translated the colour of the sitter’s skin, eyes and mouth into the Netherlandish dialect of his native Maas region. Whether one views these notes as mnemonics or simply acknowledges that written language was an integral part of Van Eyck’s creative process – and perhaps even a means of communication with his workshop – it is clear that these extensive colour notations were scrupulously followed during the application of the paint layer.30 The finished drawing served as a model for the painted portrait in Van Eyck’s workshop and its design was carefully transferred to the panel by means of a pair of dividers and a ruler. Indeed, markings from this process can be observed under magnification. This mechanical method of transfer made it possible to faithfully copy the portrait’s design from the smaller drawing onto the larger painting.31 Nevertheless, minor changes were made between the drawing and underdrawing stages, and further changes between the underdrawing and finished painting. During underdrawing, the head of the sitter was enlarged at the right and the position of the ear was altered accordingly. During painting, the proportion and perspective of the drawn likeness were altered, and the monumentality of the torso was greatly enhanced, as noted by Stefanie Buck and Thomas Ketelsen.32 When comparing the drawing with the underdrawing it can be seen that even some of the more minor features of the drawing have been meticulously transcribed to the underdrawing. These include the drawing’s dense hatching on cheek and chin, smaller parallel strokes around the nose and

remarks on character and functions in jan van eyck’s underdrawing of portraits

lips, lines around the neck, shadows in the ear. The difference in the drawing medium – thin brushes in the underdrawing as opposed to metalpoints in the drawing – accounts for the less subtle appearance of the lines, but it is worth noting that the underdrawn hatching on the right side of the face is loose in character and was more generously applied than that of the drawing. Comparing the sheet in Dresden with the portrait in Vienna and its underdrawing helps to clarify the function of the drawing. Contrary to what one might expect, the detailed modelling of the preparatory drawing in Dresden did not make the careful modelling and hatching in the underdrawing redundant. It continued to serve as guideline during painting, as some of the alterations that had been made during underdrawing were subsequently revoked. As a result, the finished painting is closer to the preparatory drawing than it is to its underdrawing in several details. This observation has consequences for our understanding of the form and function of underdrawing in portraits by Jan van Eyck. Stefanie Buck has suggested that the Dresden drawing ought to be seen as a preliminary study that represents a definable intermediary stage between the sitter and the final painted portrait in Vienna.33 I would like to argue that the use of detailed and worked up portrait drawings also must have been common working practice for Van Eyck and his workshop in the production of portraits. The detailed underdrawing of the Portrait of Margaret van Eyck – serving to increase the vividness of the likeness – does not exclude the use of a preliminary drawing, but would seem to be a prerequisite. Based on the available evidence of underdrawings in Van Eyck’s portraits, it can be argued with a high degree of likelihood that the painter and his workshop used drawings to record the physiognomy of the sitters and that elaborate drawings played an important part in the early stages of producing painted portraits. What we cannot be certain about is the number of preliminary drawings and subsequent drawing stages that were involved in making

251

portraits. Their number might actually have varied. Loose sketches made after life may have been transformed into elaborate working drawings. Alternatively, these sketches may have been used as aids to the production of detailed drawings, in turn informing the underdrawing and painting process. The existence of a few surviving highly finished portrait drawings produced in Bruges not long after Van Eyck’s death may point to the importance of such drawings for the portrait genre. These are either preliminary drawings that preceded the painted portraits or faithful copies after them.34 Van Eyck would also have adhered to specific portrait types in his commissioned work and these types were instrumental in establishing the decorum of the portrait representation. Van Eyck must have had several model drawings in his workshop that would have helped him determine the angle, scale and even the lighting of his subject. These model drawings with their established formulas could have been quickly adapted by the artist for different purposes. This might be the case of the drawing in Dresden and the donor portrait of the elderly Canon Joris van der Paele (fig. 15.13). While there are hardly any significant similarities between the portrait of Van der Paele and the panel in Vienna, there are interesting parallels to be made with the drawing. The positions of certain wrinkles seem to be common to both physiognomies, but more importantly, the similarities include the form of the lips, the characteristic placement of the neck with its double line, the position of the fur collar and the elaborate depiction of the ears. Indeed, this remarkable resemblance suggests that a common model was used, its angle slightly altered for each drawing. These common features become even more apparent when the drawing in Dresden is compared to the elaborate underdrawing of the head of Joris van der Paele (fig. 15.14).35 The resemblance between both portraits could be entirely or at least partly coincidental since they seem to follow certain patterns or portrait types that determined the representations. The similarities

252

till-holger borchert

Fig. 15.13 Jan van Eyck, Virgin and Child with Canon Joris van der Paele, 1436, oil on panel, 141 x 176.5 cm, detail, Joris van der Paele Bruges, Groeningemuseum (inv. no. 0000.GRO0161.I)

remarks on character and functions in jan van eyck’s underdrawing of portraits

Fig. 15.14 Virgin and Child with Canon Joris van der Paele, detail, Joris van der Paele, IRR

253

254

till-holger borchert

may also indicate the existence of a common model of ‘older men’ on the basis of which Van Eyck added individual features.36 In any case, the possible use of standard types and models would be entirely in agreement with the remarkable efficiency of Van Eyck and his workshop in recycling the same motifs in different pictorial contexts.37 The remarks and observations in this article regarding the form and function of underdrawing in Van Eyck’s portraiture underline the importance of preparatory drawing stages and shed new light on the production process of portraits in the Van Eyck workshop. N OTES 1 For the terminology of portraits, see Campbell 1990, pp. 1-4; Dülberg 1990, pp. 31-98. 2 A well-known example of a surviving early portrait is the Franco-Flemish Portrait of a Lady of around 1410 in the National Gallery of Art, Washington, DC, see Hand, Wolff 1986, pp. 90-97; other examples, copied after paintings as well as after sculpture, are recorded in the sixteenth century Recueil d’Arras, see Châtelet 2007. 3 On the suggestion that the word damoiselle may actually indicate the noble status of Margaret, see Reynolds 2000, p. 4. 4 See Janssens de Bisthoven 1981, p. 179 5 See Dunkerton, Morrison and Roy in this volume. 6 In 1769, the French painter and writer Jean-Baptiste Descamps recorded the panel in the archive of the Bruges painters guild in his Voyage Pittoresque, see Janssens de Bisthoven 1981, pp. 181, 190-191. 7 Janssens de Bisthoven 1981, pp. 181, 190. Descamp’s account of 1769 reveals that the painting was secured by means of chains: Tous les ans, le jour de la Saint Luc, on y voit un Tableau peint par Jean van Eyck (…), que l’on être le Portrait de sa femme: ce Tableau est attaché avec une châine et des cadenats, de crainte qu’il ne soit vole. On pretend que le pendant a été pris, sans sçavoir ce qu’il est devenue (Descamps, Voyage Pittoresque de la Flandre et du Brabant, Paris 1769, pp. 306-307. 8 The genre of conjugal or couples’ portraits seems to have already emerged in the Low Countries by the 1420s, as seen in sixteenth-century copies after portraits of Barthélémy Alatruye and his wife Marie de Pacey that were made in 1426, see Campbell 1990, pp. 211-212 and Stroo, Syfer-d’Olne 1996, pp. 87-95; on the genre of marriage portraits, see Hinz 1974, pp. 139-214. 9 Rupprich 1956, p. 168; Borchert 1998, p. 15. 10 Borchert 2012, pp. 213-232. 11 Lejeune 1972 suggested that the portrait may have been an anniversary or birthday present from the painter to his wife. 12 Borchert 2012, pp. 219-221. 13 Van Asperen de Boer 1995, pp. 81-84, esp. pp. 81-82.

14 Van Asperen de Boer, Ridderbos, Zeldenrust 1991, pp. 9-10. 15 Billinge 2000, p. 87. 16 See Billinge 2000, pp. 83-95 and Part 1 of this paper, by Billinge, in this volume. 17 On the function of the sheet as a preparatory drawing of the portrait in the Kunsthistorisches Museum in Vienna, see Ketelsen et al. 2005, pp. 170-175; Reiche et al. 2005, pp. 8-13; Dresden 2005, pp. 62-67; Borchert 2011, pp. 29-30. 18 Van Asperen de Boer 1979, pp. 149-150. 19 Campbell 1998, p. 218. 20 On the identity of the sitter, see Paviot 1995, pp. 210-215; Borchert 1997, pp. 558-562; Campbell 1998, pp. 220-222; Nys, Lievois 2002, pp. 1037-1057, Borchert 2007, pp. 35-36; London 2008, p. 98; Borchert 2012, p. 217. 21 Campbell 1998, p. 212. 22 Bruges 2011, p. 38; I would like to thank Ron Spronk and Elke Oberthaler for their kind permission to reproduce the IRR-images of the two portraits by Jan van Eyck. The paintings were documented on 17 June 2011 at the Kunsthistorisches Museum in Vienna in the Gemäldegalerie’s restoration studio by Rik Klein Gotink with Queen’s University’s OSIRIS infrared camera, acquired through the generous support of Dr Alfred Bader. The OSIRIS camera (Opus Instruments) is outfitted with an InGaAs sensor and a 6 element 150mm F/5.6 – F45 lens, and is operational in the 900-1700 nanometre range. Both panels were documented in a single (4096 ≈ 4096 pixel) image. 23 Pächt 1989, pp. 111-115. 24 Paviot 1990, pp. 89-93. 25 I will discuss the controversial identification of the portrait and question regarding its possible original function in more detail in a forthcoming article on the subject. 26 On Stevens, see Briels 1980, pp. 127-226. 27 On the biography of the Cardinal, see De Töth 1934; on his participation at the Congress of Arras, see Dickinson 1955, pp. 78-102. 28 See Briels 1980, p. 211; Stevens’s information is repeated in the inventories of the collection of Archduke Leopold, see Demus, Klauner, Schütz 1981, pp. 169-173. 29 On the underdrawing of the Vienna portrait, see Ainsworth 1989, pp. 8-9; Van Asperen de Boer 1990, pp. 8-12; Buck 2000, pp. 183-185; Reiche et al 2005, p. 11. 30 On the inscription, see Dierick 2000, pp. 79-82. 31 See Dresden 2005, pp. 62-67. 32 Buck 2000, pp. 183-185; Ketelsen et al. 2005, pp. 170-175. 33 Buck 2000, p. 184. 34 I plan to publish a systematic study of Early Netherlandish portrait drawings up to 1500 in the near future based on my research on Eyckian examples. 35 Périer-D’Ieteren 1982-1983, pp. 75-76; Fransen 2012, p. 122, see also Perier-D’Ieteren in this volume. 36 It is interesting to note that in 1851 Johann Passavant thought it possible to identify the Dresden drawing as a study for the Vienna portrait, which he believed represented Joos Vijdt, see Dresden 2005, p. 63, n. 11. 37 Reynolds 2000, pp. 5-6; Jones 2000, pp. 194-207; Borchert 2007, pp. 69-77.

Fig. 16.1 Jan van Eyck, Giovanni(?) Arnolfini and his Wife, 1434, oil on panel, 82.2 x 60 cm

16

The Speed of Illusion Lorne Campbell

ABSTRACT: Jan van Eyck was able to paint very quickly and very spontaneously. Learning from the discoveries and experiences of many generations of oil painters, he had a commanding knowledge of his vehicles and pigments. His manual dexterity and his acuteness of observation were two vital aspects of his genius. It is of course impossible to know how long he took to paint any of his pictures; but it is, I think, feasible to estimate how rapidly he executed certain details. Magnified images show not only the precision but also the speed of his brushwork, for example in the mirror, the dog and the pattens in the Portrait of Giovanni(?) Arnolfini and his Wife (NG186); and in the stubble and the eyes of the man in the Portrait of a Man (Self Portrait?) dated 21 October 1433 (NG222). There, the face could well have been painted in a single day.

—o— Haste can lead even the greatest of painters into error. The causes may be external – a deadline, an impatient and demanding patron; or internal – a feverish surge of creativity. I don’t know what distracted Bruegel’s attention when he was signing his London Adoration of the Kings. Perhaps he noticed something in the painting about which he was unhappy and he wanted to change it. His control faltered and he signed bvegel. Not noticing his mistake immediately, he finally went back and corrected it by scraping out some of the B, converting it into an R and adding a new B in a brown mixture different from that of the original signature.1 Even Dürer, in two of his most intricate engravings, the Adam and Eve of 1504 and the Melencolia I of 1514, forgot to reverse numerals and had to burnish out his errors.2

It is relatively easy to identify as mistakes errors that involve numerals and letters. If we look at the inscriptions on the three London paintings by Jan van Eyck, there are no errors but questions may still be raised. In two, Léal Souvenir3 and the Arnolfini portrait (fig. 16.1),4 there are no traces of guidelines for the inscriptions and none of them follows an absolutely straight line or is strictly horizontal or completely symmetrical. In Léal Souvenir, the Greek inscription and the Latin signature are fairly well centred and more or less horizontal; but leal sovvenir is not: its spacing is irregular and the repeated letters L, E and V are not identical. The letters, however, are so carefully painted that Jan obviously knew exactly what he was doing.5 The Arnolfini signature rises slightly from left to right.6 The date is pretty well centred but is not centred in relation to the signature. I think that Jan van Eyck felt no need to measure and rule lines and that he worked by eye to place all these inscriptions where they looked well and where they balanced his compositions. In the Portrait of a Man of 1433, however,7 which I believe to be a self-portrait, he used a ruling edge to draw guidelines for the inscriptions on the frame. als ich can is centred – it more or less had to be, because of the central peg in the frame, though the outer two pegs are not quite symmetrical. On the lower element of the frame, the central peg marks an approximate division between the signature and the date; but the ruling edge has slipped downwards above cccc and the line has been redrawn. The last two letters of octobris

258

lorne campbell

have been cramped to make them fit.8 I hesitate to call this a mistake, because the illusionisticallypainted letters are rendered with evident care, but I wonder why Jan allowed this. Perhaps he didn’t notice until the paint was too dry to remove safely from the gilding. It seems to me obvious, from the reflectograms alone, that Jan expended much time and effort on planning the Arnolfini portrait.9 It is perhaps less obvious that the process of painting need not have taken a hugely long time. Even the mirror, which seems at first such a miracle of observation and execution, could have been painted relatively quickly and indeed at one point Jan seems to have been painting at such speed that he dropped a brush loaded with blue paint onto the mirror frame. The mark would have been disguised; but time has made it visible once again.10 The reflected chandelier could have been painted in a matter of minutes by a genius with a good eye and a sure and practised hand (fig. 16.2). I imagine that Jan had been taught to paint from early in his childhood and that he had highly skilled teachers. He and they were learning from the experience of generations of painters in oil. Jan was able to paint quickly because he had mastered the properties of oil paint to such a degree that he could work with great economy. He could achieve effects of texture quickly and with complete success by dragging and feathering his paint;11 working it wetin-wet (fig. 16.7);12 blotting it with cloths,13 his fingers (fig. 16.3), or even the palm of his hand;14 making it bead by exploiting the tendency of dry oil paint to resist another application of paint bound in oil (fig. 16.3); teasing his paint with stiff brushes when it was almost dry;15 using the techniques of sgraffito.16 He probably knew many more techniques that we have not recognized and other techniques that may have fallen into disuse. It is, I think, fairly safe to assume that he knew all the possible shortcuts and used them with preternatural skill. Let us look at some of the passages where there is no underdrawing: for example the dog, which has no reserve and which is painted without any visible

preparatory work directly on top of the floorboards. Apart from its eyes (fig. 16.4) and nose (fig. 16.5), it is basically a study of relatively coarse hair in mixtures of black, vermilion, yellow and white. Some of the strands are nearly pure vermilion and the eyes have vermilion secondary lights (fig. 16.6). The shadow cast by the near back leg has been softened, its contour smudged with the artist’s finger or thumb – the work of seconds rather than minutes (fig. 16.3). The speed of execution is perhaps most apparent in the man’s pattens, again painted without any preparation on top of the floorboards (fig. 16.7); and the beads, similarly executed on top of the paint of the wall.17 The beads, their string and its tassel and the shadows that they cast have been reduced to their basic essentials. The highlights on the beads and their shadows have been dotted in with immense skill and, I am convinced, at great speed. The highlights are not always enclosed within the forms that they help to define. They themselves do not have very specific shapes. Jan has simply loaded a brush and, perhaps three or four times over, has allowed it to bounce – or perhaps dance – down the panel and to make a totally convincing, though not necessarily totally accurate, image.18 Many years of training and experience would have been necessary to develop this level of skill. Similarly, in the toe of the right patten, he has used the base colour of the floor and allowed the thinner wet paint of the brown to run into the thicker whitish paint of the light brown and has suggested in dark brown and with superb economy the mud staining the sole (fig. 16.7). In the self-portrait, the visible area of the face is just over 7.5 centimetres high and the measurement from tear duct to chin is a little over 5 centimetres. On this very small scale, however, Jan has painted with relative freedom. The stubble, for example, consists of series of dots and dashes applied with the points of small brushes. I believe that, if one were to count the dots, the number would fall short of the number of hairs in an average beard; even the pattern of growth seems to me

the speed of illusion

259

Fig. 16.2 Giovanni(?) Arnolfini and his Wife, photomicrograph of the reflection of the chandelier

Fig. 16.3 Giovanni(?) Arnolfini and his Wife, photomicrograph of the shadow cast by the dog’s back leg

Fig. 16.4 Giovanni(?) Arnolfini and his Wife, photomicrograph of the dog’s right eye

Fig. 16.5 Giovanni(?) Arnolfini and his Wife, photomicrograph of the dog’s nose

Fig. 16.6 Giovanni(?) Arnolfini and his Wife, photomicrograph of the dog’s tail

Fig. 16.7 Giovanni(?) Arnolfini and his Wife, photomicrograph of the toe of Arnolfini’s right patten

260

lorne campbell

a little eccentric, possibly because the dots are made to suggest not only the hairs of the stubble but also the relief of the face. The darker brown dots were painted before the lighter dots, which are mostly white. On the lit contour of the face, however, a very little blue is mixed into the white, so that some dots are pale blue and appear to be catching the light – as well as reflecting the background, which was once dark blue rather than black (fig. 16.8). The white of the near eye (fig. 16.9) is laid in with white mixed with minute quantities of red and blue. A very thin scumble of pink has been brought over this underlayer, which is, however, left exposed in four places to create the secondary lights. The veins are painted in vermilion into the wet scumble. The iris is ultramarine, fairly pure at its circumference but mixed with white and black towards the pupil. There are black flecks near the circumference and the pupil is painted in black on top of the blue of the iris. The principal catchlights, applied last, are four spots of lead white, one on the iris and three on the white, where they register with the four secondary lights to create a glistening effect. This may sound complicated but it is a fairly straightforward and economical way of painting. It requires, of course, astonishing dexterity; but it need not have taken very long. The picture is dated 21 October 1433, which was a Wednesday and St Ursula’s day. I wonder whether it is possible that this, and similar dates on other portraits by Jan, indicate the days on which the faces were painted or at least finished, so that they record the exact days on which the likenesses were taken. Other portraits, including the London Arnolfini, were not given exact dates and I can think of possible reasons. Is it conceivable that the self-portrait was painted in one day? I am reminded of Van Mander’s story of Anthonis Mor’s portrait of Hubert Goltzius, recounted in his biography of Hubert. The portrait was made in one hour, or at least in a very short time, and I can well believe that the face was very quickly painted.19 I haven’t examined under the microscope the Brussels portrait of Hubert Goltzius; but I have looked at the

Fig. 16.8 Jan van Eyck, Portrait of a Man (Self-Portrait?), 1433, oil on panel, 33.1 x 25.9 cm, photomicrograph of the stubble around the man’s mouth London, National Gallery (NG 222)

Fig. 16.9 Portrait of a Man (Self-Portrait?), photomicrograph of the man’s left eye

similar, though very damaged, portrait of an unidentified man in the National Gallery, London. His right eye is reasonably well preserved. In the irregular black shape of the pupil Mor has left exposed a triangle of dark brown underpaint which, with the touch of pure white, makes a secondary light and a catchlight. Rapid brushstrokes of lighter grey indicate the radial striations of the iris. The impasted brushstrokes of pure white in the wavy lines of the collar move in serpentine shapes. Working while the dark paint of the jacket was still wet, Mor has directed his brush, loaded with white, to pick up some of the darker colour along the outer contour and has made it swirl into the white to create an illusion of layers of semi-transparent fabric.20 Such an economy of means was feasible only after

the speed of illusion

the acquisition of extraordinary skills and only, of course, when those skills were practised by a painter of genius. Though it is impossible to say how long Jan van Eyck took to paint his portraits, I am prepared to hazard a guess that, in the case of the selfportrait, he may well have painted the face in one day, Wednesday 21 October 1433.

261

N OTES 1 Campbell 2014, p. 184, fig. 13. 2 Dodgson 1926, pp. 54, 94. 3 Davies 1954, pp. 132-135, pls CCCVII-CCCXIII; Campbell 1998, pp. 218-223 4 Davies 1954, pp. 117-128, pls CCLXXXI-CCCI; Campbell 1998, pp. 174-211. 5 Davies 1954, pls CCCIX, CCCX, CCCXII. 6 Davies 1954, pl. CCIC; Campbell 1998, p. 2; Dunkerton, Billinge 2005, p. 44. 7 Davies 1954, pp. 129-132, pls CCCII-CCCVI; Campbell 1998, pp. 212-217. 8 Davies 1954, pl. CCCIV. 9 Campbell 1998, p. 177. 10 Campbell 1998, p. 185. 11 Campbell 1998, p. 2; Dunkerton, Billinge 2005, p. 52. 12 Compare Dunkerton, Billinge 2005, pp. 23, 42. 13 Davies 1954, pl. CCLXXXIX. 14 Compare Dunkerton, Billinge 2005, p. 59. 15 Davies 1954, pl. CCLXXXIV, where the fur lining of the woman’s dress is so painted. 16 Campbell 1998, p. 184 fig. 16; Dunkerton, Billinge 2005, p. 57. 17 Campbell 1998, p. 185. 18 Campbell 1998, p. 183 fig. 14; Dunkerton, Billinge, 2005, p. 43. 19 Van Mander 1604, fol. 258v; Campbell 1990, pp. 64-67, 263 n. 44. 20 Campbell 2014, pp. 564-565.

Fig. 17.1 La Fontaine de vie, huile sur panneau, 181 x 119, Madrid, Musée du Prado, ici attribué pour partie à Jan van Eyck

17

Les autoportraits présumés de Jan Van Eyck et la date approximative de sa naissance Pierre Colman

ABSTRACT: The Presumed Self-portraits of Jan van Eyck and his Approximate Date of Birth The only self portrait by Jan van Eyck that has come down to us can be spotted in the Museo del Prado’s Fountain of Life (The Fountain of Grace and the Triumph of the Church over the Synagogue). This fascinating and much-discussed painting has been underestimated due to the presence of retouching and overpaint. As the painting should be dated 1424 at the latest, and not around 1455, and as the figure is likely in his mid-forties, Van Eyck must have been born around 1380. The same conclusion is reached if the Man in a Red Turban (Portrait of a Man (Self Portrait?)) (NG222) is considered a self portrait, but this identification is not credible.

—o— Un seul autoportrait de Jan van Eyck est venu jusqu’à nous. Il est à découvrir dans la Fontaine de vie du Prado (fig. 17.1). Le visage est entièrement de sa main. De sa main aussi le bassin de la fontaine, une pure merveille (fig. 17.2), au moins autant que celui de l’Agneau Mystique. Et de sa main encore le dessin sous-jacent, qui montre déjà sa stupéfiante aptitude à retravailler ses ébauches comme en se jouant.1 Le reste est d’un sous-fifre, sans doute l’un des compagnons qu’il avait lorsqu’il servait Jean de Bavière en Hollande. Le panneau a été commandé dans la perspective d’une échéance qui tombait en 1424, celle des ordonnances colonaises sur les juifs.2 Le donneur d’ordre doit être un citoyen de l’opulente métropole que Jan avait choisie en vue de parachever sa formation, une vingtaine d’années plus tôt. Ainsi se résument les

convictions que je me suis forgées au sujet de ce tableau si discuté,3 qu’il faudrait débarrasser de ses repeints et de ses surpeints. Un groupe bien ordonné de chrétiens en adoration s’oppose à un groupe de juifs en proie à la plus vive agitation. Aucun n’a fait l’objet d’une identification assurée. Le pape, selon moi Martin V, est comme de juste en tête. À sa droite, debout comme lui, un cardinal, un évêque, un homme de haute taille (un ajout ?) et un abbé qui tient une crosse. Devant eux, quatre hommes à genoux : un empereur, un roi, un homme jeune coiffé de noir (surpeint lui aussi ?) et un personnage corpulent vêtu de rouge vif comme le pape.4 Très en vue au premier plan, il est de taille plus modeste que les trois adorateurs placés devant lui. C’est assurément le donneur d’ordre.5 Derrière lui se tiennent deux hommes, distinctement en retrait des autres membres du groupe ; leur statut est de toute évidence subalterne. Ce sont, à mon idée, le concepteur et l’exécutant du tableau. Le second se tient debout en toute dernière position. Il est coiffé d’un chaperon à bourrelet de couleur noire6 et sobrement accoutré d’un justaucorps foncé garni de fourrure.7 Il lève d’ostensible façon la main gauche (repeinte ou surpeinte ?). Il l’approche du front de son voisin, à mon sens pour évoquer leur collaboration.8 Il est le représentant de tous ceux qui travaillent de leurs mains, comme le concepteur présumé est celui des intellectuels et le donneur d’ordre présumé celui des aristocrates, tout comme

264

pierre colman

Fig. 17.2 La Fontaine de vie, detail : le bassin de la fontaine, ici considéré comme entièrement du pinceau de Jan van Eyck

les autoportraits présumés de jan van eyck et la date approximative de sa naissance 265

le roi, le cardinal, l’évêque et l’abbé sont ceux de leurs pareils. Ils forment avec le pape et l’empereur un microcosme de la chrétienté. L’homme au chaperon noir montre un visage extraordinairement prenant.9 Il a le regard baissé comme s’il était perdu dans ses pensées. Il fait mine de se tenir en dehors de la scène, en somme. Peuton parler d’Humilitas gestus, au sens où l’entend Justus Müller-Hofstede ?10 Rien, quoi qu’il en soit, de la Selbstverkleinerung que Jan s’imposera ou se verra imposer dans le Double portrait de la National Gallery et dans La Vierge au chanoine van der Paele. A-t-il pu se mettre en valeur à ce point ?11 Oui, car l’éclat de Jean de Bavière rejaillissait sur lui. En se mettant ainsi en scène, avec une relative discrétion, au prix d’une compensation financière peut-être, le peintre pouvait tout à la fois participer aux bénéfices spirituels et assurer sa notoriété. Il entretenait avec son client, rien n’empêche de le supposer, des relations privilégiées. Il allait bien en avoir avec Giovanni (?) Arnolfini, on est fort tenté de le croire,12 et avec Philippe le Bon, on ne s’en ébahira jamais assez. Un arrangement de ce genre-là est à envisager dans un autre cas, celui du célèbre Retable du Saint Sacrement de Louvain.13 Dirk Bouts y a installé le portrait de Rase van Baussele, en place d’honneur bien entendu. Il y a mis aussi le sien propre,14 avec un subtil mélange de modestie et de fierté : encore qu’il se tienne à l’écart, il est fort en évidence. Il s’abstient de joindre les mains et de baisser les yeux, à l’exemple de Rase. Il reste étranger à la Cène. Il ne porte pas plus que l’homme au chaperon noir le regard sur le spectateur.15 Ils ont un comportement analogue. Ces deux œuvres bonnes à confronter, une quarantaine d’années les séparent. Dans l’intervalle, le statut social de l’artiste s’est grandement amélioré. C’est avec Jan van Eyck que la voie s’est ouverte. Le visage dans lequel je reconnais le sien a été rendu avec une attention impitoyable aux irréparables outrages du temps. Les yeux sont fatigués. Le cou est flétri (fig. 17.3). L’âge de l’homme se situe entre quarante et soixante ans, de l’avis de celles et

Fig. 17.3 La Fontaine de vie, détail : le visage de l’homme en noir, ici présenté comme un autoportrait de Jan van Eyck

ceux à qui je l’ai demandé ; l’accord s’est fait, après discussion, sur le milieu de la quarantaine. Si l’on admet qu’il s’agit bien du « prince des peintres » et qu’il a peint là son propre visage en 1424 au plus tard, sa naissance est à situer vers 1380, et non pas vers 1390-1400, comme on le ressasse, le plus souvent avec la plus grande assurance.16 La conviction a des racines pourries : une « vieille tradition », diffusée par les humanistes du XVIe siècle. Selon elle, on doit reconnaître Jan van Eyck dans celui des Juges intègres de l’Agneau Mystique qui a sur les épaules un rosaire. Plus personne ne lui accorde crédit.17 Erwin Panofsky s’est singularisé en situant la naissance aux alentours de 1390 au plus tard.18 Albert Châtelet en juge tout autrement : « aucune

266

pierre colman

raison n’impose vraiment de la situer nettement avant 1400, même si nombre d’auteurs songent plutôt à 1390 » ; or, il situe en 1419, voire en 1412 des miniatures qu’il attribue à Jan, non sans hardiesse.19 C’est dans l’Homme au turban qu’il faut voir un autoportrait, maints auteurs en ont acquis la conviction sans disposer d’aucun argument décisif.20 S’y accrocher ne permettrait pas de maintenir la date que j’ai l’audace de contester.21 Si Jan avait vu le jour entre 1390 et 1400, il aurait eu entre 33 et 43 ans en 1433, date du portrait. Le modèle est nettement plus âgé : la paupière de l’œil droit est tombante ; des rides marquées s’étagent sous l’oeil gauche. La supposition gratuite d’un vieillissement prématuré est la seule échappatoire. Lucas de Heere et Marcus van Vaernewyck veulent que Jan soit mort jeune. Leurs témoignages n’en font qu’un, car le second dépend assurément du premier. Ils sont à la fois tardifs et vagues. Et le fantôme de Raphaël plane sur eux. La naissance d’Hubert est, quant à elle, située d’ordinaire vers 1366 ou vers 1370, une vingtaine d’années avant la date admise pour celle de Jan. Pareil écart sort de la vraisemblance. Ceci dit, on se base derechef sur de fragiles témoignages et l’on n’a pas la preuve formelle que Hubert et Jan étaient frères. L’enquête m’a fait reprendre le chemin du Centre des Primitifs flamands, needless to say. Je m’y suis senti chez moi. Un peu parce que j’ai eu quelque temps le redoutable honneur d’assumer la présidence. Beaucoup du fait de l’accueil qui m’a été réservé. Mes remerciements sont chaleureux à proportion.22 Le mot de la fin, je me plais à l’emprunter à Catherine Reynolds : « It is only through constant reexamination and questioning that understanding increases ».23

Annexe : Discussion paper posté sur kunstgeschichte e-journal

Il faut sortir résolument de l’ornière creusée de trop convaincante façon par Josua Bruyn. Le tableau n’a pas été commandé par le roi de Castille Enrico IV, qui l’a offert à un couvent de Ségovie entre 1455 et 1459. Il l’a été par un Colonais qui se proposait de convertir les juifs et nullement de les vouer à l’exécration ; et cela en 1424 au plus tard. Ayant très vite perdu sa raison d’être première, il a été vendu et a pris le chemin de la péninsule ibérique. Bien loin d’être postérieur à la mort du peintre, il est de la sorte antérieur à l’Agneau Mystique et à tous les Van Eyck indiscutablement datés.24

Le donneur d’ordre présumé portraituré au milieu du groupe des chrétiens porte un vêtement qui oriente clairement vers le monde germanique et donne les alentours de 1425 comme terminus ad quem. La coiffure en bourrelet, quant à elle, est repérable dans la célèbre miniature intitulée, peu judicieusement, , qu’il faut dater, selon moi, de 1417-1418 : elle est portée par le cavalier dans lequel je reconnais Guillaume, l’aîné des deux bâtards de Jean de Bavière.25

On ne saurait se fonder sur la récurrence de divers motifs dans différentes œuvres de Jan Van Eyck pour les ranger dans l’ordre chronologique.

Le dessin sous-jacent prouve que le panneau du Prado n’est pas une copie : aucun copiste ne dessine ronde une fontaine qu’il a pour tâche de peindre octogonale. Libre à souhait, travaillé par les changements de composition, il ne diffère pas de ceux de Jan. Il est assurément de sa main.

Le visage de l’homme sobrement vêtu de noir l’est aussi ; il a tout d’un autoportrait. Et l’est aussi l’eau de la fontaine où nagent les hosties, une pure merveille. En revanche, la lumière étale, dénuée de subtilité, ne saurait l’être, ni les têtes présentées en profil absolu. L’exécution, réputée faible, est en fait

les autoportraits présumés de jan van eyck et la date approximative de sa naissance 267

très inégale. Selon moi, le tableau a été peint par le plus capable des membres de l’atelier au départ d’une ébauche mise en place par son chef, qui l’a parachevé par endroits. L’exécutant pourrait bien se confondre avec le « knecht » dont le salaire est mentionné dans les comptes de Jean de Bavière.26 Ce prince ne laissait-il à son peintre qu’une marge de liberté très étriquée, sans doute. Peut-être le donneur d’ordre entendait-il limiter la dépense.

L’enlèvement des surpeints évoqués trop laconiquement pourrait bien réserver d’heureuses surprises.

Merci à Volker Herzner et à Bart Fransen, avec qui j’ai eu des échanges répétés aussi agréables qu’enrichissants. N OTES 1 Voir Billinge 2000 et Silva Maroto 2006 pour une confrontation éloquente à souhait. À mon bien vif regret, j’ai dû renoncer à donner une reproduction du bassin en IR, les courriels adressés au Prado étant restés sans réponse. Et de même pour le copyright. 2 Le tableau n’a nullement été peint vers 1455, comme le voulait Josua Bruyn, dont les convictions ont fait tache d’huile, en dépit des objections fortes d’Otto Pächt, de Charles Sterling et de Volker Herzner : Colman 2009, p. 76 et 130. 3 Colman 2009, p. 73-78 et 130-131. Comptes rendus : Henderiks 2009, Châtelet 2010, Elsig 2010, Frémelet 2011 et Martens 2011 et 2012. Dans le sien, Albert Châtelet m’adresse une critique qui m’a fait remettre mon ouvrage sur le métier. C’est « sans argument précis », estime mon pugnace collègue, que je propose de reconnaître dans le personnage vêtu de noir un autoportrait in assistenza. Il me reproche aussi de ne pas « tenir compte des identifications proposées pour les personnages de la société chrétienne ». Il m’a fait connaître, et je l’en remercie encore, un texte (Châtelet 1985, p.  232) dans lequel il invoque des ressemblances indétectables pour moi. Il est revenu sur le sujet encore plus hardiment voici peu (Châtelet 2011, p. 200-202 et 296-297, P/7). Il cultive des convictions qui ne sont en rien conciliables avec les miennes. Il situe le tableau vers 1445-1450 et l’attribue à Jean de Pestivien. Bart Fransen a donné au moment où mon Mémoire est sorti de presse une synthèse nourrie à souhait sans avoir l’ambition de mettre de nouvelles propositions sur le tapis : Fransen 2009. Mon petit livre a échappé à l’attention de Luc Dequeker 2011 ; le théologien louvaniste reste dans l’ornière creusée par Josua Bruyn (Bruyn 1957, p. 73-83). En réaction à un envoi de Volker Herzner, j’ai brièvement résumé et amplifié mes arguments dans le journal en ligne Kunstgeschichte en 2011 (voir annexe). 4 C’est la couleur « escarlatte vermeille » que les princes souverains ne parvenaient pas à se réserver : Madou 1986, p. 59 ; Piponnier, Mame 1995, p. 89 et 202 ; Monnas 2000, p. 147. 5 Un chercheur colonais parviendra quelque jour à le sortir de l’anonymat, je ne renonce pas à l’espérer. En tout cas, ce n’est ni Hubert van Eyck (Crowe, Cavalcaselle 1857, p. 98), ni Philippe le Hardi (Post 1922, p. 122.), ni Jean de Bavière, comme le prétendait Jean Lejeune (Liège 1968, cat. exp., n° 164, p. 162-163), sans voir que le collier énigmatique posé sur ses épaules diffère totalement de celui de l’ordre de saint Antoine en Barbefosse (voir Colman 2009, p. 112). Ce

n’est pas davantage le roi René (Châtelet 2011, p. 200-202 et 296297) : la ressemblance générale n’est pas niable, mais sourcils, bouches et mentons sont différents. 6 Le chaperon n’est pas du tout en désaccord chronologique avec le costume de l’homme en rouge : il est parfaitement comparable avec celui qu’a sur la tête Jean de Touraine, passé de vie à trépas en 1417, d’une part (De Vos 2011, p.  197-200) et d’autre part avec celui de l’énigmatique Timotheos, dont le portrait est daté de 1432. encore qu’il dégage le front là, alors qu’il est ici enfoncé bas. 7 L’orfèvre Jan de Leeuw en porte un du même genre (Châtelet 2011, p. 176 et p. 281, IX/5), mais aussi Dirk Bouts, si c’est bien lui (Périer-D’Ieteren 2005, fig. 9). 8 Il replie un peu les doigts comme s’il se proposait de saisir un objet. Selon Albert Châtelet, qui veut reconnaître en lui Jacques Cœur dans le rôle de « commanditaire », il désigne les agenouillés « comme s’il nous les présentait » (Châtelet 2011, p. 200). 9 Quatre des autres visages, vus en profil absolu, sont bien peu eyckien. Celui du pape, vu de face, manque d’accent. Les autres sont de meilleure venue, et pas peu. Celui du roi se singularise par un rictus qui m’intrigue fort. Crowe et Cavalcaselle identifiaient l’homme avec celui des Juges intègres dont la tradition faisait en leur temps, il y a plus de 150 ans, un autoportrait de Jan van Eyck ; ils admettaient que la ressemblance est loin d’être probante : Crowe, Cavalcaselle 1857, p. 98. Rejet de l’idée : Post 1922, p.  122. Rappel expéditif sans référence : Châtelet et Faggin 1969, p. 83. 10 Müller-Hofstede 1998, p. 38-68. 11 Châtelet 2011, p. 200. 12 Campbell 1998, p. 201. 13 La preuve s’en cherche en vain dans les informations dont on dispose à son sujet, exceptionnellement abondantes : Comblen-Sonkes 1996, p.  18-19 et 178 ; Müller-Hofstede1998, p.  58 ; Martens 1999, p. 401-403 et 411 ; Périer- D’Ieteren 2005, p. 36 et 124. Le triptyque dont le panneau central contient le portrait présumé de Bouts avait sa place au-dessus d’un autel. Le panneau de la Fontaine de vie l’a eue, lui, en Espagne, au voisinage d’une fontaine, en accord avec le thème traité. Peut-être l’avait-il à l’origine déjà. 14 Van Gelder 1951, p. 51-52, Cardon 1998, p. 521-523 et Smeyers 1998, p. 43 rejettent cette conviction, pourtant répandue. Ils veulent reconnaître, eux, les quatre délégués de la confrérie du Saint Sacrement. Ils ne s’inquiètent aucunement de la spectaculaire inégalité qu’organise la mise en scène. 15 On peut faire le même constat devant de très nombreux autoportraits : Calabrese 2006, p. 162. Comme témoins pour le XVe siècle, dans cette somme, seulement deux dessins, attribués l’un à Dirk Bouts et situé vers 1460-1470, l’autre à Gentile Bellini et daté de 1480 (fig. 134 et170). Pour ce qui est de Bouts, l’attribution est confirmée, mais non l’interprétation : Périer-D’Ieteren 2005, p. 120-121. Plus probant : l’autoportrait de Hans Memling dans le triptyque de John Donne (Müller-Hofstede1998, p. 62). 16 La fourchette 1390-1400 est proposée dans De Patoul, Van Schoute 1998, p. 288 et la date de 1395 environ dans Van Buren 1996, p. 705), sans argumentation. 17 Cela a été dûment souligné voici un siècle : Friedländer 1915, p. 129. Il s’agit à peu près certainement de Philippe le Bon : Post 1921, p. 67-81 ; Dequeker 2011, p. 152. À l’issue de mon exposé, il m’a été demandé de préciser ce que Jan avait fait, à mon avis, au début d’une carrière que je propose de faire commencer au moins dix ans plus tôt. Si j’ai vu clair (Colman 2009, p 58-73 et 79-101), avant d’être au service de Jean de Bavière de 1422 au plus tard à 1425, il servait son frère aîné Guillaume en 1417, et ce n’est assurément pas le fait d’un débutant. Vers 1410, il avait peint le triptyque Norfolk, à Cologne, où il avait vécu une bonne dizaine d’années, et d’abord pour y parachever son apprentissage. De son côté, Châtelet le voit à Paris vers 1412-1419. 18 Panofsky 1953, p.  178 (not later than ca 1390) ; voir aussi p. 427, n. 178/1 et 2. Il ne s’est nullement arrêté à 1395-1400, Dhanens 1980, p. 19. Belle prudence sur ce point-là : Dhanens 1977, p. 23.

268

pierre colman

19 Châtelet 2000, p. 89 ; Châtelet 2011, p. 29-34, 42 et 222. 20 Comme Martin Davies et Lorne Campbell l’ont exemplairement reconnu l’un comme l’autre. Voir Colman 2012. Avant de m’expliquer là de façon approfondie au sujet de NG 222, je passe en revue les autres candidats. 21 Observation faite sans commentaire : Campbell 1998, p. 174. Les deux savants auteurs qui le notent aussi (Panofsky 1953, p. 198 et Van Buren 1996, p. 709) ne reculent pas pour autant la date de naissance. Voir encore Belting, Kruse 1994, p. 151. 22 Merci aussi à Volker Herzner pour son amicale assistance et ses commentaires critiques.

23 Reynolds 2000, p. 12. Catherine vient de mettre en évidence le statut prééminent de la peinture : Reynolds 2012. Le même recueil savant comporte une contribution dont l’illustration aurait gagné à s’enrichir d’une reproduction du bassin : Bol, Lehmann 2012. La légende de la fig. 17/3 attribue à un miniaturiste parisien une enluminure célèbre qui appartient à l’art mosan. Qu’on me pardonne de m’en étonner. 24 Colman 2009, p. 73-78 et 130-131. 25 Ibid., p. 109 et 118. 26 Ibid., p. 121-122.

Fig. 18.1 Jan van Eyck, Portrait of Margaret van Eyck, 1439, oil on panel, 32.6 x 25.8 cm, Bruges, Groeningemuseum, (inv. no. 0000.GRO0162.I), photographed in visible light, during varnish removal

18

Pigments, Media and Varnish Layers on the Portrait of Margaret van Eyck Jill Dunkerton, Rachel Morrison and Ashok Roy

ABSTRACT: During the cleaning and restoration undertaken at the National Gallery of the Portrait of Margaret Van Eyck (Bruges, Groeningenmuseum), it was possible to take a few small samples from the edges of the painting, the frame mouldings and the marbled decoration on the reverse. In spite of the limited possibilities for sampling, the samples proved to be exceptionally instructive, supplying information not only about the pigments, paint medium and layer structure, but also about Van Eyck’s use of unpigmented organic layers, which appear as surface and intermediate layers on both portrait and marbled reverse. The initial results of the technical examination have been presented on the websites of the National Gallery and the Groeningenmuseum, but since then the samples have been examined by further methods, including SEM-EDX, ATR-FTIR (attenuated total reflectionFTIR) and py-GC-MS (pyrolysis gas chromatography mass spectrometry). The results are presented in this article.

—o— In 2008 the Groeningemuseum generously agreed to lend Jan van Eyck’s great portrait of his wife, painted in 1439, to the Renaissance Faces: Van Eyck to Titian exhibition which took place first at the Prado, Madrid, and then at the National Gallery in London.1 Before the loan was agreed, the lenders expressed misgivings about the general appearance of the painting, with its discoloured varnish, patchy grey toning layers on the veil and numerous small surface accretions, which did indeed make the work appear shabby and the sitter seem older than her thirty-three years. Any cleaning of the portrait before the Madrid and London exhibitions was out

of the question, but, since the exhibition schedule at the Groeningemuseum meant that the permanent display areas would not be available for a considerable period, it was arranged that the Portrait of Margaret van Eyck would remain in London for treatment in the Conservation Department at the National Gallery. An important factor in the decision was the fact that we have three celebrated portraits by Van Eyck in our collection, as well as two more by his followers, and that we hold data on them collected by our Scientific Department.2 As well as their experience in the identification and analysis of pigments, the Scientific Department has a long and pioneering tradition in the analysis of paint binding media and other organic materials, always an important issue in the study of Eyckian painting. Although it was claimed that the portrait had not been touched by restorers since the seventeenth century,3 this is unlikely to be the case. The discoloured varnish, toning layers, and also the previous restoration, including the largely repainted black background, were easily soluble in normal cleaning solvents. The cleaning, while delicate, was entirely straightforward. Moreover, the condition of the painting is very much better than has sometimes been believed – in the Corpus for the Groeningemuseum it is described as being in no more than ‘Fairly good state, notable wearing in several areas, especially the shadowed parts of the

272

jill dunkerton, rachel morrison and ashok roy

face and neck etc.’4 In fact, the only significant problem was the chipping of the edges and corners of the distinctive craquelure, most marked in the black background and to a lesser extent in parts of the white veil. Large parts of the work are in an almost perfect state of preservation. Following restoration, the portrait was displayed at the National Gallery for several months in the company of the Gallery’s own works by Van Eyck. Simultaneously an account of the treatment was put up on the websites of both the Groeningemuseum and the National Gallery.5 This included the initial results of the scientific and technical examination. The Van Eyck Studies colloquium presented an opportunity to elaborate on a few aspects of those discoveries by re-examining the very few paint samples taken (all from the edges), including by methods such as atr-ftir (attenuated total reflection-Fourier transform infrared imaging),6 a technique obtained by the Scientific Department only after the restoration of the portrait and the subsequent web publication. As is usual practice, the progress of varnish removal was regularly evaluated by ultraviolet fluorescence, here documented at the point when the presumably nineteenth-century soft resin varnish had been removed from the central and right part of the painting surface (figs 18.1, 18.2). The varnish that remains on the left side has the strong greenish-yellow fluorescence typical of an aged natural resin varnish such as dammar or mastic. On the cleaned side (the little patches at the edges are where paint consolidation needed to be carried out before varnish removal) there remains a thin fluorescent layer, interrupted in places, but clearly visible on Margaret’s face, headdress and much of her red dress. This is clearly very old; an indication that this might be the remains of an original surface coating that had survived past cleaning is the remarkable state of preservation of most of the fine detail on the painting.7 The only disruption to the image is that caused by the cracking. Just as interesting, however, are those areas which do not exhibit fluorescence. The interpreta-

tion of ultraviolet images can be problematic but here the quenching of any fluorescence by the copper green pigments contained in the sash is typical. More surprising is the dark appearance of the red dress towards the lower edge and in a curved area in the lower right corner. An initial assumption might be that this area had simply been cleaned more in the past, losing the old surface coating. Alternatively, the dull areas could be taken as old retouching. A single paint sample, taken from the lower edge of the painting where there is damage and loss as a result of the cracking and shrinking apart of the oak panel and the attached frame mouldings, provided another explanation as well as a considerable amount of information about Van Eyck’s technique. Unfortunately the sample split and had to be mounted as two separate cross sections (figs 18.3, 18.4). Immediately over the white chalk ground is a thin layer of vermilion. The fragment of sample used for medium analysis by gc-ms identified heatbodied linseed oil in this layer,8 which was brushed on freely as an unmodulated base colour, extending slightly onto the frame mouldings. The next layer in the structure appears at the bottom of the upper cross section and consists of a thin warm grey layer containing black, lead white and lead-tin yellow.9 The presence of a more substantial layer of similar composition in a sample of the black background, also taken at the conjunction between frame and painting, supports the probability that this is paint from the greyish marbled frame which in turn slightly overlapped onto the painted area. The frame decoration was therefore applied as one with the painting of the portrait. The folds of Margaret’s red dress are modelled over the vermilion base layer in typically Eyckian fashion with layers of translucent red lake glaze. Signs of blotting, probably with a cloth pad, are visible in several places. A small amount of bone black was included with the red lake,10 and the deepest folds were emphasised with a final touch of ultramarine, now blanched as is so often the case in works by Van Eyck. sem-edx analysis

pigments, media and varnish layers on the portrait of margaret van eyck

Fig. 18.2 Portrait of Margaret van Eyck, photographed in ultraviolet light, during varnish removal

273

274

jill dunkerton, rachel morrison and ashok roy

50 μm

Fig. 18.3 Cross section containing the lower portion of the sample taken from the red dress at the bottom edge, photographed in visible light. The cross section contains only the ground and the opaque red underlayer

50 μm

0.14

0.08 0.06

2926

0.10 3268

Absorbance Units

0.12

1394

1533

1644

50 μm

0.04 0.02 0.00 3500

3000

2500 2000 Wavenumber cm-1

1500

1000

Fig. 18.4 Cross section containing the upper portion of the sample taken from the red dress at the bottom edge, photographed in visible and ultraviolet light. The cross section contains a pale grey layer, probably related to the decoration of the frame, followed by several layers of red lake glaze. The ATR-FTIR image, produced by integration of the amide II band between 1572 and 1498 cm-1, is illustrated at the lower left next to a detail from the ultraviolet image of the cross section showing the analysed area. The red parts of the ATR-FTIR image correspond to areas rich in protein and match well with the fluorescent interlayer visible under ultraviolet light. An FTIR spectrum extracted from the intermediate layer confirms the presence of protein and is illustrated at the lower right

pigments, media and varnish layers on the portrait of margaret van eyck

also confirmed the presence of a little ultramarine in the red lake glaze as well as a tiny amount of glass.11 In the cross section three layers of red lake glaze are visible. They are most clearly seen in the ultraviolet light image. The lowest has a paler, pinkish orange fluorescence, distinctly different from the two subsequent layers, which is partly due to the fact that more lead white was included with the initial application of lake. The upper layers of lake glaze are separated by a thin fluorescent layer which is not quite continuous through the whole sample, to which we shall return. hplc analysis of a sample from the red glaze identified two red dyestuffs. The major component is kermes from the scale insect, Kermes vermilio Planchon, but a smaller quantity of madder, Rubia tinctorum L, was also detected. Examination of pigment recipes, and indeed the previous analyses of lake samples from fifteenth- and sixteenth-century paintings at the National Gallery, suggests that the source of the dye for making both lakes is textile shearings.12 It is just possible that both dyestuffs were incorporated into the same lake pigment. However, examination of the cross section in ultraviolet light shows that some particles have a slightly different fluorescence than others, suggesting that two separate lakes were used. The presence of protein, detected by ftir, within some of the red lake particles indicates that at least one of the dyestuffs was extracted from a woollen textile. Based on the results of previous analyses, it seems most likely that this is the case for the madder lake.13 The medium of the red lake glaze is again heatbodied linseed oil.14 This is also likely to be the medium of the black background, but here the sample was contaminated with beeswax from past consolidation of the panel and frame and so it was not possible to distinguish precisely the oil type, but it too is heat-bodied.15 Given the similarity of composition of the red lake layers above and beneath the strongly fluorescent unpigmented layer in the cross section – and this also includes the presence of zinc throughout16

275

– there can be little doubt that all the paint is original. The days of believing that all layers above any hypothetical varnish layer must be later repaint are long past. Interspersed varnish layers, especially within layers of red lake or green glaze, have been reported on works that predate Van Eyck, including areas of red lake paint from the Westminster Retable, painted in England (in oil) in about 1270,17 and on later Netherlandish and also Italian paintings. Since the publication of the restoration of the Portrait of Margaret van Eyck on the websites, when we suggested that the fluorescent interlayer might be an oil resin varnish, we have been able to apply atr-ftir analysis to this cross section. Unexpectedly, the organic layer turns out to be proteinaceous, the type of protein as yet unidentified. Contamination as a result of flake laying at the edge seems unlikely as there were no signs of cleavage within the red lake layers. It cannot be excluded that the layer represents another overlap with the frame decoration, but gc-ms analysis confirms that the medium of this is also heat-bodied linseed oil. A final application of glue size, for instance, to the frame would not be impossible. However, this would not explain the difference in surface fluorescence between the main body of the painting and the apparently reworked area in the lower right corner. For the time being, at least, we have to work with the hypothesis that Van Eyck applied a thin proteinaceous coating to his finished, or rather almost finished, painting, perhaps even as a preliminary to a later varnishing with an oil resin varnish, long since removed in a past cleaning. Moreover, the contraction and cracking of the layer, so evident in ultraviolet in the cross section, a feature that appears odd for an oleo-resinous layer, raises the possibility that the application of such a layer might explain, at least in part, the very distinctive craquelure of the paint surface.18 That Van Eyck did indeed use protein-based media in the execution of the portrait is confirmed by the examination of the decorated reverse (fig. 18.5). Here cleaning was restricted to the removal

276

jill dunkerton, rachel morrison and ashok roy

Fig. 18.5 Portrait of Margaret van Eyck, reverse

pigments, media and varnish layers on the portrait of margaret van eyck

of a thick wax coating which had imbibed a considerable amount of grey surface dirt. The brown paint on the mouldings is not original. At the bottom of a cross section (which otherwise shows only layers of new ground and repaint) there is a trace of black paint and in a few places where the repaint had flaked away it can be seen that the mouldings were painted black. In one loss there was a suggestion of red marbling. It seems likely that little of this original decoration has survived and so the later (and familiar) repainting has been retained. Conversely, the marbled back of the panel has survived in such outstanding condition that its originality might easily be questioned. The absence of cracking is remarkable, and it has been suggested that parchment may have been laid over the wood before the application of the ground layer.19 This consists of the usual chalk; in the cross section it has a very similar appearance to the ground on the front of the portrait. As with the moulding, the base layer is black; atr-ftir and sem-edx analysis confirm that this is coal black. Black seems to have been a common and practical base layer for marbling, appearing for example in samples from two Eyckian portraits in the National Gallery. In the Portrait of a Young Man Holding a Ring (ng2602) the marbling was then applied with a single layer of vermilion, red lake and probably red lead. On the reverse of the Portrait of Marco Barbarigo (ng696) a copper green pigment is present in addition to the vermilion and red lake.20 The marbling on the reverse of the Portrait of Margaret Van Eyck, on the other hand, is astonishing for its sense of depth, giving the illusion of a polished stone, albeit imaginary (fig. 18.6). The cross section shows its multilayered construction (fig. 18.7). Above the black layer a dark purplish layer has been applied. In cross section this layer has a slight fluorescence under ultraviolet light and contains some large particles of red lake, very fine particles of vermilion, some white and a few particles of black. atr-ftir analysis of the cross section indicates that the large red lake particles are proteinaceous and have been prepared from the

277

Fig. 18.6 Photomicrograph from the marbled reverse showing spattered paint

shearings of woollen textiles, similar to at least one of the red lakes in the portrait itself and absolutely typical for lakes of this period. Given the appearance of the particles under UV light it is highly likely that a madder lake was used. Over this there is a translucent layer, which fluoresces more strongly under ultraviolet light. This layer is slightly milky and was found by sem-edx analysis to contain a small amount of calcium, almost certainly chalk, and a little pigment which may have been picked up from the lower layer during application. Nevertheless, it is evident that this translucent layer was intended to create the sense of depth since some of the dots and spatters have been applied below it and are seen through the layer while others have been applied on top. The cross section was taken from one of the bright red vermilion spatters at the edge and shows that above the translucent layer there are a few particles of white and a relatively thick layer of vermilion.21 Since oil paint is too thick and viscous to be a suitable medium for spattered decoration it was no surprise that atr-ftir of the cross section and transmission ftir of unmounted fragments have convincingly shown that the binding medium of all the paint layers, including the translucent one, is proteinaceous. As yet we have not determined whether the protein is egg, glue or perhaps casein. The binder used for the lead white spatters on the

278

jill dunkerton, rachel morrison and ashok roy

1394

0.16

0.10 0.08

2922

0.12 3306

Absorbance Units

0.14

1072

1540

1644

50 μm

0.06 0.04 0.02 0.00 3500

3000

2500 2000 Wavenumber cm-1

1500

50 μm

Fig. 18.7 Cross section from one of the red spatters on the marbled reverse of the panel, photographed in visible and ultraviolet light. The ATR-FTIR image, produced by integration of the amide II band between 1573 and 1499 cm-1 is illustrated at the lower left next to a detail from the ultraviolet image of the cross section showing the analysed area. The ATR-FTIR image indicates that the paint layers are rich in protein. The crack running through the sample is clearly visible. An FTIR spectrum extracted from the porphyry-coloured paint layer confirms the presence of protein and is illustrated at the lower left

1000

reverse of the National Gallery’s panel by the Master of Veronica, painted in Cologne probably in the first decade of the fifteenth century,22 is also protein-based. Just as with the Portrait of Margaret van Eyck, the Veronica image on the front is painted in oil. In the case of the Master of Saint Veronica panel, it can be seen under magnification that the paint of the spatters was aerated, resulting in craters from burst air bubbles. In this it resembles the spattering carried out in egg tempera by many Italian painters, notably Andrea Mantegna.23 Van Eyck’s amoeba-like spatters are free of such bubbles, and may well have been made with a different pro-

teinaceous medium, something we hope to resolve as our analytical techniques become more refined. In a reversal of the situation on the front of the panel, the protein-based layers of marbling were completed with a final application of an oil-based varnish. Analysis by gc-ms found that the major component was heat-bodied linseed oil with only traces of diterpenoid resin acids.24 Further analysis of this varnish layer by pyrolysis gc-ms also detected a peak for succinic acid dimethyl ester, again with small quantities of diterpenoid resin acids. This result may suggest that the resin component of the varnish layer is in fact amber.25 However, the

pigments, media and varnish layers on the portrait of margaret van eyck

1 mm

Fig. 18.8 Photomicrograph from the marbled reverse showing small losses in the varnish layer

sample was small and it was not possible to detect the monoterpenoid alcohols fenchol and borneol sometimes observed in samples containing amber, so this identification can only be tentatively postulated.26 The layer is strongly fluorescent in ultraviolet light, except where it is missing as a result of old scratches and flake losses (fig. 18.8). It has retained its translucency and is no more than slightly discoloured. There is no reason to believe that this is not Van Eyck’s original varnish. Its survival may also account for the superb condition of the marbling. Much of this paper has centred on the discussion of just two paint samples. With the application of an ever-widening range of analytical techniques, they have proved to be a rich source of information, clarifying aspects of Van Eyck’s technique, but also, and perhaps inevitably, introducing complications which have yet to be resolved. Fortunately the samples still exist, ready for analysis with the techniques of the future and for reinterpretation in the light of the discoveries that will surely be made during the re-examination and restoration of the Ghent Altarpiece. NOTES 1 Campbell et al. 2008, pp. 180-181. 2 See Campbell 1998, pp. 174-231, and for recent examination and analysis of paint samples from the National Gallery’s works by Van Eyck, see Spring and Morrison in this volume. 3 Note in the file on the Portrait of Margaret van Eyck, held at the Groeningemuseum, Bruges.

279

4 Janssens de Bisthoven 1983, p. 175. 5 http://www.nationalgallery.org.uk/paintings/research/ the-restoration-of-margaret-the-artists-wife/introduction 6 See for example Spring et al. 2008, pp. 37-45. 7 For example, in a photomicrograph of a detail such as Margaret’s left eyebrow it can be seen that every fine ginger hair is present; the little gap is not the result of damage but rather a deliberate omission, either because Margaret had a small scar or had the sparse brows typical of a redhead. 8 The identification of heat-bodied linseed oil was confirmed by the detection of fatty acid methyl esters with the following ratios: A/P 1.2, P/S 1.3, A/Sub 3.0. 9 Pigment identification was carried out by energy dispersive X-ray analysis in the scanning electron microscope (SEM-EDX). 10 Bone black was identified in a small unmounted fragment of red lake glaze by transmission FTIR analysis. The presence of calcium and phosphorus was also confirmed by SEM-EDX analysis. 11 See Spring and Morrison in this volume. 12 Kirby, Spring, Higgitt 2005, pp. 71-87. 13 See Spring and Morrison in this volume. 14 GC-MS analysis of a sample of red lake glaze identified heat-bodied linseed oil with fatty acid methyl esters detected with the following ratios: A/P 1.4, P/S 1.8, A/Sub 3.0. 15 GC-MS analysis of a sample from the black background gave fatty acids as follows: A/P 0.6, P/S 3.9, A/Sub 2.6. Contamination of the sample with beeswax contributed to a larger than expected peak for the palmitic acid methyl ester which has increased the palmitate to stearate ratio. For this reason the type of oil cannot be determined, however the ratio of the diacids azelate and suberate indicate that the oil is heat-bodied. 16 See Spring and Morrison in this volume. 17 Sauerberg et al. 2009, pp. 244-246. 18 In some areas, for example the headdress, it is possible to see small patches of a finely cracked translucent orange-coloured material, perhaps traces of this coating. 19 Verougstraete, Van Schoute 2000, p. 111. 20 Campbell 1998, p. 228 for NG 2602 and p. 224 for NG 696. 21 An old paint sample held at the KIK-IRPA has recently been re-examined by Jana Sanyova and shows a very similar structure to our sample, as indeed does the other viable cross section, from the frame on the front. 22 Spring et al. 2012, pp. 92-93. 23 For a microphotograph showing the air bubbles in paint spatters in a work by Andrea Mantegna see Dunkerton, Billinge 2005, p. 13. 24 GC-MS analysis of a sample of varnish from the marbled reverse identified heat-bodied linseed oil with fatty acid methyl ester ratios as follows: A/P 1.2, P/S 1.3, A/Sub 2.2. 25 In the website publication it was postulated that the traces of dehydroabietic acid methyl ester and its oxidation product 7-oxodehydroabietic acid methyl ester detected by GC-MS analysis could indicate that the varnish included a little pine resin, although the levels of these components were extremely low. However, the subsequent analysis by pyrolysis GC-MS which detected succinic acid dimethyl ester suggests that the resin is more likely to be amber, which also contains minor amounts of diterpenoid resin acids in addition to the polymeric fraction. 26 Sauerberg et al. 2012, pp. 252-253. It should be borne in mind that succinic acid can form to some extent in aged oil paint films through chain scission reactions. Without the detection of other marker compounds for amber it is difficult to reach a definitive identification of this resin.

Fig. 19.1 Jan van Eyck, The Annunciation, c.1434-1436, oil on panel transferred to canvas, 90.2 x 27.1 cm Washington, National Gallery of Art, Andrew W. Mellon Collection (inv. no. 1937.1.39)

19

New Findings on the Painting Medium of the Washington Annunciation Melanie Gifford, John K. Delaney, Suzanne Quillen Lomax, Rachel Morrison and Marika Spring

ABSTRACT: Scientific identification of Jan van Eyck’s painting medium has proved challenging, with apparently contradictory results. Previously published interpretation of GC-MS, HPLC and histological staining suggested that Van Eyck had applied the blue glaze of The Annunciation in a proteinaceous medium. In 2012 further analysis using new methods led to a reinterpretation of the medium of the blue glaze. Fibre-optic reflectance spectrometry (FORS), ATR-FTIR microspectroscopic imaging and GC-MS using a new protocol all identified oil as the primary medium in the final blue glaze on the Virgin’s robe with no clear evidence of a proteinaceous binder. Evidence of calcium oxalate predominantly in the glaze layer, however, suggests that the oil medium is badly degraded in ultramarine-rich paints. The blanched appearance of the surviving blue glazes in The Annunciation support this suggestion. The painting’s appearance today is reinterpreted in light of this degradation.

—o— Jan van Eyck’s painting medium has long been the object of scholarly speculation and research. It is now well established that despite Giorgio Vasari’s report that Van Eyck had invented the technique, oil painting had been in use for centuries before his lifetime.1 Nonetheless, Van Eyck’s technical brilliance has invited speculation that he must somehow have adapted the painting practices that his contemporaries used in order to facilitate his extraordinary rendering of the visual qualities of the material world.2 Van Eyck’s Annunciation at the National Gallery of Art in Washington (fig. 19.1),

with its vivid evocation of jewels, velvet brocades and living flesh in a light-filled church, is a remarkable exemplar of the artist’s work. Technical study of The Annunciation was originally undertaken in conjunction with conservation treatment during 1991-1992.3 This research has continued and recently the painting and paint samples made during the original study were re-examined. In this new study the painting was extensively analysed in ways not available in the 1990s, including in situ methods that do not require samples. In addition, two of the original paint cross sections and three further micro-samples were investigated using recently developed analytical techniques. While the complete findings of the research on the materials and techniques of The Annunciation have recently been published elsewhere,4 the present paper expands on one aspect of this study. As a result of new research we have been able to revaluate and revise our previously published conclusions on the analysis of the paint medium. We also offer a more detailed understanding of how the appearance of The Annunciation today differs from the effects Van Eyck sought to create. Restoration History and Conservation Treatment 1991-1992 An important goal of the technical study at the time of the conservation treatment was to investi-

282 melanie gifford, john k. delaney, suzanne quillen lomax, rachel morrison and marika spring gate the painting’s condition and its history. The evidence of the restoration history uncovered during the study guided conservation decisions. More recently this evidence proved crucial in reinterpreting analysis of the paint medium, and warrants a brief discussion in the present paper. In 1930 Andrew W. Mellon purchased The Annunciation when the Soviet government sold works from the collections of the Hermitage Museum, and the painting was a key part of the collection Mellon donated at the founding of the National Gallery of Art.5 While The Annunciation was in Russia restorers transferred the painting: removing the original wood panel support, washing off the original chalk ground and remounting the painting onto a coarse fabric using thick, leadwhite adhesive.6 After the transfer, the painting was heavily varnished and damages were overpainted. At the National Gallery in 1991 David Bull undertook conservation treatment to remove badly yellowed varnish and extensive repaint, which had been untouched in the almost 150 years since the nineteenth-century transfer. Cleaning revealed a painting which, despite damage from the coarse transfer fabric, was remarkably well preserved in many areas.7 The area of the Virgin’s robe, however, had been a source of concern even before conservation treatment was begun. Because the blue robe had been completely overpainted and that repaint was now badly degraded, the Virgin’s robe appeared qualitatively different from all other areas of the painting. The presence of such extensive repaint suggested that it had been applied to cover damage that was far more severe in this area than in the rest of the painting. Close examination of the painting uncovered strong circumstantial evidence that the original surface of the Virgin’s robe had been damaged earlier, probably during the Russian transfer: particles of ultramarine blue pigment that seem likely to have originated in the paint of the Virgin’s robe were found trapped in damages in the foreground

near the figure of the Virgin.8 Analysis of a paint cross section taken from a location where the paint of the Virgin’s hair had protected the underlying blue paint of the robe revealed the original paint structure of the Virgin’s blue robe. As in his Ghent Altarpiece,9 Van Eyck had applied multiple layers of blue paint; here he used two underpaint layers and a final ultramarine blue glaze. Because removal of overpaint from the robe using mild solvents in the recent treatment revealed only the blue underpaint layers it seemed clear that the glaze of the Virgin’s blue robe had been lost in the earlier restoration while the underlying paint layers have survived. The original blue glaze layer was already missing when the early restorers overpainted the Virgin’s robe, almost certainly in order to hide its loss.10 A particular focus of the technical study in the 1990s was to understand why the blue glaze of the robe had suffered far more damage than any other paint in The Annunciation. Medium Analysis in the 1990s In the 1990s we analysed the organic components of the paint medium using histological staining of paint cross sections as well as two different methods of instrumental analysis: gas chromatographymass spectrometry (gc-ms)11 to analyse for drying oils and resins and high-performance liquid chromatography (hplc)12 to analyse for proteins (at that time there was no single analytical protocol that could test for both oil and protein). In the samples analysed for proteins using hplc, amino acids were found in only trace amounts, far too low to be indicative of a protein paint medium.13 gc-ms analysis for evidence of oil, however, was less consistent. A sample of tan paint in the architectural background gave clear evidence of a drying oil medium. The specific oil present was identified by measuring the ratios of fatty acids, yielding ratios typical of a linseed oil medium: the ratio of palmitic to stearic acids (P/S ratio) was 1.99 and the ratio of azelaic to palmitic acids (A/P ratio) was 0.95. By contrast, the results from a sample of the blue underpaint of the Virgin’s robe were far more

new findings on the painting medium of the washington annunciation

ambiguous, yielding a P/S ratio of 1.25 and an extremely low A/P ratio of 0.06. At the time we suggested that these atypical ratios of fatty acids, which are inconsistent with standard values for most drying oils, might be evidence of a degraded oil medium in the underpaint. The existence of Van Eyck’s now-lost ultramarine-blue glaze is documented only by the paint cross section taken from the Virgin’s hair. Below the hair, this sample preserved all three original blue paint layers: both the layers of underpaint and the final glaze (fig. 19.2a). Van Eyck’s only blue pigment in The Annunciation was ultramarine; in the modelling underlayers of the robe he mixed this with other pigments but the glaze is pigmented only with finely ground ultramarine.14 Because the glaze is now entirely missing from the rest of the Virgin’s robe no samples could be taken for analysis by gc-ms and hplc. However, limited analysis of the glaze was possible through histological staining of this one paint cross section. After the sample was stained with Ponceau S, a histological stain for protein, the glaze layer was strongly red-pink but little staining was apparent in the two underpaint layers, suggesting a positive reaction for protein in the glaze (fig. 19.2b).15 Previous to our analysis in the 1990s, published studies of other works by Van Eyck also had suggested that the artist painted with ultramarine in a medium different from the drying oil he had used with other pigments. Coremans et al. suggested that in the Ghent Altarpiece ultramarine layers in the Virgin’s mantle and in blue jewels were painted using a water-based paint medium.16 Kockaert and Verrier, using histological staining of paint cross sections, concluded that ultramarine blue paint layers were bound in a protein medium, probably egg white.17 Based on the analytical methods available to us in the 1990s, the earlier comparative research that had been published, and early sources that suggested the use of an aqueous medium to preserve the colour of blue pigments,18 we proposed that Van Eyck might have applied the ultramarine glaze

283

a

IV III II I

50 μm b

IV III II I 50 μm

Figs 19.2a-b Paint cross section from a strand of the Virgin’s hair where it overlaps her blue robe: (a) visible light image showing layers i) light blue underpaint, ii) mid-tone blue, iii) blue glaze, iv) the Virgin’s hair; (b) the same sample, after histological staining with Ponceau S, appeared to show a positive reaction for protein in the ultramarine blue glaze layer

using a protein-based, water-soluble medium and that the original glaze could have been lost accidentally when the painting had been treated with water during the nineteenth-century transfer.19 We also noted that not only was the blue glaze lost, but also that the surviving blue underpaint appeared to be more degraded than the paint in other areas of the painting. Medium Analysis in 2012 Recently, we had the opportunity to carry out further analysis of the paint medium using methods developed since the 1990s. Three additional microscopic paint samples were removed from the edges of the painting and investigated using a recently developed gc-ms protocol that analyses simultaneously for both fatty acids and amino acids.20 As previously, analysis of tan paint from the side wall of the church found fatty acid ratios consistent with linseed oil. The analysis results from a sample of

284 melanie gifford, john k. delaney, suzanne quillen lomax, rachel morrison and marika spring

0.001

0.000 dR/dw

the dark red glaze on the cushioned seat in the foreground were also consistent with drying oil. However, as in the earlier analysis, a sample of the underpaint in the blue robe again gave anomalous results that were suggestive of a degraded paint: in this case no stearic acid was identified in the blue underpaint. As in the earlier analysis, only trace amounts of amino acids were identified in all three samples.21 It was noted in gc-ms analysis, however, that the sample of the red glaze also included additional traces of other amino acids that were not present in the blue and tan paints. Analysis was also carried out with fibre-optic reflectance spectrometry (fors), an in situ methodology that has recently been used to identify paint media without requiring samples.22 In the tan paint of the side wall of the church the fors results were consistent with the findings of gc-ms: fors identified vibrational features characteristic of a drying oil medium, but did not show evidence of protein. In red glaze paint, where gc-ms had found evidence of an oil medium but also minute traces of amino acids not present in other samples, fors identified not only features characteristic of oil but also the amide features of protein. Because fors does not require samples, it was possible to repeat the analysis at multiple points on the painting surface. Evidence of protein was found in every area of red glaze analysed, but was not found in any other areas of the composition. In particular, protein was not identified in any blue paint, neither in the exposed underpaint of the Virgin’s robe, nor in areas where the final blue paint layers are still present, such as within the halos of the archaic wall paintings in the background of The Annunciation. Figure 19.3 shows the first derivatives of two fors spectra, one from a blue area and the other from red glaze. Both show the transition points characteristic of a drying oil, but amide features appear only in the spectrum of the red glaze.23 Attenuated total reflection Fourier-transform infrared microspectroscopic (atr-ftir) imaging (a technology that had not been developed when the paint cross sections were made in the 1990s)

-0.001

-0.002 2000

2100

2200 Wavelenght (nm)

2300

2400

Fig. 19.3 First derivatives of FORS spectra for red and blue areas in The Annunciation: a blue halo depicted on the rear wall (shown in blue) and red glazes (shown in red). In both the red and blue paints, FORS analysis shows transition points at 2305 nm and 2348 nm, characteristic of drying oils. Absorption features at 2053 nm and 2168 nm, characteristic of protein, are observed only in the red glazes and not in the blue paints

yielded additional information from the original cross sections without requiring new samples.24 With this microscopic technique a large number of atr-ftir spectra are collected from regularly spaced points across the surface of the cross section, so that in addition to characterizing the materials at any single point, the data can be processed to produce images that map these materials, showing their location within the paint layer structure. atr-ftir imaging of one sample from the red brocade of Gabriel’s cope identified protein in the upper red lake-based layer (fig. 19.4). Integration of the amide I absorption band produced an image indicating that it was located within the particles of red lake; areas of high intensity in the atr-ftir image correspond to the large red particles. The fors analyses, which suggested the presence of protein more generally in all areas of red lake glaze in The Annunciation, are consistent with the atr-ftir imaging results, which demonstrate that protein is a component of the pigment rather than the paint medium. This finding can be understood by examining fifteenth century historical recipes for the produc-

new findings on the painting medium of the washington annunciation

a III II

I

b

III II

I c

1537

1644

50 μm

0.14

1452

Absorbance Units

0.12 0.10 0.08 0.06 0.04 0.02 0.00 d

2200

2000

1800

1600 1400 Wave number cm-1

1200

1000

Figs 19.4a-d Cross section from the red brocade of Gabriel’s cope: (a) visible light image showing the area analysed by ATR-FTIR imaging, with layers i) red underpaint, ii) red glaze, iii) dark shadow; (b) ATR-FTIR image produced by integration of the amide I band between 1684 and 1590 cm-1; (c) UV- induced autofluorescence image showing the area analysed; (d) average ATR-FTIR spectrum from large red lake particle showing absorption bands characteristic of protein

tion of red lake pigments, several of which show that shearings of dyed wool textiles could be used as the source of the red dyestuff. The dyestuff was usually extracted with an alkali such as lye, which was sometimes strong enough to dissolve the wool protein so that it became incorporated into the pigment particles.25 Most importantly, atr-ftir imaging was also used to gain new information from the only exist-

285

ing sample of the blue glaze on the Virgin’s robe (fig. 19.5). No convincing evidence of protein in the blue paint layers was found, even in the upper ultramarine glaze, where histological staining in the 1990s had seemed to show a positive reaction for protein.26 Furthermore, an absorption band characteristic of glyceride esters was observed, suggesting that the paint medium is likely to be a drying oil.27 Spectra from the ultramarine-containing paint also showed absorption bands characteristic of calcium oxalate and the atr-ftir images indicated that this material is concentrated in the ultramarine glaze at the uppermost part of the sample.28 The reasons and circumstances of calcium oxalate formation in paints are not well understood, but it seems that its presence might indicate a breakdown of the medium, and therefore of degradation: in particular it has been noted in association with ultramarine-rich paints, both within the layer and on the surface as a crust.29 Furthermore, of the three blue layers the glaze has the highest concentration of ultramarine and, because the particles in the glaze are more finely ground, has a higher relative surface area of ultramarine exposed to the medium. Both factors would likely increase the reactivity of the pigment with the oil medium. Reinterpretation of Medium Analysis The greater understanding of the influence of ultramarine on the ageing of the paint medium that has developed since the 1990s examination, and evidence of deterioration such as the presence of calcium oxalate confirmed by atr-ftir imaging, helps to explain the earlier results from the blue glaze layer. The degradation is likely to have made this paint layer both more sensitive to aqueous solutions and more porous, thereby increasing the potential for obtaining a false positive result by histological staining. It now seems that Ponceau S did not actually yield a positive reaction, which would stain the blue glaze layer through binding with a protein medium. Instead, the darker colour of the glaze layer after staining (see fig. 19.2b) was probably due to an excess of stain trapped within its

286 melanie gifford, john k. delaney, suzanne quillen lomax, rachel morrison and marika spring irregular surface. In addition, it is now understood that gc-ms analysis of ultramarine in oil can give anomalous results, as in this study, due to the effect of the pigment on the chemistry of the medium as it ages and degrades.30 The present reinterpretation of the Washington Annunciation paint medium, based on new analytical results, suggests that it may be fruitful to re-examine other, earlier publications that also proposed that Van Eyck used an aqueous, protein-based medium when painting with ultramarine blue.31

a IV III

II

IV III II II

b

c

1720

0.10

0.08 Absorbance Units

1320

1626

50 μm

0.06

0.04 0.02

d

1900

1800

1700

1500 1600 Wave number cm-1

1400

1300

1200

Figs 19.5a-d Cross section from a strand of the Virgin’s hair where it overlaps her blue robe: (a) visible light image showing the area analysed by ATR-FTIR imaging (layers identified in fig. 19.2a); (b) ATR-FTIR image produced by integration of the band between 1332 and 1305 cm-1, characteristic of calcium oxalate; (c) UV- induced autofluorescence image showing the area analysed; (d) detail of the averaged ATR-FTIR spectrum from a calcium oxalate-rich part of the upper blue paint layer (shown in blue). The blue spectrum shows an absorption band at 1720 cm-1 characteristic of glyceride esters, and two absorption bands characteristic of calcium oxalate: compare to a standard spectrum of calcium oxalate (shown in red)

Re-evaluation of the Painting’s Appearance The recognition that the ultramarine-blue glaze of the Virgin’s mantle has degraded calls for a re-evaluation of The Annunciation’s present appearance. Although historical documentary sources do suggest that in wall paintings and polychrome sculpture (which were otherwise painted in an oil medium) the blue pigments azurite and ultramarine were sometimes applied in an aqueous, proteinbased medium, there is little evidence that this was typical in panel paintings, except perhaps in Spain.32 When varied media were used on the same work of art, a paint layer of coarsely ground blue pigment in an aqueous medium would have a distinctly different appearance when compared to other colours on the same work executed in an oil medium. The blue paint would be lighter and also far more opaque, both because the irregular surface would scatter light, and because the refractive index of the paint medium would effectively be much lower than the pigment once the water evaporated from the glue medium leaving air-filled voids in the paint layer. This difference in appearance was confirmed in the case of the Santa Marina retable of the late fifteenth century, and indeed in Spain historical sources give evidence regarding such use of mixed media.33 In that retable, analysis identified oil paint or egg tempera in most areas (particularly red and green glazes) but indicated a glue medium in blue passages.34 Close examination of the paintings out of their frames revealed exactly the predicted dichotomy in appearance: oil-bound

new findings on the painting medium of the washington annunciation

287

Fig. 19.6 The Annunciation, macrophotograph of a blue halo in the background wall paintings showing a hazy, whitish appearance typical of degraded ultramarine-blue-containing paint

paints have saturated colours, but where the removal of the frames revealed intact blue passages the paint is a light and unsaturated blue. In oilmedium passages, the fall of light on drapery folds was modelled with gradations of tone while the glue-bound blue draperies were modelled with a system of black hatched lines. Moreover, the artists seem to have valued the unsaturated appearance of the blue areas because they refrained from varnishing these exact areas.35 While the generalized depiction of draperies that resulted from using mixed media is successful in the context of large ensembles such as Spanish retables, Jan van Eyck’s genius was his use of oil paint to create an astonishing evocation of the visual qualities of the natural world. Van Eyck’s illusion would have been ruptured if he had incorporated passages of light-coloured, glue-bound paints into an image otherwise composed of richly saturated colours. When ultramarine blue pigment is bound in oil (rather than in glue) the refractive indices of pigment and medium are almost the same, giving a deep, transparent colour well-suited to the dark shadows of Van Eyck’s blue draperies and his translucent jewels.36 This optical effect makes it improbable from an aesthetic, as well as a

technical, perspective that Van Eyck applied final ultramarine glazes in an aqueous medium. In The Annunciation, however, the degradation specific to the ultramarine-rich glazes has caused those areas to lose their deep saturated colour. The recognition that the present-day light blue colour was not intended revises our understanding of the painting’s original appearance. On the rear wall of the church interior Van Eyck depicted wall paintings with scenes from the life of Moses. He worked in a deliberately archaic style, with a dim, muted palette that recedes into the shadows beneath the equally archaic timber roof. The blue paint of the haloes was applied as a single layer of ultramarine. Microscopic examination of the haloes revealed visual evidence of degradation: a hazy, whitish appearance that is most pronounced where the paint is most thickly built up (fig. 19.6). As a result, the light blue haloes stand out prominently in the vaguely depicted wall paintings. Although the blue glaze of the Virgin’s robe has been lost, The Annunciation offers one area in which Van Eyck’s complex layering of glaze and underpaint to depict blue fabric survives. A glimpse of the lining of Gabriel’s cope offers a deep blue contrast to the rich red and green of his vestments.

288 melanie gifford, john k. delaney, suzanne quillen lomax, rachel morrison and marika spring where on the painting; only the glaze of the Virgin’s robe was lost. It seems most likely that early restorers deliberately removed the degraded blue glaze from the robe, whether because they misinterpreted it as earlier repaint, or that they deemed the degradation aesthetically unacceptable. Other blue areas may have survived because the early restorers simply overlooked similar degradation in areas seen as less essential to the composition.

Fig. 19.7 The Annunciation, macrophotograph of the dark blue lining of Gabriel’s cope showing a bluish-black underpaint; the final brushstrokes of ultramarine glaze appear degraded and light-coloured

Microscopic examination shows a bluish-black underpaint with freely handled final blue brushstrokes (fig. 19.7). Magnification also reveals that only those free brush strokes appear degraded and light-coloured. Although the degraded light blue strokes now disrupt the shadowed recess behind Gabriel’s sleeve, those final strokes of paint must originally have been intended to be a deep, translucent blue glaze. In light of this new research it seems that loss of the blue glaze on the Virgin’s robe, probably during a nineteenth century restoration, was not due to the use of a different paint medium for the blue glazes. Survival of the blue halos and the lining of Gabriel’s cope makes it clear that the cleaning and transfer during the nineteenth century restoration did not remove the final blue paint layers every-

Conclusion The analysis of aged organic materials such as the paint medium is a challenging task. The interpretation of analytical results is strengthened by consideration of the historical and aesthetic context, but new research technologies can also open new perspectives. Renewed technical study of Van Eyck’s Washington Annunciation, using analytical methods that were not previously available, has prompted the re-examination of earlier conclusions on the paint medium. The new analysis suggests that Van Eyck did not use a different medium to paint his ultramarine-blue glazes, but used drying oil as the primary paint medium throughout the painting. The altered appearance of the final blue glazes seems due to degradation of the oil medium in the most ultramarine-rich paints. In the case of The Annunciation the new analysis allowed for reconsideration of what the degradation of blue glazes means for the painting’s appearance today (see fig. 19.1). When these glazes were fresh, the Virgin’s robe would have complemented the red and green glazes of Gabriel’s vestments with a deep blue that restoration can evoke, but can never fully recreate. The celestial rainbows of the archangel’s wings would not have ended in bands of pale, opaque blue, as they do now. Van Eyck’s visual sequence from brilliant orange and yellow at the centre of the wings, deepening to dark copper green above and dark red lake below, would have merged into a clear, sapphire blue at the margins of the wings that echoed the nearby jewels.

new findings on the painting medium of the washington annunciation

NOTES 1 Vasari 1568, pp. 425-426. On the earlier use of the oil medium see Roy 2000, White 2000. 2 For an overview of theories on Van Eyck’s painting medium see Effmann 2006. 3 Bull 2013. 4 Gifford et al. 2013. 5 Hand, Wolff 1986, pp. 76. 6 Bull 2013 details the treatment history. 7 For a detailed discussion of the conservation treatment see Bull 2013. 8 Bull 2013, p. 163; Gifford et al. 2013, p. 140. 9 Coremans 1953, p. 70. 10 Conservation treatment included compensation of the missing glaze. For a full discussion of the treatment decisions and for photographic documentation, see Bull 2013. 11 Lomax 1993.The samples were placed in vials and a solution of C13 and C15 fatty acids were added as internal standards. The samples were hydrolysed by the addition of a 10% KOH solution. Upon reacidification, the organics were removed by extraction with ether. The ethereal extracts were dried, redissolved in a methanol/ether solution and methylated with diazomethane. Analysis was performed on a Varian 3500 gas chromatograph interfaced with a Finnigan 800 ion trap mass spectrometer. In all three samples analysed peaks were found corresponding to fatty acids consistent with drying oil as well as diterpenes, including one attributable to 7-oxodehydroabietic acid that suggests the presence of small amounts of aged pine resin. 12 Halpine 1993. Norleucine, an internal standard was added to each sample followed by cold water. The samples were sonicated and centrifuged and the supernatant and precipitate were independently analysed. The samples were hydrolysed with 6N HCl under vacuum for 24 hours at 110o C and then derivatized with phenylisothiocyanate (PITC). Analysis by reverse phase HPLC was accomplished with a Picotag column and Picotag eluent on a Hewlett Packard HPLC. Diode array detection at 254 nm was used. 13 The trace amino acids were not identified as having derived from a specific protein. They were not consistent with egg protein, and there was no evidence of hydroxyproline, an amino acid that is specifically characteristic of glue. 14 Gifford et al. 2013, p. 140. 15 The sample was stained for twenty minutes with a saturated solution of Ponceau S (Acid Red 112) in 3% acetic acid, rinsed with water, then examined at high magnification in visible light. 16 Coremans 1953, pp. 70-71. 17 Kockaert, Verrier 1978-1979 stained for protein with the red stain, acid fuchsin. In the published images ultramarine blue layers exhibit pink-red staining consistent with the results for the Washington Annunciation. 18 Raft 1982. 19 Gifford, 1999, p. 108. 20 Lomax 2000, Lomax 2012. Norleucine was added to each sample as internal standard. Samples were hydrolysed for 24 hours under vacuum with 6N HCl at 100o C. After evaporating to dryness, the samples were silylated with a solution of 30% MTBSTFA/TMDMCS in pyridine. After heating to 60o C for 30 minutes, then 105o C for five hours, the samples were cooled and analysed by GC-MS. A Varian CP3800 GC and Saturn 2200 MS were used.

289

21 As previously, hydroxyproline, an amino acid indicative of animal glue, was not identified in any of the samples in this analysis. 22 FORS analysis was performed using a spectroradiometer operating from 350 to 2500 nm with approximately 2 nm sampling. Sites analysed were 3 mm in diameter with acquisition times of 5 seconds and light intensity of approximately 4000 lux. On the use of FORS in medium identification see Ricciardi et al. 2012. 23 In the first derivatives of the FORS spectra for both red and blue areas transition points were observed at 2305nm and 2348nm, consistent with published values for drying oils (Vagnini et al. 2009, p. 2110); absorption features at 2053 nm and 2168 nm, observed only in red glazes, are consistent with published values for proteins (Vagnini et al. 2009, p. 2111). For the spectrum of red glaze, spectra from the red cushioned seat and Gabriel’s red cope were averaged. 24 ATR-FTIR imaging was carried out using a Bruker Hyperion 3200 microscope equipped with a germanium ATR objective and a 64 ≈ 64 pixel focal plane array (FPA) detector coupled to a Bruker Tensor 27 spectrometer. 25 Kirby, Spring, Higgitt 2005, pp. 76-78. 26 There also was no evidence that the ultramarine particles had been pre-coated with a protein, such as glair or glue, before being ground in an oil medium: a preparation method that has been observed using FTIR microscopy in Norwegian altar frontals of the thirteenth and fourteenth centuries. See White 2000, p. 104; Plahter 1997, p. 346. 27 An absorption band around 1720 cm-1 was observed in most spectra from the blue paint layers. This is characteristic of glyceride esters; however, it is difficult to confirm the presence of a drying oil on this basis alone. The fats present in egg tempera also give rise to the same FTIR absorption, although a clearer indication of protein bands would be expected if an egg tempera binding medium had been used. 28 A broad absorption band around 1626 cm-1 and a sharp band around 1320 cm-1 are characteristic of calcium oxalate, matching most closely reference spectra of whewellite, the monohydrate phase of calcium oxalate. Cariati et al. 2000, p.184. 29 See Spring and Morrison in this volume; also Higgitt, White 2005, pp. 93-94; Spring, Higgitt 2006. 30 See Spring and Morrison in this volume, and references therein. 31 As above, nn. 16, 17. 32 Hodge, Spring, Marchant 2000, p. 24. Interestingly, the seventeenth-century Spanish artist and writer Francisco Pacheco reflected this tradition when he asserted that Van Eyck had used a tempera for the blues of the Ghent Altarpiece: Arte de la Pintura, 1649, cited by Raft 1982, p. 116. 33 Hodge, Spring, Marchant 2000, pp. 15, 24; Veliz 2000, p. 36. 34 Hodge, Spring, Marchant 1998, p. 74: analysis by GC-MS and FTIR indicated the use of oil and egg tempera except in blue areas, which were probably bound with glue. 35 In the Santa Marina retable the paintings were otherwise completed and were completely varnished before the glue-bound blue paint was applied: Hodge, Spring, Marchant 2000, pp. 25-26. 36 White, Kirby 2006 made this same point, p. 215.

Fig. 20.1 Jan van Eyck, Portrait of a Man (Self Portrait?), 1433, oil on oak panel, 33.1 x 25.9 cm (with frame), detail of the frame, London, National Gallery (NG 222)

20

Jan van Eyck’s Greek, Hebrew and Trilingual Inscriptions Susan Frances Jones

ABSTRACT: Van Eyck probably learnt Greek transliteration not from scribal colophons, as has been supposed, but from his training as a painter. He did not copy existing systems, however, but developed his own practice – and he may have drawn on more than one source. His transposition of Greek letters from sacred names within the pictorial field to his personal motto on the frame confirms his intellectuality as a painter. He must have used Greek letters partly for their intellectual and social value. That the motto also had a religious dimension is highly probable; arguably, it was deliberately multilayered and complex. This made it apt for small-scale paintings intended to be viewed in private settings by privileged viewers who were highly alert to visual codes and signs. Along with Van Eyck’s signature, the motto was also part of Van Eyck’s broader revival of the ‘Romanesque’, which placed him within a regional history of image-making.

—o— Van Eyck’s use of Greek has seemed wholly exceptional for a painter: no other painter in the Burgundian Netherlands of his day exploited Greek letters in such complex ways. Greek was a language of learning but in the West, in the Middle Ages, grammatical knowledge of the language was rare, even among scholars. The imperative to learn Latin pushed Greek into second place: in the words of Walter Berschin, Greek was ‘more honoured than studied’.1 This was because, along with Hebrew and Latin, it was one of the original languages of Scripture. The combination of these three ‘sacred languages’ (as they were termed by

Isidore of Seville) had an antecedent in Christ’s lifetime, moreover, since according to the Gospel of John the tituli affixed to the Cross, mockingly identifying Christ as ‘Jesus of Nazareth King of the Jews’, were written in Hebrew, in Greek and in Latin. This article aims to shed light on how and why Van Eyck used Greek, Hebrew and trilingual inscriptions in his works and on the sources of his knowledge. The most enigmatic Greek letters in Van Eyck’s oeuvre are those in his personal motto, aΛ𐅝 ixh xan (als ich can), depicted most prominently on the upper frame of the Portrait of a Man (Self-Portrait?) (fig. 20.1): there, the Greek letter lambda (Λ) is a substitute for the Roman letter L, the square sigma (𐅝) for S and chi (X) for C. When Van Eyck adopted the motto is unknown. The earliest surviving example is on the upper frame of the Portrait of a Man (Self-Portrait?), dated 21 October 1433 (London, National Gallery), by which time he was already established at the Burgundian court (from May 1425) and indeed in Bruges (where he settled probably after his return from Portugal in January 1430). It appears on the frames of three other surviving paintings: the Virgin and Child with Saints Catherine and Michael and a Donor (1437, Dresden); the Portrait of Margaret van Eyck (1439) and the Virgin by a Fountain (1439), and probably also on the frame of a lost Holy Face (figs 20.2a-e).2

292

susan frances jones

into rex regu[m] while delta, chi and lunate sigma appear in the words dominus dominancium. On the Deity’s stole, in addition, an uncial omega is substituted for the letter O in the Hebrew name Sabaoth, which refers to the Deity as the Lord of Hosts. From these and other examples in his signed and accepted painting we can determine that Van Eyck employed a system of Greek transliteration, using the following equivalents:

a

Lambda (Λ) – L Chi (Χ) – C Square sigma (𐅝) or lunate sigma (C) – S Rho (Ρ) – R Gamma (Γ) – G Delta (Δ) – D Upsilon (Υ) – U Omega (󰰴) – O3

b

c

d

e Figs 20.2a-e Jan van Eyck’s motto as it appears on: (a) the Dresden Triptych (1437; Dresden Gemäldegalerie Alte Meister, inv. no. 799); (b) the Portrait of Margaret van Eyck (1439; Bruges, Groeningemuseum, inv. no. 0000.GRO0162.I); (c) the Virgin by a Fountain (1439; Antwerp, KMSKA, inv. no. 411), (d) copy of a lost Holy Face (Bruges, Groeningemuseum, inv. no. 0000.GRO0206.1), (e) copy of a lost Holy Face (Berlin, Staatliche Museen zu Berlin, Gemäldegalerie, inv. no. 528).

Van Eyck also deployed Greek letters in the inscription along the edge of the mantle of the Deity Enthroned in the Ghent Altarpiece, which reads rex regum et dominus dominancium (King of Kings and Lord of Lords). There, the Latin words are partly transliterated into Greek: the letters rho, gamma and upsilon are incorporated

What theories have been put forward to explain such knowledge? In an influential article of 1968, Robert Scheller argued that Van Eyck derived his personal motto from scribal colophons. Scheller suggested that his source was a long-standing modesty formula that originated in the rhetoric of classical antiquity and was taken up by medieval copyists and emendators: ut potui (non sicut volui); in Dutch, Zoals ich kan (niet zoals ik zou willen); in English, ‘As I can (not as I would)’.4 The sentiment appears to be that he has made the painting ‘As best as I can’ or ‘As well as I can’. In addition, Scheller contended that the Greek letters in the motto also came from scribal colophons, in which a scribe might use Greek letters to conceal his name, making himself appear at once learned and modest.5 In support of this, he referred the reader to an article by Bernhard Bischoff, published in 1951, which addressed colophons by scribes connected to the Carolingian and Ottonian courts, written between the ninth century and the eleventh or twelfth centuries.6 An example is that written by the scribe Adalbald, working at Tours in the ninth century:7

jan van eyck’s greek, hebrew and trilingual inscriptions

ΔC ΑΔΗCΘ󰰴 CΚΡΙΒΗΝΘΗ ΑΔΑΛΒΑΛΔ󰰴 D[EU]S ADESTO SCRIBENTI ADALBALDO GOD BE PRESENT TO THE SCRIBE ADALBALD To pursue this enquiry we can reproduce here another inscription, which appears in the dedication miniature of the Gospel Book of Abbess Svanhild of Essen, of the eleventh century: 8 CXΑ MΑΡΥΑ ΑΔ ΠR󰰴ΠΡΥΜ ΝΑΤU ΦΕΡ NPM VΥΡΓ󰰴 ΠΡEΧΑΤ

S[AN]C[T]A MARIA, AD PROPRIUM NATU[M], FER N[OST]R[U]M, VIRGO, PRECAT[UM] HOLY MARY, TO THINE OWN SON, BRING, O VIRGIN, OUR PRAYER These inscriptions use mostly the same substitutions as Van Eyck, and they have the same perceived inaccuracies. In each case, the underlying text is in Latin but the individual characters have been transliterated into Greek, letter by letter. The choice of Greek letters was guided by their phonetic similarity to the Roman letters (rather than, say, their visual likeness); however, there was also a desire to choose Greek letters that were as little like Roman letters as possible: upsilon (Υ) over iota (I), theta (Θ) over tau (T) and chi (X) over kappa (K). To describe this phenomenon, Berschin used the term ‘hyper-Greek’.9 In his motto, Van Eyck appears to have used the Greek letter chi (X) to represent the Roman letter C, although the letter kappa (K) would be a more accurate phonetic choice. His use of chi (X) there has been regarded as a mistake.10 Nonetheless, chi (X) is used to represent C in the second inscription shown above: it appears in the first word s[an] c[t]a and in the final word precatu[m]. As Scheller has already observed, Van Eyck’s choice of chi (X) to represent C was in keeping with long-established tradition and is correct in its own terms.11 The Greek letters in Van Eyck’s motto are no more than a straight-

293

forward transliteration of selected letters from a Middle Dutch text that reads als ich can. That the central word is ich makes it likely that the motto represents Van Eyck’s individual dialect and way of speaking.12 Scheller’s interpretation of the motto and related inscriptions has various implications. It suggests that among fifteenth-century painters only Jan van Eyck had such knowledge and that he obtained it directly from the world of scribes. Indeed, Scheller went on to suggest that Van Eyck’s mode of signing and dating, including particular verb forms, various scripts and unusual terms such as actum, were derived from scribal and notarial practice. Maurits Smeyers agreed that Van Eyck’s manner of signing and dating was in a sense related to the practice of scribes, and that the motto was borrowed specifically from scribal colophons.13 By these means, according to those scholars, Van Eyck presented himself to the ‘reader’ of his works as a learned painter. For Scheller, it was the perception of a rigour derived from learning and even from legal knowledge that led Bartolomeo Fazio, writing in 1456, to say of Van Eyck litterarum nonnihil doctus – ‘he was not unlettered’.14 Where did Van Eyck employ Greek? In every instance aside from his motto, it is used for the nomina sacra – the sacred names of God: xP̅ c; agla; yecyc; sabaot[h]; eloy; rex regum et dominus dominancium; iesus nazarenus rex iudaeorum. The Greek inscription in the Portrait of a Man (‘Léal Souvenir’) is admittedly a complex case, but, as Lorne Campbell has noted, the inscription tΥm. 󰰴ΘΕΟc contains the name of God 󰰴ΘΕΟc (in Latin, otheos), deliberately separated from the first three letters by a point on the baseline (fig. 20.3).15 The idea that the inscription in the portrait was an attempt at writing Τιμόθεος (Timothy) gave rise to the idea that the sitter wanted to compare himself to the ancient Greek lyre-player and musical innovator, Timotheos of Miletus (d. 360 bce).16 But the interpretation of the inscription remains open to question. If we apply Van Eyck’s system as

294

susan frances jones

Fig. 20.3 Jan van Eyck, Portrait of a Man (‘Leal Souvenir’), 1432, oil on oak panel, 33.3 x 18.9 cm, London, National Gallery (inv. no. NG 290), detail

explicated above, in which Greek upsilon (Υ) is a substitute for Roman U, and take account of the fact that there is a point in the inscription and a large space after the first three letters (tΥm.), then the text can be reconstructed in Latin as a twoword inscription tum. otheos (‘Then God’).17 There is precedent in medieval systems of transliteration, however, for the Greek letter upsilon (Υ) to represent the Roman letter I, as seen in the eleventh-century inscription above, in which maΡΥΑ is used for maria, among other instances. This makes it possible that Van Eyck decided to use upsilon (Υ) here, against his usual custom, because it seemed preferable to Greek iota (I), which would have been indistinguishable from the Roman letter I. Like the scribes discussed above, he could have chosen upsilon for its obvious ‘Greekness’.18 In that case, Van Eyck may well have been thinking of the name Τιμόθεος. Although the name was not current in the Burgundian Netherlands in Van Eyck’s time, as Erwin Panofsky observed, this does not exclude a sitter from outside the Netherlands. We find the name Timoteo in Italy and the name Tymotheus in Germany.19 Two copies of the painting have Italian provenances, raising the possibility that the original was at one time in Italy, perhaps in the seventeenth century.20 With respect to the

painting’s meaning and purpose, it is clearly important that Van Eyck used punctuation and spacing to draw attention to the sacred name 󰰴ΘΕΟc. The Greek letters here are probably best studied in the context of Christianity rather than pagan antiquity. Such sacred names were believed to be powerful. There was some debate as to whether their power was intrinsic to the word or was derived from God through the word; but divine aid was sought for a wide variety of practices that we might call magic. Names of God were used in curing and healing (they might be read out over curative herbs, for example), to help in pregnancy and childbirth, to secure divine protection and assistance against enemies and, in exorcisms, to drive out demons. In the eyes of those who saw, read or uttered the names, and used them in everyday life, these various activities were not divisible into distinct categories of magic, religion and medicine, as they would be now.21 Van Eyck depicted such inscriptions on armour, weapons and floor tiles, which suggests that they were inscribed on such objects in actuality. In the Ghent Altarpiece, one of the Knights of Christ carries a shield in which divine names are formed into a cross: at its centre is a Greek tau. This accords

jan van eyck’s greek, hebrew and trilingual inscriptions

295

Figs 20.4a-b Ghent Altarpiece, Musician Angels, details of the tiled floor

with the belief that sacred names and letters would increase the power of objects that could mean the difference between life and death. In the Ghent Altarpiece, the tiled floors in the panels of the Musician and Singing Angels display various names and monograms of God, some of which incorporate Greek letters. The three final letters of the so-called yesus tile are lunate sigma (C), upsilon (Υ) and lunate sigma (C), which must represent Latin sus; the initial letter is perhaps meant to be a Latin Y; it does not seem to be a Latin I (fig. 20.4a). Importantly, the name is inscribed as if on a piece of paper or parchment that is curled up at the edges on both sides: this is not just a word on a tile but a representation of a textual amulet, which might be used for prayer, protection or healing.

Another of the repeats in these floors (which appears on only two tiles) is the name agla: in one of the tiles it contains a Greek gamma (Γ) for the letter G (fig. 20.4b). As is well known, this name is said to derive from words of Hebrew origin meaning ‘Thou art mighty for ever, Lord’ (ate gebir leilam adonai). These words have been made into an acrostic written in Roman characters, allowing them to be grasped, read and spoken without difficulty. This sort of adaptation was important because although Hebrew was the least comprehensible of the three sacred languages, at least to Van Eyck’s viewers, it was also the highest and most holy. Sacred names and titles in general might be transliterated from the original language, but they were not translated, as it was believed that this would diminish their power.22 agla was

296

susan frances jones

especially popular on jewellery, suggesting that always keeping it on the person was seen as important. 23 It was precisely in Van Eyck’s day that the production of textual amulets, written on paper or parchment, reached a peak. These texts did not need to be read in order to be effective, although this was increasingly possible for courtly, noble and indeed middle-class users of the late medieval period. Such amulets had complex, multiple functions: they might be read, contemplated, worn for protection or applied directly to a wound or afflicted area in an attempt to heal or to cure.24 Among the sacred names in such amulets we sometimes find that a Greek letter has been substituted for a Roman one: in one example, in the name Adonai, an uncial omega appears in place of the Roman letter o (ad󰰴nai), the same substitution that was made by Van Eyck in the name Sabaoth (saba󰰴th) in the Ghent Altarpiece. Evidently, these textual materials provide parallels for Van Eyck’s practice. In all these examples, the Greek letters in the sacred names served to increase their existing power, as did the laying out of names in a symbolic shape such as a cross or a circle. The inscriptions made it possible for the viewer to harness the power of God by seeing and remembering the sacred names and taking them away on the lips or in the heart; but, apt for the medium of panel painting, they also simply showed that power. Indeed, the densest concentration of Greek in Van Eyck’s oeuvre is in the upper register of the Ghent Altarpiece where it is an essential component of the revelation of the divine. In the titles rex regum et dominus dominancium on the mantle of the Deity, Van Eyck employed not only Greek letters but some long-obsolete Western letters: the freely drawn X with tightly-curled tips (at the end of the word rex) and the letter A with a fully-rounded bow (in the word dominancium) are uncials, letters that were used for display script in the Caroline system of scripts and in proto-Gothic manuscripts from the ninth to the twelfth century. For contemporary viewers, these archaic Western

letters must have seemed as ancient, strange and distant as Greek. Van Eyck thus created a sequence of letters that was perfectly calibrated to suggest mystery and incomprehensibility, and, by mingling east and west, the cosmic, boundless nature of God. Greek also appeared in trilingual inscriptions on the Cross: according to the Gospel of John, ‘Jesus of Nazareth King of the Jews’ was written in Hebrew, Greek and Latin. In several images of the Crucifixion by Eyckian painters the titles are depicted in all three languages, suggesting a consistent practice originating with Hubert and Jan van Eyck. On the titulus of the cross held by one of the angels around the altar in the Ghent Altarpiece not only the Greek line but the Hebrew line, which is not very legible at the end, has been transliterated from the Latin (fig. 20.5). The Hebrew letters, in Ashkenazi script, read, from right to left: yod (‫)י‬- a’yin (‫)ע‬- shin (‫ )ש‬/ gimel (‫ –)?( )ג‬aleph (‫)א‬/ yod (‫ )י‬and a final, unidentified letter.25 Although the letters are difficult to make out, the text is clearly meant to read, in Latin: Ies na i [?].26 The choice of gimel was most likely erroneous: it is similar in form to nun (‫)נ‬, which would be suitable to represent N, as it does in other Eyckian paintings.27 It is the consistent use of Hebrew letters that are clearly meant to represent vowels which reveals that this is a system of transliteration and not an attempt to write the language itself. The system is as follows: yod (‫ – )י‬I a’yin (‫ – )ע‬E shin (‫ – )ש‬S nun (‫ – )נ‬N aleph (‫ – )א‬A As with the Greek letters, the Hebrew letters were chosen for their broad phonetic similarity to the Latin equivalents, and the substitutions are not strictly ‘correct’ (although it has a sh-sound, the painter uses shin for a single S, for example). With respect to the origin of this system of Hebrew transliteration, the letter a’yin (‫ )ע‬is not used to represent the vowel

jan van eyck’s greek, hebrew and trilingual inscriptions

297

Fig. 20.5 Ghent Altarpiece, The Adoration of the Lamb, detail of the titles of the cross

E in Hebrew but it is in Yiddish, which suggests that the system follows Yiddish orthography.28 In the Ghent Altarpiece, the titles of the Cross are displayed on the cross held by an angel as one of the arma Christi or Instruments of the Passion. They are presented as the foremost signs of Christ’s victory and were believed to have extraordinary powers of protection. This is how they were understood by Ludolph of Saxony, who, in his Vita Christi, provided a prayer to the titulus crucis which asked for the strength to fight valiantly under the standard of the Lord to combat attacks by the devil.29 Ludolph asserted that every Christian ought to carry the titles of the Cross in his heart, on his lips and even in writing – and indeed, they were a common choice for textual amulets, finger rings and brooches.30 In Eyckian painting, Greek letters not only indicated the presence and power of God but extended to the archangels and warrior saints who acted and fought in his name. They showed the power of the saints and their efficacy as intercessors, encouraging veneration and prayer in the beholder. Thus they appear on the shield of the one of the Knights of Christ in the Ghent Altarpiece

(as noted above) and – though only at the underdrawing stage – in the costume of the Archangel Gabriel in the Washington Annunciation (fig. 20.6). In that picture, the inscription on the edge of the angel’s dalmatic reads agla + elo[y?]: in the name agla, gamma (Γ) and lambda (Λ) are used for Roman G and L and in the name elo[y?], lambda (Λ) and uncial omega (󰰴) for L and O. The omission of these words at the painting stage brings an improvement in visual clarity, suggesting that the decision was made by Van Eyck himself. What was the source for Van Eyck’s knowledge? Greek transliteration appears to have come down to late-medieval Europe along different pathways, partly from the monasteries of early-medieval Europe but also, beginning in the twelfth century, from the universities. At least by the late fourteenth century, however, this kind of knowledge did not depend on a university education, as demonstrated by the English master surgeon John Arderne (1307/8-1377), who operated at the intersection of magic, religion and medicine. Arderne would not have gone to university but he was a learned man with sufficient Latin to write several medical treatises in that language. He had in his

298

susan frances jones

Fig. 20.6 Jan van Eyck, The Annunciation, c.1434-1436, canvas transferred from wood, 92.7 x 36.7 cm, Washington, National Gallery of Art, Andrew W. Mellon Collection (inv. no. 1937.1.39), IRR assembly, detail of the Angel Annunciate’s dalmatic

possession a charm against spasm and cramp, which appears in some copies of his Treatise of Fistula in Ano, Haemorrhoids and Clysters (issued in 1376) and which he ‘used to write in Greek letters that it might not be understanded of the people’ (in the original text, written in Latin, ... scribere istud literas grecis, ne a laicis perspicietur).31 This is the charm: In nomine patris + et filii + et Spiritus sancti + Amen + Thebal + Enthe + Enthanay + In nomine Patris + et Filii + et Spiritus Sancti + Amen + Ihesu Nazarenus + Maria + Iohannes + Michael + Gabriel + Raphael + Verbum caro factum est + Arderne had acquired it from a certain knight, the son of Lord Reginald de Grey de Schirlond near

Chesterfield, who was in Milan in 1368 for the wedding of Lionel of Antwerp, Duke of Clarence, son of Edward III. On that occasion the knight had successfully used the charm on a gentleman who was so afflicted by spasm that he was ‘almost dead from the pain and starvation’. The charm, written on parchment and put in a purse, was placed upon the man’s neck, while bystanders said ‘the Lord’s Prayer and one to Our Lady’. Arderne’s explanation for writing down the charm in Greek letters was to prevent it becoming widely known and used, which he believed would reduce its power (‘lest perchance it should lose the virtues given by God’).32 From the twelfth century, this was a common belief and literary theme among Western authors of works on astrology, alchemy and other occult sciences, which were based primarily on imported Arabic texts.33 Such was the perceived

jan van eyck’s greek, hebrew and trilingual inscriptions

value of this new knowledge that it brought in its wake an insistence on secrecy, and thus on practices such as cryptography.34 According to Richard Kieckhefer, the two main groups involved in the new learning were clerics and physicians.35 This provides a context for a passage by the Franciscan theologian and philosopher Roger Bacon (c.12141294), which speaks of wise men obscuring their thoughts by mixing letters from different alphabets, Hebrew, Greek and Latin, so that they could be understood only by the most ‘diligent and learned’.36 In this practice, various kinds of cryptographic systems occur, using Latin, Greek, Hebrew and even runes.37 In the thirteenth century, a set of recipes inscribed by a student physician on a medical manuscript, including one for the widely-feared incendiary weapon known as ‘Greek fire’, were encoded in Greek and pseudo-Greek characters.38 In the late fourteenth century Greek transliteration was in use at the court of Charles V of France (1338-1380). Between around 1370 and around 1378, as an aid to his personal devotions, the king inserted a series of folios in a book of hours that he had inherited, known as the Savoy Hours. The book’s provenance was both royal and saintly, having been made for Blanche of Burgundy (d. 1348), Countess of Savoy and granddaughter of St Louis of France. The additions included two pseudo-Greek texts, one a prayer, the underlying text of which was French, and the other a series of precepts of good kingship, the underlying text of which was Latin (both are now lost). This is the prayer, as transcribed and translated by Paul Durrieu:39 󰰴 BIEPE MEPE TV MIΛITE BEP’ ΔIEV ΛE ΠEPE A VNITE XAPAEITE KA MIΛITE ΠAP MOP AME’ ΛENΦEP Ȣ EPE ΛVMANITE

O Vier[g]e Mère Tu milite(s) Vers Dieu le Père A Unité, Charité, Qu’a milité Par mor[t] [des] âmes L’enfer, où er[r]e L’humanité

299

The text was produced by the following system: Omega (󰰴) – O Beta (B) – V Rho (Ρ) – R Lambda (Λ) – L Delta (Δ) – D Pi (Π) – P Kappa ( K) – QU Chi (X) – C Phi (Φ) – F Ou ligature (Ȣ) – OU Although the texts are different in kind, their meaning in each case was highly personal: the king’s own vernacular is used for the prayer to the Virgin, and the precepts apply only to him. The texts evoke the two main connotations of Greek – religious piety and ancient wisdom – bringing them together in a single, very exclusive book. The presence of the precepts is in keeping with the belief at the Valois courts that ancient histories could provide exemplars of good government.40 But in the prayer, for the first time in our discussion, we find the combination of Greek letters with a vernacular language, exemplifying the spread of learning outside the universities to a broader public, a process that became increasingly significant in the fourteenth and fifteenth centuries. The Savoy Hours was also historically an extensively personalised book, containing twenty-five pictures of Blanche in prayer, and a ‘Prayer for Myself’ – and this usage was continued by Charles, who added numerous pictures of himself in prayer. The addition of the ‘Greek’ inscriptions to the book shows Charles promoting the image of the wise and pious ruler.41 A pseudo-Greek text, this time in minuscule script, can be found in another (currently untraced) book of hours: one made for a craftsman, the French sculptor Michel Colombe (c.1430, Bourges – Tours, c.1513).42 The colophon, composed by one Petrus Fabri, commemorates the making of the book in 1487 with both a Latin text and a pseudoGreek text transliterated from French – the user’s

300

susan frances jones

native tongue. In addition, the Latin text is signed with his initials (P. F.) transliterated into Greek (Π. Φ.). It is plausible that the scribe Petrus Fabri, who calls himself (in the pseudo-Greek text) pour le temps resident a Bourges, can be identified with the poet and rhetorician Pierre le Fèvre, called Fabri (c.1450-c.1535), who was a native of Rouen. Fabri’s knowledge of Latin is clear from the fact that the chapter on letter-writing in his Le Grant et Vray Art de pleine Rhétorique, published at Rouen in 1521, is a literal translation from Latin sources.43 Fabri was a leading member of Rouen’s Confraternity of the Immaculate Conception, which was famed, among other things, for its annual puy or poetry competition. In 1487, the year that this colophon was written, Fabri was the head of the confraternity and sponsor or ‘prince’ of the puy.44 The book thus represents an interaction – or a social relationship – between a manual craftsman and a man of letters. Colombe clearly regarded it as a sign of his success, since he is celebrated in the Latin text as regni Francie supremi sculptoris.

That Greek cryptography is found in panel painting by the early fifteenth century has been overlooked hitherto. In the so-called Norfolk Triptych, now in the Museum Boijmans Van Beuningen in Rotterdam, dateable to sometime within or around the years 1410 to 1420, the tituli, provided in all three sacred languages, deploy both Greek and Hebrew transliteration (fig. 20.7). Working on a minute scale, the painter has depicted one word on each line: Ihesus in Hebrew, Nazarenus in Greek and Iudeorum in Latin. From right to left the Hebrew line reads yod (‫)י‬, he (‫)ה‬, a’yin (‫)ע‬, shin (‫)ש‬, waw (‫)ו‬, shin (‫)ש‬, clearly meant for ihesus. Unlike Van Eyck, the painter has spelled the word ihesus with an H, but otherwise this system uses the same Hebrew equivalents. In the text of the Greek line, reading [n]asarenus, the painter adopted a similar but not identical system to that Van Eyck would use, substituting lunate sigma (C) for S; rho (Ρ) for R and an unusual sign for the Latin letter N, to which we will return shortly.

a

b Figs 20.7a-b Southern Netherlands, Maastricht or Liège?, (a) Norfolk Triptych, centre panel, Christ as the Man of Sorrows, Christ and the Virgin, and Saints, c.1410-20, oil on oak panel, 33.1 x 16.35 cm, (b) detail, the titles of the cross Rotterdam, Museum Boijmans Van Beuningen (inv. no. 2466)

jan van eyck’s greek, hebrew and trilingual inscriptions

An image of the Last Judgement formerly in Diest (Brabant), also displays a system of Greek transliteration but since it can only be dated broadly, perhaps around 1425-1435, it is not clear whether the system, or the idea, was derived from the Ghent Altarpiece, which bears the date 6 May 1432 (fig. 20.8).45 The painter used Greek letters in the titles rex regum et dominus dominancium on the hem of Christ’s mantle – the same titles that appear, also in Greek, on the mantle of the Deity Enthroned in the Ghent Altarpiece. An inspection of the Greek and Hebrew letters in the Last Judgement, however, shows that for the Greek letters the painter used both different forms and a different range of forms from Van Eyck; and, furthermore, that he knew a system of Hebrew transliteration which he is unlikely to have worked out on the basis of the few Hebrew letters in the Ghent Altarpiece.

301

The word on the robe of John the Evangelist, in red at the centre of the picture, reads, from right to left: ihesus (yod (‫)י‬, he (‫)ה‬, a’yin (‫)ע‬, shin (‫)ש‬, waw (‫)ו‬, shin (‫ – ))ש‬the same formula that was used in the Rotterdam picture. Moreover, the letters running vertically down the edge of the mantle of one of the other Apostles, probably Peter, spell out the names ihesus and maria alternately (see fig. 20.8). ihesus is once more spelled as above (compare fig. 20.7), and maria is also written in Hebrew characters. We can plausibly interpret the four final letters as alef (‫)א‬, resh (‫)ר‬, yod (‫ )י‬and alef (‫)א‬, standing for the Latin A R I A; in that case, the first letter should be a mem (‫)מ‬, standing for Latin M, but clearly it is not. While the inscriptions undoubtedly follow a system, then, it is probable that the painter copied the letters, which are very badly drawn, from a model of some kind, and that he did so without any real understanding.

Right to left: yod, he, a’yin, shin, waw, shin, or IHESUS

Right to left: mem(?), alef, resh, yod, alef, or MARIA

Fig. 20.8 Southern Netherlands, Last Judgement, 1425-1435, oil on panel, 231.5 x 186.5 cm, Brussels KMSKB-MRBAB, (inv. no. 4658), The system of Hebrew transliteration used for the sacred names on the mantle of St Peter (?) is shown to the right

302

susan frances jones

From these examples we can infer that other artists from this period, whether before or after Van Eyck, had independent knowledge of systems of transliteration. Intriguingly, there is evidence that one of the two paintings, the Norfolk triptych, was made in the Mosan region from which the Van Eyck family originated, suggesting that Greek and Hebrew transliteration may have been part of a regional tradition that was available to Van Eyck as part of his workshop training. The question of how Netherlandish painters obtained knowledge of Greek and Hebrew letters is thus pushed further back in time, to the generation before Van Eyck. The introduction of Greek transliteration to panel painting was not the work of a single, exceptionally learned painter (i.e. Jan van Eyck) but reflected the interests of members of the patron class. The use of transliteration systems to depict the titles of the Cross by 1410 or 1420 presumably responded to a desire for authenticity in the representation of the titulus, which was one of the most holy relics of Christianity and one that, in the early fifteenth century, was lost to view (it would not be rediscovered until 1492). To show the trilingual titles in full became characteristic of Eyckian painting. Indeed, a model for the titulus in which the Hebrew and Greek lines were transliterated was arguably available in Van Eyck’s workshop.46 Examination of these earlier systems of transliteration indicates that Van Eyck’s excluded certain

medieval Greek forms that were standard in the West: in an illustration of the five letters domin from the beginning of the word dominus, inscribed on the edge of Christ’s mantle in the Diest Last Judgement (fig. 20.9), the last three signs are the so-called M-siglum, sometimes called ‘western M’ , which looks like two letter Cs positioned back-toback, and joined with a crossbar; a curved-Y form representing Latin I, and finally a sign representing N.47 The painter of the Norfolk Triptych used the last sign to represent the Roman letter N in the word [n]azarenus (compare figs 20.9 and 20.10a). Earlier texts and images indicate that these equivalents were of very long-standing: the medical cryptogram of the thirteenth century, mentioned above, employs both the M-siglum for M and the sign -( for N. These forms were in use in Van Eyck’s day: his contemporary, Rogier van der Weyden, deployed the same curved-Y form to represent I in the title inri in his Descent from the Cross (compare figs 20.9 and 20.10b). They do not, however, feature in the sacred names and titles that Van Eyck partly transliterated into Greek, in which the letters have phonetic value. The only example of M-siglum in Eyckian painting (known to this writer) occurs in the Ghent Altarpiece, where it is shown on a very small scale in the decoration of the red hat worn by one of the characters in the group of Old Testament prophets in the Adoration of the Lamb panel (fig. 20.11).

Fig. 20.9 Detail of Fig. 20.8 showing the first five letters of the word DOMINUS on the edge of Christ’s mantle

jan van eyck’s greek, hebrew and trilingual inscriptions

Figs 20.10a-b (a) detail of Fig. 20.7a; (b) Rogier van der Weyden, Descent from the Cross, detail, titulus, c.1435, oil on oak panel, 220 x 262 cm, Madrid, Museo Nacional del Prado, cat. P-2825

Together with the larger Hebrew letters on the hat, these Greek letters are ornamental forms which are presumably a socio-historical marker for this group of figures. It is therefore possible to construct an argument that Jan van Eyck developed a Greek alphabet that was different from those already available to him, and that he omitted to use particular medieval Greek signs because they did not meet his needs. This would be in keeping with his intellectual curiosity. He could also have found opportunities at the Burgundian court and in Bruges to make contact with men of letters.48 Broadly speaking his Greek alphabet is in keeping with medieval norms, as shown by his use of uncial omega (󰰴), lunate sigma (C) and – notably in the Portrait of a Man (Self-Portrait?) – a rather inaccurate form of lambda, in which the left leg is too short (fig. 20.1).

303

From all these observations we can draw some preliminary conclusions, focusing mainly on Van Eyck’s motto. First, in Van Eyck’s day, the practice of Greek transliteration was probably less of a rarity than is often supposed. It was not the preserve of any one single group or profession but had filtered into various groups, including, but not limited to, scribes, clerics, physicians, surgeons and painters, and those groups themselves represented different levels of knowledge and a wide variety of practices. Clerics, for example, were probably among those who composed and wrote textual amulets, in which sacred names of God feature so prominently.49 The Greek alphabet had also found its way into the realm of literature written in the vernacular. In the mid-fourteenth century a Greek alphabetical table complete with Latin equivalents appeared in copies of the Travels of Sir John Mandeville, which was composed around 1352, in French, and, in each of two recensions (Insular and Continental), offered a body of six alphabets: Greek, Egyptian, Hebrew, Saracen, Persian and Chaldean (admittedly, the alphabets are usually mangled).50 While it remains difficult to make an assessment of just how many people possessed this kind of knowledge, it is clear that it did not denote a particular social sphere but must have extended across the boundaries of the university, church, court and town. In Van Eyck’s day, moreover, it did not necessarily indicate a university education: within the group of medical practitioners, for example, physicians were university-educated but master surgeons were not. That this kind of knowledge could have circulated somewhat more widely than is often supposed is possible because the level of Greek it represented was basic. All that was needed was a simple alphabetical table. In the Carolingian period, tables of this kind might be appended to bilingual manuscripts in Greek and Latin as an aid to understanding. Some offered a simple series of letters, others a complete alphabetic table, with a phonetic transcription and equivalent letters from the Roman alphabet.51 The Greek and Hebrew alphabets in late-medieval texts of Mandeville’s Travels were

304

susan frances jones

Fig. 20.11 Ghent Altarpiece, The Adoration of the Lamb, detail, ‘M-siglum’ in a band of Greek and pseudo-Greek letters on the hat of one of the Prophets of the Old Dispensation

apparently derived from medieval compendia which ultimately went back to alphabets of the early medieval period.52 Secondly, Van Eyck was not exceptional in using Greek transliteration even among painters. On the available evidence, it is likely that he learnt Greek cryptography in the first instance not from scribal practice or books but as part of his training in the craft. This is affirmed by his ability to transliterate Latin phonetically into Hebrew letters, which was not a feature of scribal colophons. It is a reminder that the most apt context for the study of Van Eyck’s motto is the world of painting and painters, not of scribes. Painters (and other craftsmen) had long been required to make and to depict letters and inscriptions, making Van Eyck’s painted inscriptions a development of existing practice. The evidence suggests that in the late-medieval period inscriptions transliterated into Greek were regarded as suitable for books of hours. These books

were essentially for personal devotion. The Savoy Hours, which survives only in part, was a particularly luxurious example, made in Paris in the workshop of Jean Pucelle.53A description of the book in an inventory of 11 April 1380 (known only from a copy) tells us that it had a splendid binding embroidered with pearls and decorated with seven gold fleurs-de-lys; each of the gold clasps was ornamented with two balas rubies, two sapphires and five large pearls.54 At the time, the book was in the estude of Charles V in his private apartments at his château at Vincennes.55 As noted above, the ‘Greek’ writings in this book were there for the king’s eyes: they denoted texts that were highly personal or even secret. That pseudo-Greek inscriptions appeared in such a book sheds light on Van Eyck’s practice of reserving his motto for small-scale, autonomous panel paintings rather than larger, more public works such as the Ghent Altarpiece, the Virgin and Child with Canon Joris van der Paele

jan van eyck’s greek, hebrew and trilingual inscriptions

and the lost Virgin of Nicholas van Maelbeke (which was probably a large painting). It suggests that complex word puzzles of this kind were considered to be appropriate to objects intended for private settings. The text of the motto is very short and the code is deliberately transparent. This is in contrast to the scribal colophons, prayer and precepts referred to above, which are relatively long and contain numerous Greek letters. In Van Eyck’s motto, the Greek letters do not obscure the Middle Dutch text but rather co-exist with it in a delicate visual and linguistic balance. The fact that the text is visibly in a vernacular language; the fact that only three Greek letters occur, and the familiarity of the Greek letters themselves – sigma (C) and chi (X) were probably the most widely known of all Greek letters because they are in the monogram of Christ (ihc xpc) – all indicate that Van Eyck wanted the code to be breakable. While he must have anticipated an educated audience, he must also have wanted to reach the maximum number of people within that group: his audience would not necessarily exclude men of learning, but it would not be composed only of such men. Since the motto undoubtedly drew attention to the intellectual refinement of the painter and his audience, it must have had a social value. That Van Eyck wanted to be perceived as socially elevated and well educated is evident from his portrait of his wife Margaret, who is shown in expensive attire, and, furthermore, ‘speaking’ Latin – implicitly an ability that Van Eyck shared. This mode of representation may have been Margaret’s due, but there may also be an element of social pretension in this portrait.56 The Greek letters in Jan van Eyck’s motto were clearly not meant to keep the text secret, even if they hint at secrecy or mystery in relation to Van Eyck’s ability. It was clearly important to Van Eyck that the viewer could at once see the Greek letters and read the words als ich can, deriving meaning from the combination of the two. The Greek letters showed the text to be personally meaningful

305

to Van Eyck as an individual whilst charging it with associations of intellectual sophistication and social status. Given that Greek letters were used consistently in Van Eyck’s painting to represent sacred names, it is very plausible that the motto had a religious meaning.57 According to Josef Koerner, writing on Van Eyck’s lost Holy Face, the chi (X) in the centre of the word ixh in the motto was a reference to Christ, adding yet another layer to an existing play on ixh and eyck. In Koerner’s view, Van Eyck’s motto was intended to evoke a comparison between the painter and Christ, and between divine and human image-making.58 The layout of the motto, consisting of three blocks of three Greek letters, is strongly evocative of the sacred monogram of Christ, ihc xpc. This is relevant to its interpretation because in Van Eyck’s inscriptional practice as a whole, the external form of an inscription tends to reinforce the meaning of the words. His motto, furthermore, was not literary but emblematic: like the monogram of Christ, it could be ‘made’ in any one of a diverse range of materials, from ink to stone to gold, but it always remained constant in form.59 The design, with regular interruptions between short ‘Greek’ words, recalls the Eyckian way of painting the titles of the Cross (see fig. 20.5) – or, indeed, the three sacred names displayed in three sacred languages on the robe of Christ in a lost Eyckian Holy Face.60 In support of such a reading, Van Eyck would certainly have perceived the two languages in the motto to have different hierarchical value. Hebrew and Greek were the two most sacred languages, with Hebrew the higher in dignity, and thus closer to God; Latin was the third sacred language and the vernacular languages were the lowest of all. Seen in this light, the motto is an inscription that deliberately combines languages from opposite ends of a hierarchy: the painter’s personal vernacular with a high-grade language of abstraction and of divinity. It could reasonably be interpreted to mean that he has made the painting to the best of his ability and with God’s grace.

306

susan frances jones

Fig. 20.12 Ghent Altarpiece, John the Baptist Enthroned, detail

Van Eyck’s motto is thus complex and multilayered: it was no doubt a tool for social positioning, but it was arguably also an expression of ideas and beliefs about his individual ability that were complex, and deeply personal.61 Other texts show Van Eyck cultivating the potential for a single sign to be read in a multiplicity of ways. In the panel of St John the Baptist in the Ghent Altarpiece, for example, the saint’s raised finger seems to activate the cross-shaped punctuation mark at the beginning of the inscription, so that it reads simultaneously as textual punctuation and as the Cross of Christ (fig. 20.12). The same cross can be interpreted as the trace of a gesture made in the air, perhaps referring to the gesture of blessing made by a priest at the altar, to consecrate the Eucharistic elements.62 In the Portrait of Jan de Leeuw, moreover, the inscription on the frame contains a double chronogram, in which particular Roman letters (V, X, L, C and so forth) can also be seen as numerals. Who were the viewers and audiences of Van Eyck’s motto? That the text is in Middle Dutch indicates that Van Eyck designed it first and foremost for viewers in the Netherlands, who knew the language. Since the motto appears on the Portrait of Margaret van Eyck and the Portrait of a Man (Self-

Portrait?), his intended audiences may have included members of his own social circle. Nonetheless, the heraldry of the Marian triptych now in Dresden, which also bears the motto, has been linked to the Genoese merchant family Giustiniani.63 The small scale and portability of such works meant that potentially they could be used to convey ideas about painting to eminent viewers throughout Europe. It is widely accepted that the words als ich can were a modesty formula, allowing Van Eyck to appear modest whilst in fact asserting his superiority over every other painter. This may be the case, for he must have recognised how exceptional his ability was. A different motivation for signing and dating his work and adopting a motto would be to set an example to other painters. For the beholder, the motto was an encouragement to perceive the object as a product of individual aptitudes and skills. That such ideas were discussed and emulated among painters is clear from the addition of the motto als ich chun to a painting of the Crucifixion (Vienna, Belvedere, 1449) by the Swabian painter Conrad Laib.64 While Van Eyck’s practice of signing and dating does appear to have been new in panel painting of the region – there is no surviving evidence that

jan van eyck’s greek, hebrew and trilingual inscriptions

earlier painters systematically signed and dated their works, or had personal mottoes – that novelty is not explained solely by the theory that Van Eyck borrowed from scribal or notarial practice. As suggested by his deliberate revival of a late-twelfthcentury mixed-hand majuscule script, shown in the presumed self-portrait (fig. 20.1), it is plausible that Van Eyck derived the practice from the signatures of esteemed craftsmen of the Romanesque period, whose names he saw at the edges of old and venerable objects that interested him. These old signatures were crafted in the traditional epigraphic materials of gold, bronze and stone – the same materials that Van Eyck imitated on the frames of his paintings. The motto never appears separately from Van Eyck’s name and the date, and, importantly, it contains a square form of sigma (see figs 20.1, 20.2). In other ‘Greek’ inscriptions – such as the titles of the Cross, the hem of the Enthroned Deity’s mantle, and the yecyc tiles in the Ghent Altarpiece – he used a lunate sigma. This suggests that he wanted his motto, like his signature, to hark back to a distant, pre-Gothic era. Van Eyck’s aim was arguably to create a continuum between his ‘modern’ painting and artefacts from the fardistant past made in his own region. By interpreting Van Eyck’s motto and his signatures in the light of ancient and humanistic art ‘theory’ rather than the context of Bruges and the Burgundian court we risk misidentifying the sources of his knowledge and misunderstanding the contemporary perception of his oeuvre. NOTES * For their help in reading and interpreting painted texts in Greek and Hebrew I am deeply grateful to Irene Zwiep, Robert Ireland and Orly Moshenberg. I owe special gratitude to the late Professor Chimen Abramsky for looking at Van Eyck’s Hebrew inscriptions with me. I would also like to thank Lorne Campbell, Stephan Kemperdick, John Lowden, Didier Martens, Catherine Reynolds, Gervais Rosser and Cyriel Stroo for informative conversations and Lee Preedy for her careful editing of the text. 1 Berschin 1988, p. 19. 2 The sizes of the signed and dated works are: Portrait of a Man (Self-Portrait?) (1433, The National Gallery, London), 33.1 ≈ 25.9 cm; Virgin and Child with Saints Catherine and Michael and a Donor (1437, Dresden, Gemäldegalerie Alte Meister), panel 33.2 ≈ 27.2 cm (centre panel with original frame); the Virgin by a Fountain (1439, Koninklijk

307

Museum voor Schone Kunsten, Antwerp), 24.8 ≈ 18.1 cm and the Portrait of Margaret Van Eyck (1439, Groeningemuseum, Bruges), 41.3 ≈ 34.5 cm. 3 The system as described here is based on the Ghent Altarpiece and the signed and dated, and generally accepted works of Jan Van Eyck; it does not include the work of Hand G of the Turin-Milan Hours, which deserves separate consideration. Van Eyck used the forms of Greek available to him in his own culture. Van Eyck’s majuscule omega is shaped like a large, rounded W, as the majuscule omega shaped like an inverted horseshoe (Ω), was unknown in the West in the Middle Ages (for which see Berschin 1988, p. 30). Similarly, Van Eyck’s sigma is either a square form or a lunate form, rather than the fourstroke sigma (Σ). The Greek alphabet was almost always written in majuscule form in the West. 4 Scheller 1968, p. 136. 5 Ibid., p. 137. 6 Bischoff 1951, pp. 27-55, esp. pp 32-39. 7 Bénédictins de Bouveret 1965, no. 94, p. 12. 8 Ibid., no. 2355, p. 295; Kahsnitz 1971, p. 373 (the inscription is arranged in two columns). 9 For this aspect of transcription, see Berschin 1988, p. 30. Kaczynski 1988, pp. 28-30; Bischoff 1951, p. 36, n. 4. 10 Smeyers 1996, p. 404. 11 Scheller 1968, p. 137 provides a coherent explanation for the use of chi (X) in the motto. 12 See Dhanens 1980, p. 180. Similarly, in the only surviving example of his handwriting, on the drawing of a man in Dresden (Kupferstichkabinett), Van Eyck wrote den auge for ‘the eyes’, but on the frame of the Portrait of Jan de Leeuw, he used the word oghen. 13 Smeyers 1996, p. 404. 14 Scheller 1968, p. 139. For Fazio, see Baxandall 1964, pp. 102, 103. 15 Campbell 1998, p. 222. For Van Eyck’s Greek and Hebrew inscriptions in relation to sacred names and formulae, see Mély 1921, pp. 1-16; the most comprehensive and up-to-date text, however, is Paviot, Goren 2006, pp. 68-72. 16 For an overview of the arguments see Campbell 1998, p. 220. 17 Ibid., p. 222. 18 Dr Robert Ireland (University College London) suggested to me that Van Eyck may deliberately have decided to use upsilon here to make the inscription appear more Greek (in conversation). 19 Among the deaths announced to the General Chapter of the Carthusian Order in 1468 was that of the prior of the Charterhouse of the Hanseatic city of Rostock ‘Domnus Tymotheus’: see Sargent, Hogg 1985, p. 55. I thank Lorne Campbell for this reference. 20 Campbell 1998, p. 220. 21 I am following here Kieckhefer 1989, pp. 8-17 and passim. 22 Ibid., 1989, p. 40. 23 Thomas, Pavitt 1922, p. 107. 24 For textual amulets, see Skemer 2006, pp. 78-79, 125-27, 144151 and passim; Kieckhefer 1989, pp. 75-80. 25 In the Ghent Altarpiece, the plaque is painted partly on top of pre-existing paint layers; see the ‘closer to van eyck’ website (http:// closertovaneyck.kikirpa.be). 26 That this is a system of transliteration was recognized by Irene Zwiep, Professor of Hebrew and Jewish Studies at the University of Amsterdam, when at the Warburg Institute in 1996. The same solution was proposed earlier by Eugen Schiltz in relation to the Eyckian Crucifixion and Last Judgement in The Metropolitan Museum, New York, for which see Paviot, Goren 2006, pp. 59-60. Paviot and Goren, in contrast, have focused on reading Hebrew words in this and other Hebrew texts in Eyckian painting, for which see Paviot, Goren 2006, pp. 56-60. 27 This reading is supported by the Hebrew lines in the tituli in both the Eyckian Crucifixion and Last Judgement in The Metropolitan Museum, New York, in which the texts are more extensive. The

308

susan frances jones

Hebrew line in the Crucifixion reads (from right to left): yod - shin / nun - aleph - zayin / resh - a’yin - zadi / yod - waw (?) - daleth (?) – and an unidentified letter. This is intended for the Latin text: IS / NAZ / REX / IUD. The Hebrew line in the Last Judgement reads: yod - a’yin / nun - aleph – peh (?) / resh - a’yin - zadi / yod - waw - daleth - a’yin – waw (?). This can be ‘read’ in Latin as: IE / NAP / REX / IUDEU. 28 I owe this information to Irene Zwiep; for the use of this equivalent see also Fuks 1957, pp. xxxvii and xxxiv. 29 See Bodenstedt 1944, pp. 134-35. 30 To give but one example, see Dalton 1912, no. 881, p. 139 (fourteenth century). 31 The description of the charm, how Arderne acquired it, and its content, are all recorded in a fifteenth-century Latin copy of Arderne’s Practica de Fistula in Ano, &c. (The British Library, Sloane Ms 2002, fols 79-80v). Another version of the same treatise, written in English in the early fifteenth century (British Library, Sloane Ms 6, fols 141-154v) lacks the description of the charm. For the 1910 edition of the English version the editor, D’Arcy Power, copied the text about the charm from Sloane Ms 2002, providing both the Latin text and an English translation. It was presumably D’Arcy Power who introduced the archaic term ‘understanded’ rather than using ‘understood’; see Power 1910, pp. 102-103 (Latin), pp. 103-104 (English) and p. 135, n. 102/8. For Arderne’s charm, see also Skemer 2006, pp. 144-145. 32 Power 1910, p. 104. 33 Kieckhefer 1989, pp. 117-118. 34 Ibid., pp. 140-43. 35 Ibid., pp. 119. 36 For this text in relation to Van Eyck, see Smeyers 1996, p. 406 and Paviot, Goren 2006, pp. 53-54. 37 For runic and other cryptography, see Kieckhefer 1989, p. 141 and nn. 28 and 29. 38 Corner 1936, pp. 745-750. 39 For the transcriptions, see Durrieu 1911, pp. 514, 536-539. The relevant part of the manuscript was destroyed in the fire at the Biblioteca Nazionale Turin in 1904. The surviving fragment is in New Haven, Yale University, Beinecke Rare Book and Manuscript Library, Ms 390. 40 Buettner 1992, p. 80. 41 For the use of Hebrew to extol the king’s wisdom, see Kupfer 2008, p. 84 and passim. 42 Grandmaison 1912, pp. 77-79. I am grateful to Lorne Campbell for bringing this article to my attention. 43 Clark 1986, p. 123. 44 Mantovani 2000, p. 41. 45 The system of Greek transliteration was worked out by De Ridder; see De Ridder 1989-1991, pp. 119-121.

46 This would explain the texts in the Crucifixion in the Gemäldegalerie, Berlin and The Metropolitan Museum, New York. A different usage was adopted in the Ca’ d’Oro Crucifixion. 47 On these forms, see Kaczynski 1988, p. 29. 48 The pursuit of correct forms of Hebrew preoccupied learned men at the court of Charles V of France, resulting in a genuine Hebrew inscription in a Mandeville’s Travels made for the king; Kupfer 2008, pp. 60-91. 49 Skemer 2006, pp. 128-129. 50 See Letts 1949, pp. 151-160; Kupfer 2008, pp. 59-60; Paviot, Goren 2006, p. 54. 51 Berschin 1988, pp. 101, 128, 142, 209-210. 52 Kupfer 2008, p. 59. 53 Baltimore 1998, pp. 28-29, 31-32, 39, 176-178. The extant pages measure 20.1 ≈ 14.7 cm, trimmed up to the margins; Durrieu 1911, p. 534 suggested that the original size was as large as 25.0 ≈ 17.0 cm. 54 Durrieu 1911, p. 519. 55 Durrieu 1911, pp. 518-519, 534. 56 These ideas – including the notion that the portrait may be somewhat pretentious – were discussed by Catherine Reynolds in ‘Gossaert and the Netherlandish Tradition’, lecture, The National Gallery, London, 4 March 2011. 57 This was the opinion of De Vos 1983, p. 3 and Gludovatz 2005, p. 134. 58 Koerner 1993, p. 107. A visual association with Christ has been perceived by Susie Nash in Van Eyck’s presumed self-portrait, in which the number 33 of the date, written in Arabic numerals, is the age of Christ at his death; see Nash, 2008, p. 154. 59 Gludovatz 2005, p. 135 points out the emblematic quality of the motto. 60 The version closest to the lost work – one of several Eyckian archetypes – is probably that sold at Sotheby’s New York, 16 May 1996, no. 181. 61 For the notion of a ‘explicit layered quality’ in the Renaissance notion of the individual, manifest as early as the fourteenth century, see Martin 2000, pp. 11-31. 62 With respect to a prayer to the archangels, Eamon Duffy observed that the sign of the cross in the text of a prayer was a prompt for the reader to make the sign in actuality; Duffy 1992, p. 271. 63 For the Dresden Triptych, see Neidhardt, Schölzel 2005, pp. 20-21. 64 The full text is: d PFENNING / 1449/ ALS ICH CHUN. For Laib see Kemperdick 2010, p. 56.

Fig. 21.1 After Jan van Eyck, Portrait of Isabella of Portugal, 17th century after an original of 1429, pen and brush on paper, 45 x 41 cm, Lisbon, Arquivo Nacional da Torre do Tombo

21

Early Texts on Some Portraits by Jan van Eyck Stephan Kemperdick

ABSTRACT: Jan van Eyck’s 1429 portrait of Isabella of Portugal, his earliest documented work, is known through a written account of the Burgundian excursion to Portugal and through a drawn copy of much later date. Both these sources can shed some light on aspects of the lost original, among them the number of its copies. Likewise, the descriptions given by the art collector Peeter Stevens on two preserved portraits by Van Eyck, the Man in a Red Turban in the National Gallery, London and the so-called Cardinal Niccolò Albergati in Vienna, are examined with respect to the possible identities of the sitters. It can be shown that Stevens’s information on the Vienna portrait is not based on an inscription of the now lost original frame, and further that the sitter is not a Carthusian and thus not Albergati.

—o— The earliest known work by Jan van Eyck was a portrait, made in 1429 at Avis in Portugal. It showed Isabella, daughter of John I of Portugal and prospective bride of Philip the Good, Duke of Burgundy. The Duke had sent a delegation of trusted men, among them his court painter Jan van Eyck, to Portugal in order to negotiate the marriage. Thus the portrait was no doubt made by order of the duke, and sent off to Flanders immediately, in order to present the appearance of the infanta to her future husband. Our information about this portrait is provided by an account of the Portuguese mission, probably written shortly after its accomplishment and published in a French version by H.W. James Weale in 1908.1 According to the account, the Burgundian ambassadors reached Avis on 12 January and

opened negotiations with the king and his sons. Two days later the ambassadors finally received une cedulle par escript, and with that they were admitted to the princess, and had her likeness painted au vif by maistre Jehan de Eyk, varlet de chambre de mon dit signeur de Bourgoingne et excellent maistre en art de painture (‘master Jehan de Eyk, valet de chambre of my said Lord of Burgundy and excellent master in the art of painting’).2 The report suggests that it was not easy to see the princess and thus Van Eyck would not have had much time to work in her presence. He probably only made a sketch of her features and her dress, and worked on the painting afterwards in a studio. Meanwhile, the ambassadors were collecting information on the reputation of the princess, which was finally written down and dispatched in four copies to the duke on 12 February by four messengers; two travelling by land and two by sea. At the same time, Isabella’s portrait was also sent, and since Weale’s publication in 1908,3 scholars have become used to assuming that Jan had painted two versions, one to go by ship, another to go by land. This, however, is in no way substantiated by the account. There it is emphasized that each of the four messages contained the complete report on the princess, and it continues: Aussi luy [the duke] envoyerent ilz [the ambassadors] la figure de la dicte dame faicte par painctre, comme dit est (They [the ambassadors] also sent him [the duke] the picture of the said lady, made by the painter, as it is said’).4 Nothing is said about two copies, or that

312

stephan kemperdick

each message was accompanied by one – which would have made four portraits, of course. My first point, therefore, is that we have to abandon the idea that there were two versions of the Isabella portrait of 1429, let alone two different versions. While the original painting is lost without a trace, there is a well-known drawn copy, probably dating from the seventeenth century, which seems to reproduce the work faithfully. After being in private hands, it entered the Arquivo Nacional da Torre do Tombo in Lisbon, where it still remains (fig. 21.1).5 The drawing clearly shows the original portrait of the princess set into a frame which could only have been added later, after the wedding had taken place in 1430, for it bears the combined initials of Philip and Isabella, the flint and steel of the Order of the Golden Fleece and, above all, a French inscription which in translation reads: ‘This is the portrait which was sent to Philip, Duke of Burgundy and Brabant, of Lady Isabelle, daughter of King John of Portugal and Algarve and Lord of Ceuta by his conquest, who was since then wife and spouse of the said Duke Philip.’ The inscription clearly confirms that this is a reproduction of the portrait of 1429. The frame around the portrait is unlike those known from Jan van Eyck’s paintings in both its moulding and decoration, which strongly resembles the borders in illuminated manuscripts. However, it is obviously not a kind of passepartout made of paper or parchment but a real wooden frame, as the shaded inner moulding suggests.6 The tendrils with leaves on the inner band, and the acanthus on the outer, ending in umbels, can be compared to Flemish illumination, for example some manuscripts of the Bruges Gold Scroll Group (fig. 21.2).7 Therefore we can conclude that this frame was made roughly around the middle of the fifteenth century. Obviously, it was still in existence when the copy shown here was drawn maybe two hundred years later. The Eyckian picture itself can clearly be distinguished from the later frame; the painted columns and lintel belong to the former, while the shaded

inner mouldings already form a part of the added frame. The inscription LINFANTE  ·  DAME ISABIEL on the lintel refers to the still unmarried princess and thus must be part of Van Eyck’s original work. As the caption is French it can only have been addressed to Duke Philip himself. The portrait has rightly been compared to some other early Eyckian works, especially the figure of the Cumaean Sybil from the Ghent Altarpiece, which was clearly inspired by the foreign dress of the princess and thus must have been conceived and executed by Van Eyck after his return from Portugal.8 The parapet on which Isabella rests her hand resembles the so-called Timotheos in London, a portrait dated 10 October 1432.9 However, the way light seems to fall into the painting, making the space around the figure look like a rounded, shallow niche that opens between two columns, is rather unusual in Eyckian portraits. Instead, it strongly resembles the niches of the donors on the outside of the Ghent Altarpiece. These dark niches, which are lit, like the Isabella portrait, on one side by a diffuse light, are essential for the structure of the whole outside of the altarpiece, especially with respect to the painted statues of the two St Johns. In the isolated portrait of Isabella, the motif seems to be more accidental. Therefore, I would like to propose that the outside of the Ghent Altarpiece served as a model for the concept of Isabella’s portrait and that, accordingly, this part of the retable was already executed before Jan van Eyck embarked on his Portuguese journey in October 1428. As mentioned above, Van Eyck probably made a drawing in front of Isabella, but what was sent to Flanders must have been a painting: even in the drawn copy we sense that the original rendered not only different colours and textures – of skin, fabric, fur and stone – but also shades and cast shadows in the manner that is characteristic of Van Eyck’s oil paintings. However, it is unlikely that he made a panel painting in oils; not so much because it would have been such a laborious and time-consuming task but because an oil painting may well not have been dry when the time came for it to be sent. As

early texts on some portraits by jan van eyck

Fig. 21.2 Master of the Gold Scrolls Group, Souls carried into Heaven, Book of Hours, Bruges, c.1430-1440 Private collection

it turned out, Jan had four weeks to execute the painting, but if by chance the accounts of the character of the princess had been ready at an earlier date, it would certainly have been necessary to have the likeness ready to ship as well. Thus the portrait of Isabella might have been executed in some kind of tempera – maybe even on parchment or fine canvas. And it was small: the overall size of the drawn copy is around 45 ≈ 41 centimetres, so the picture at the centre measures around 26.5 ≈ 20 centimetres, thus corresponding to the size of some later portraits by Van Eyck, such as those of Baudouin de Lannoy (26 ≈ 19.5 cm), the Man in a Red Turban (25.7 ≈ 19 cm) or Jan de Leeuw (24.5 ≈ 19 cm). Obviously, the drawing in Lisbon is a life-size copy of the Eyckian painting and its frame. The corresponding measurements further confirm that the lost original of the Portrait of Isa-

313

bella of Portugal was indeed a work by Jan van Eyck who seems to have already used his standard format for small portraits in the late 1420s. The later fate of the original 1429 portrait is completely unknown. Certainly it cannot be identified as a portrait mentioned in the Mechelen inventories of Margaret of Austria. That work is described in the 1516 inventory as: Ung moien tableau de la face d’une Portugaloise que Madame a eu de don Diego. Fait de la main de Johannes, et est fait sans huille et sur toille sans aucun couverte ne feullet (‘A painting of medium size of the face of a Portuguese woman which Madame had received from Don Diego. Made by the hand of Johannes, and made without oils and on canvas, without any lid or cover’).10 In the 1523 inventory it is described thus: Item, ung aultre tableau de une jeusne dame accoustrée à la mode de Portugal, son habit rouge fouré de martre, tenant en sa main dextre ung roulet avec ung petit sainct Nicolas en hault, nommé la belle Portugaloise (‘Item, another painting of a young lady dressed in Portuguese fashion, her red gown trimmed with marten, who in her right hand is holding a roll with a little Saint Nicholas at the top, called the beautiful Portuguese woman’).11 As Isabella’s identity was very clearly stated on her 1429 portrait, and Margaret’s inventories always give the names of the noblemen and noblewomen portrayed, they would hardly have referred to Margaret’s great-grandmother merely as ‘a beautiful Portuguese woman’. Thus it must be a different painting, but if the attribution to ‘Johannes’ – to Van Eyck, that is – is correct, and the picture indeed showed a Portuguese lady, it might have been another work that Jan had carried out during his trip to Portugal, especially as it was painted on cloth without oils – exactly what we can assume for the Isabella portrait. Let us now turn towards two extant portrait paintings by Jan van Eyck, both of which were first described in the same seventeenth-century text. One of them is the famous Man with a Red Turban of 1433 in the National Gallery London (fig. 20.1).12

314

stephan kemperdick

It is first recorded in the Antwerp collection of Thomas Howard (1585-1646), Earl of Arundel, where the merchant and collector Peeter Stevens (1590-1668) saw it around 1642-1644. In the annotations Stevens made to his copy of Karel van Mander’s Schilderboeck, he mentioned the panel in the following words:13 Noch byden Grave van Arundel, hebbe hier tot Antwerpen gesien een cleyn conterfaisel vande Hertoch van Barlaumont dato 1433, seer curieux gedaen. ‘Barlaumont’ is undoubtedly a corrupted form of Berlaimont, the name of a noble Hainault family. There were several seigneurs among them, but never a duke, and thus Lorne Campbell’s suggestion that Stevens confused hertoch (duke) with heer (seigneur) is convincing.14 The name Berlaimont cannot be found on the preserved gilded frame of the portrait, which bears only the line ‘ioh[ann]es de eyck me fecit an[n]o m°cccc° 33° 21 octobris’ and the artist’s motto ‘aΛc ixh xan’. Consequently, there are only two plausible explanations for how Stevens came to identify the sitter as a Barlaumont: either the name was written on the back of the panel, for example on a piece of paper,15 or the information was provided by the owner of the painting, the Earl of Arundel, which is the more likely in my view. However, this identification of the man portrayed has received very little attention. This is certainly due to the fact that the painting is often considered to be a self-portrait of Jan van Eyck, an assumption that is hypothetical and poses a certain problem with the age of the sitter: W.H. James Weale calls the man ‘a well-to-do merchant of about sixty-five years of age’,16 and this estimate does not seem too far out, given his wrinkles and his aged-looking eyes. At any rate, the man can hardly be younger than fifty, and if one retains the identification with the painter himself, one has to push back Jan van Eyck’s birth to around 1380.17 Of the Berlaimonts, on the other hand, little is known, apart from the interesting fact that they had family ties with the Lannoys, their neighbours in Hainault: Marie, daughter of Fastré de Berlaimont, married Hugues, seigneur de Lannoy, in

1405, while Adrienne, daughter of Jacques de Berlaimont, seigneur d’Anseroul et de Solre-le-Château, became the wife of Baudouin de Lannoy in 1433.18 In turn, Baudouin de Lannoy was acquainted with Jan van Eyck from at least 1428 on, when both of them were members of the Burgundian party sent to Portugal in order to ask for the hand of the Infanta Isabella. Several years later, after 1432, Jan painted the portrait of Baudouin de Lannoy that is now in Berlin.19 Thus there would have been at least an indirect connection between the Berlaimont family and Jan van Eyck, and members of that family certainly had ample opportunity to hear the praise of the painter’s skills and maybe even to admire paintings by his hand. Furthermore, the abovementioned Jacques de Berlaimont (c.1380-1445), the father-in-law of Baudouin de Lannoy, was a counsellor to Philip the Good, Duke of Burgundy, Van Eyck’s patron, and was also one of the witnesses at the betrothal of Jacqueline of Bavaria and John of Brabant in Biervliet in 1417, where John of Bavaria, Jan van Eyck’s first documented patron from at least 1422 to 1425, was present as well. Dieter Jansen has further suggested that the portrait in question might have passed into the Arundel collection through a later Berlaimont heir: Isabelle-Claire, daughter of Florent, comte de Berlaimont, had married Philippe-Charles d’Arenberg, duc d’Arschot, in 1620, and Arundel might have bought paintings from Arenberg’s estate – which is again an unproven assumption.20 In summary, Stevens’s remark on the Man with a Red Turban might as easily be erroneous as reliable – at present, further data to check it are lacking.21 But since the identification of the sitter as Van Eyck himself also remains hypothetical, the alternative identification as a member of the Berlaimont family cannot be ruled out completely. While Peeter Stevens’s comments on the London portrait have found very little response in the scholarly debate, exactly the opposite is true of his remarks on another male portrait, the likeness of an elderly man, that is now in Vienna (fig. 21.3),

early texts on some portraits by jan van eyck

preserved without an original frame or any old inscriptions. It lost a frame around 1720, when all four edges of the panel were cut into an octagonal shape in order to fit it into an oval that would be integrated into the Baroque wall decorations of the Vienna Stallburg.22 In the late nineteenth century, the work regained its rectangular format, but the restorations at the edges and lower margin of the picture are clearly recognizable. Yet there is a portion of the original red marbling preserved at the centre of the back of the panel (fig. 21.4), which looks similar to the reverse of the Portrait of Margaret van Eyck, dated 1439.23 The Vienna portrait was acquired by Archduke Leopold of Austria in 1648 and was listed in his 1659 Brussels inventory as: Ein Contrafait von Öhlfarb auf Holcz des Cardinal Santa Cruce. In einer halb grünen vnndt halb zier verguldten Ramen, hoch 2 Span vnnd 1 Span 6 Finger braidt. Original von Johann von Eyckh, welcher die Öhlfarb erst gefundten (‘A portrait in oil on wood of the Cardinal Santa Croce. In a frame half green and half gilded, 2 palms high and 1 palm 6 fingers wide. Original by Johann von Eyckh who first invented oil paint’).24 From 1426 to 1443 and thus in Jan van Eyck’s time, the cardinal priest of Santa Croce in Gerusalemme in Rome was Niccolò Albergati (c.13751443), General of the Carthusian Order since 1407 and Bishop of Bologna since 1417, who was of crucial importance to the peace treaty between Burgundy and France, concluded at Arras in 1435.25 On the strength of the inventory of the Archduke’s collection, the sitter in the Vienna portrait has almost always been called Albergati in the literature, beginning with Max Rooses and W.H. James Weale in 1902 and 1904 respectively.26 A number of essays that tried to prove, primarily on the basis of the costume, that the person portrayed could not be the famous cleric have not really dented this identification.27 A major cause of its persistence in the last three decades is those very notes that Stevens added to the chapter on Van Eyck in his copy of the Schilderboeck, discovered and subsequently published by Jan Briels in 1980.28 Stevens,

315

who had owned the painting in question before 1648, also called it a likeness of the Cardinal Santa Croce, painted by Jan van Eyck. It is remarkable, especially from a methodological point of view, that scholars have readily accepted this as a confirmation of the 1659 inventory. Almost all assume that Stevens transcribed a text on the original frame, also known to the writer of the inventory, and that the many details in his narrative confirm the reliability of this assumption.29 However, Stevens’s remarks do not confirm the inventory of Archduke Leopold because the two sources are not independent of each other: Stevens himself had sold the painting to Leopold, and he had certainly sold it as a portrait of the famous cardinal by the famous painter. Accordingly, the entry in the Archduke’s inventory is most probably based on Stevens’s information, the more so as both indicate only the Cardinal’s title, but not the name of the sitter. In fact, Stevens’s remarks rather contradict the identification with Albergati, as we shall shortly see. They read as follows: Noch bij Peeter Stevens een fraey conterfaysel van Jan van Eyck met dato 1438, wesende den Cardinael Santa Croce, die alsdoen tot Brugge was gesonden vanden Paus om de peys te maecken met Hertoch Philips over syn vaders doot met den dolphyn van Franckryck. Ditto stuck is nu in handen vanden Ertshertoch Leopoldus, die het nu gecocht hadde v. April 1648 (‘Also at Peeter Stevens’s, a fine portrait by Jan van Eyck, dated 1438, being the Cardinal Santa Croce who at that time was sent to Bruges by the Pope in order to make peace between Duke Philip and the French Dauphin in the matter of Philip’s father’s death. This work is now in the possession of Archduke Leopold who bought it on 5 April 1648’).30 The general assumption that this narrative derived from the inscription on an original frame fails to holds water when we check it against historical evidence. First, it does not seem likely that the frame which housed the panel around 1650 was the original one. In a painting of a fictitious picture gallery we can see the portrait in question in the foreground, leaning against a chair; it was probably

316

stephan kemperdick

Fig. 21.3 Jan van Eyck, Portrait of a Man, 1438 (?), oil on oak, 34.1 x 27.3 cm Vienna, Kunsthistorisches Museum (GG 975)

early texts on some portraits by jan van eyck

Fig. 21.4 Reverse of fig. 21.3

317

318

stephan kemperdick

added around 1650 by David Teniers the Younger to the already existing composition.31 Here, the frame depicted is obviously not a fifteenth century specimen (fig. 21.5). The portrait itself is rendered with some precision, and accordingly the frame does not suggest itself to be an invention – as its colour is of an olive tint it might well be the frame described in the 1659 Austrian inventory as gilt and green. But whatever Teniers might have rendered here, the halb grün vnndt halb zier verguldt Ramen mentioned in 1659 could hardly have been the original one, when we take the red marbling on the reverse of the panel into consideration. This marbling would have been combined with a frame imitating stone – as in the aforementioned Portrait of Margaret van Eyck of 1439 – and in Van Eyck’s surviving oeuvre, frames polychromed with a combination of stone imitation and gilding do not exist; indeed the idea of a simulated stone object does not fit too well with gilded mouldings. Instead, the original frame of the Vienna portrait probably looked similar to that of the Margaret van Eyck that in all likelihood dates approximately from the same period. If we now turn to the information Stevens provides about the Vienna portrait, we can say for certain that it is highly unusual for inscriptions on frames to tell such a historical story. What one would expect to find on a frame from Van Eyck’s time are the indispensable Christian name and the titles of the sitter – in this case Niccolò, his status as bishop and Carthusian – but they are not mentioned. Instead, and this is of paramount importance, the information given by Stevens is indeed full of errors: the peace congress took place at Arras, not at Bruges; it was in 1435, not 1438; and the French prince was not dauphin but king, crowned in 1429 as Charles VII. Thus, these partially correct details seem to be based rather on Stevens’s somewhat imprecise historical knowledge than on an original inscription. Only the date 1438 seems reliable because Stevens used the term dato in several other instances where there is indeed a date inscribed on the painting or its frame – as in

Fig. 21.5 David Teniers the Younger, Interior of a Picture Gallery, detail, c. 1650, oil on panel, 58.8 x 79 cm London, Courtauld Institute of Art Gallery (P.1978.PG 137)

the case of Jan van Eyck’s Man with a Red Turban: dato 1433. Consequently, this date must have been fixed (somewhere) to the painting now in Vienna, either on the frame, on the painted surface, or maybe on a note on the back of the panel. As Stevens had no doubts about the authorship of the portrait, Van Eyck’s name seems also to have been connected with the work, be it by an inscription or an oral tradition. The date 1438, the most reliable information provided by Stevens, would place the painting at a time when Albergati was not in the Netherlands; he came there only in 1431, when he visited Bruges among other places, and in 1435, to attend the peace conference in Arras. Accordingly scholars have been forced and were only too ready to separate the well-known preparatory drawing in Dresden, certainly made in front of the sitter,32 from the painting by three or even seven years. Obviously, scholars have seen few problems with such a separation. In fact, it is not very probable. If

early texts on some portraits by jan van eyck

one makes a detailed drawing with colour notes as a preparation for a painted portrait, there is no reason one should wait for years to execute it on a panel. Moreover, the date on a portrait not only records the creation of the work but also documents the likeness and status of the person portrayed in that year. The frequent indications of the age of the sitter exemplify this, and one may take as an example Van Eyck’s Portrait of Jan de Leeuw, also in Vienna,33 which is dated 1436; in combination with the date of the sitter’s birth, St Ursula’s Day 1401, which is also inscribed, one can easily calculate that Jan de Leeuw is shown at the age of thirtyfive. Thus, the date 1438 is not an argument pro but rather contra an identification of the man on the Vienna panel as Albergati. Much more decisive, however, are the observations on his dress and attire. They have already been put forward in the literature, but they still can be expanded upon somewhat. 34 His dress is indeed not that of a cardinal; it is more or less identical – up to the two buttons at the neck – to the gown worn by Jean Chevrot in the 1447 miniature from the Chroniques de Hainault.35 Chevrot was a bishop, not a cardinal, but likewise it is not a regular bishop’s robe. Rather, it seems to qualify him as counsellor to the Duke of Burgundy or as a scholar or both. It appears to have been a kind of garment that was worn by scholars, for example university professors; it might be the so-called cappa clausa, often mentioned in academic dress regulations.36 A scholar with this gown, in purple instead of red, can be seen in a Cologne painting from around 1490 as a bystander to the right of St Bruno receiving the Carthusian habit (fig. 21.6)37 – and this man is definitely not a Carthusian but a layman. Niccolò Albergati was recognized by his contemporaries as a near saint, very strictly adhering to the ascetic rules of his order and even refusing to wear a bishop’s attire.38 With respect to the Eyckian portrait, it has been suggested that he had to wear the red gown for representational reasons at the Arras conference, and that he would have

319

worn his Carthusian habit under it, in the way his corpse was found when his tomb was opened in 1633: he was clothed with pontifical vestments over the dress of his order.39 However, this red garment is not a pontifical vestment, as we have seen, and it would obviously have been impossible to wear the hooded habit of a Carthusian under it. The most important point, however, is the hairstyle of the man portrayed. It has already been pointed out by Edwin Hall that as a cleric he should have a tonsure.40 Regardless, this remark has either been completely ignored or it has been answered by the pro-Albergati camp with two objections: first, that there might be a small tonsure, not visible from the front, and second, that Petrus Christus’s so-called Portrait of a Carthusian of 1446, now in New York (fig. 21.7), also shows no visible tonsure41 – a large tonsure thus not having been a requirement for a Carthusian. Nevertheless, this argument turns out to be one against the traditional identification of the Vienna likeness. For the man on Petrus Christus’s panel is indeed not a Carthusian monk but a conversus or lay brother, who does not have the higher ordination and who is not a priest. In contemporary paintings these people are clearly distinguishable from the monks, who were priests: we see one lay brother with a beard and no tonsure in the left foreground of the aforementioned Cologne scene of St Bruno the Carthusian (fig. 21.6). Even clearer is another Cologne painting of around 1480,42 showing the Virgin with Carthusian saints and members of the Charterhouse (fig. 21.8): according to hierarchy, the monks, three on each side, are closer to the Virgin, whereas four lay brothers are in the second row. Like the sitter in Petrus Christus’s portrait, the latter figures have no tonsure, but a beard with clean-shaven upper lip. Their scapula lacks the bands that mark the garment of the Carthusian monks, and according to the statutes of the order, the lay brothers did indeed not wear them.43 What distinguishes each and every Carthusian is the extreme tonsure, the so-called corona that can be seen in all the representations of true

320

stephan kemperdick

Fig. 21.6 Cologne Master, St Bruno receives the Carthusian Habit from St Hughes, detail, c.1490, oil on canvas, 160 x 268 cm Paris, Musée du Louvre (MNR 972)

Carthusians, for example in the previously cited Cologne paintings (figs 21.6, 21.8) or in Petrus Christus’s Exeter Madonna, in which the Carthusian prior Jan de Vos kneels before the Virgin.44 Beyond a doubt the lack of the tonsure excludes the possibility that the individual in the Vienna portrait is a Carthusian, let alone the very strict general of the order; consequently, he is not Albergati. Maybe Peeter Stevens came to this mistaken identification because of the red garments, which he, a man of the seventeenth century, took at face value as a cardinal’s dress. Maybe he combined this assumption with the authorship of Jan van Eyck

and the date 1438, which he erroneously took as the date of the Franco-Burgundian peace treaty, and concluded that the sitter must be the famous Cardinal of Santa Croce. But whatever Stevens’s considerations might have been, in all probability, there was no longer an original frame fixed to the portrait around 1650, and the information Stevens provides is so inaccurate that an original inscription can be excluded as his source. The dress and hairstyle of the sitter are in no way admissible for a Carthusian, especially not for a man who was known for his ascetic attitude and his strict adherence to the order’s

early texts on some portraits by jan van eyck

Fig. 21.7 Petrus Christus, Portrait of a Carthusian Lay Brother, 1446, oil on oak, 29.2 x 21.6 cm New York, The Metropolitan Museum of Art, The Jules Bache Collection 1949 (inv. no. 49.7.19)

321

322

stephan kemperdick

Fig. 21.8 Cologne master, The Virgin Mary with Carthusians, c.1480, oil on wood, transferred to canvas, 200 x 170 cm Cologne, Wallraf-Richartz-Museum (WRM 153)

early texts on some portraits by jan van eyck

rules. Thus we have to abandon the name Niccolò Albergati for the Vienna portrait and all the suggestions it has brought along in scholarship. It does not make sense to prefer a wrong name to anonymity. On the contrary, to retain the identification with Albergati for the Vienna portrait is an obstacle to understanding. It not only leads to the unlikely chronological separation of the preliminary drawing and the execution of the painting but also blocks the research into the possible function and meaning of the portrait, which might represent a scholar, a counsellor or other learned person. The three panels discussed here mark different stages in Jan van Eyck’s career. The likeness of Isabella of Portugal is his earliest known portrait and at the same time his only work which was no doubt made on the order of his patron, Philip the Good, Duke of Burgundy. Its general type as a halflength with hands was preferred by the painter and occurs up to his latest extant portrait, Margaret van Eyck, of 1439. The 1433 Man with a Red Turban is exceptional in its quality – outstanding even by the standards of Van Eyck – but also by the fact that its frame is completely gilded, a feature that still lacks any explanation. It is also remarkable that the panel, almost identical in size to the Jan de Leeuw of 1436, shows the sitter at a smaller scale than the latter, as if further removed in space, and without hands. This last feature reoccurs only in the portrait of an unknown old man in Vienna. Probably made in 1438, it is among the latest surviving works of the painter. One may doubt whether there is a development recognizable in a little more than ten years of portrait painting by Jan van Eyck. However, the different types of portrait he employed are still a problem that has hardly been tackled. For example, could the age-old type of a shoulderlength or bust without hands have been intended to conform to an already existing gallery of portraits? Or was it merely the choice of the sitter, or even the painter, without further relevance? Jan van Eyck’s portraits have not ceased to pose a multitude of questions.

323

N OTES * Among the friends and colleagues who have helped, I would especially like to thank Valentine Henderiks who made it possible to give the talk in Brussels, and Douglas Kline, Berlin, who had a look at the English text. 1 Weale 1908, pp. lv-lxxii. 2 Weale 1908, p. lix. 3 Weale 1908, p. 15. Only Herzner 1995, p. 119, n. 21, has already realized that the report speaks of just one painting. 4 Weale 1908, p. lix. 5 The most thorough analysis so far is by Bauch 1961-1962, pp. 102-109. I am grateful to Bart Fransen for the information regarding the present whereabouts of the sheet. 6 A frame painted with acanthus scrolls in the manner of a manuscript also shows, for example, a small triptych of the Pietà, probably painted in Tournai around 1425 (Cologne, Wallraf-RichartzMuseum, see Rotterdam 2012, exh. cat., no. 43). 7 Book of Hours, Bruges, around 1430-1440, private collection, see Cologne 1987, exh. cat., no. 53, fol. 141r. 8 Bauch 1961-1962, p. 108-109, Herzner 1995, pp. 118-122, figs 41, 42. 9 Campbell 1998, pp.218-223. 10 Le Glay 1839, vol. 2, p. 480. 11 Michelant 1871, pp. 85-86. 12 Campbell 1998, pp. 212-217. The ‘turban’ is in fact a chaperon. 13 Briels 1980, p. 211. 14 Campbell 1998, p. 214. 15 There are remnants of sealing wax on the back that once probably fixed a paper; Campbell 1998, p. 212. 16 Weale 1908, p. 68. 17 As suggested by Campbell 1998, p. 174. 18 Goethals 1849, entry ‘Berlaymont’, unpaginated. Jansen 1989, pp. 45-47, identifies the sitter of the London portrait explicitly with Jacques de Berlaimont. Campbell 1998, p. 217, n. 20, objects that he was not a Seigneur de Berlaimont – however, he was a Berlaimont, and the knowledge of the individual titles of the members of the family might have been somewhat imprecise in the seventeenth century. 19 Gemäldegalerie SMB, Dhanens 1980, pp. 329-332. 20 Jansen 1989, pp. 45-47, cites a monograph on Arundel, Howarth 1985, p. 232, n. 8, but the duc d’Arschot mentioned there as former owner of paintings is a different person and already dead by 1613. Probably it was Charles III de Croÿ, duc d’Aerschot (1560-1612), also a collector of art, who in 1580 had himself married Marie de Brimeu, widow of Lancelot de Berlaymont. At any rate, there were several Berlaimonts in the family tree of the Croÿs, Arenbergs and dukes of Aerschot. 21 The dress of the man has been judged ‘not nearly rich enough’ for ‘a leading nobleman’ by Campbell 1998, p. 214. However, this is hard to estimate in my opinion: His robe is black or very dark purple, both well imaginable, even without brocade pattern, for a nobleman; see for example the posthumous portrait of John the Fearless, c.1450, in a plain black robe, Rotterdam 2012, exh. cat., no. 64. The London man also has no jewellery, but the same is true for instance for the portrait of Wenzel of Brabant in the Thyssen Collection Madrid, c.1410, who is clad in undecorated blue cloth, ibid., no. 63. 22 A watercolour by Ferdinand Storffer, made in 1733, shows the portrait, with others attributed to Dürer, Holbein etc., integrated into the wall decorations; Vienna, Kunsthistorisches Museum Storffer Album III, fol. 16. 23 Bruges, Groeningemuseum, Scott 2000, p. 133, Verougstraete, Van Schoute 2000, pp. 108-110, pl. 46, Bruges 2010, no. 20 (Till-Holger Borchert).

324

stephan kemperdick

24 After Demus, Klauner, Schütz 1991, p. 170. 25 For Albergati see Weale 1904; Hall 1968, pp. 3-5; Hunter 1993, pp. 209-212. 26 Rooses 1902, pp. 4-5; Weale 1904, pp. 190-192. 27 Weiss 1955, Hall 1968, Hunter 1993. 28 Briels 1980, p. 211. Stevens’s Schilderboeck is today in the Biblioteca Hertziana in Rome. 29 Stevens’s notes were accepted as confirmation, for example by Briels 1980, pp. 180-182, Dhanens 1980, pp. 282-291, Demus, Klauner, Schütz 1981, pp. 169-173; Jansen 1988, pp. 132-140; Belting, Kruse 1994, pp. 148-149, Till-Holger Borchert in Bruges 2002, exh. cat., p. 235, no. 24; Thomas Ketelsen in Dresden 2005, exh. cat., pp. 62-67, no. 12, Châtelet 2011, pp. 278-279. 30 Briels 1980, p. 211. 31 London 2006, exh. cat., no. 1. 32 Dresden 2005, exh. cat., no. 11. 33 Demus, Klauner, Schütz 1981, pp. 173-175. 34 See n. 26. A decisive argument is also the fact that Carthusians were not allowed to wear fur, and the man in Vienna has a fur collar; Weiss 1955, p. 146.

35 As was already observed by Hall 1968, p. 30. For the manuscript, Brussels, KBR, MS 9242, fol. 1, Brussels/Paris 2011, pp. 174-177 (Lorne Campbell); the same dress is also shown in variants of the Chroniques miniature, see ibid., figs 31-34. 36 Hargreaves-Mawdsley 1963, pp. 23, 155, 180, Bringemeier 1974, fig. 20. 37 Paris, Musée du Louvre, Schäfke 1991, pp. 279-280 (Ulrike Mader). 38 See n. 24. 39 Hall 1968, p. 6. 40 Hall 1968, pp. 6-7. 41 Thomas Ketelsen in Dresden 2005, exh. cat., p. 66. On the painting see New York 1994, exh. cat., no. 5. 42 Cologne, Wallraf-Richartz-Museum, Zehnder 1990, pp. 165170. 43 Geschichte einiger Geistlichen Orden 1783, pp. 77-79; Biedenfeld 1837, p. 72. 44 Berlin, Gemäldegalerie SMB; see New York 1994, exh. cat., no. 7.

Figs 22.1a-b (a) Old Roman fresco, detail, with empirical perspective effects, Metropolitan Museum of Art; (b) Detail of a room in perspective, Perspective, by Vredeman de Vriese, Leiden, 1604

22

Jan van Eyck and his ‘Stereoscopic’ Approach to Painting Renzo Leonardi

ABSTRACT: The apparent realism and illusion of depth, characteristic of Jan van Eyck’s paintings, have been extensively discussed in the literature. Perspective is a major tool that produces a three-dimensional illusion in a two-dimensional plane. In his paintings, Jan van Eyck intensively exploited his empirical perception of perspective, but not the linear geometrical ‘construction’. First, I will briefly discuss the ‘perceptual difficulties’ that have to be overcome in order to switch from an empirical use of perspective to a geometrically rigorous one. I will then focus on an interesting technique deployed by Van Eyck to enhance the illusion of depth, by flanking his use of empirical perspective and his extremely refined painterly techniques (aerial perspective, multiple layering and glazing, light handling, and so forth), through what I would call a ‘stereoscopic’ approach to his subjects, by combining views taken from different points and angles. From the Ghent Altarpiece to the Arnolfini Portrait and its mirror, a number of illuminating examples will be identified and discussed.

—o— From an Empirically-perceived Perspective toward the ‘One-point Construction’ (or put another way, from Jan van Eyck to Brunelleschi) In rendering the spatial depth of his paintings through his exceptional (empirical) sense of perspective, Jan van Eyck did not use the rigorous prescription of geometric perspective construction. I do not think that this was a deliberate choice. The abstract conceptual elaborations, leading to the ‘central one-point construction’ formalized in Italy in the first quarter of the fifteenth century,

were probably not yet adopted within the frame of the Nordic countries. Instead, Jan van Eyck implemented his empirical approach to perspective by deploying a wide number of expedients to reach, nonetheless, extraordinary three-dimensional illusions of depth. I will address some of these ploys in the next sections. In the following, I briefly comment on the ‘perceptual difficulties’ that must be overcome in order to switch from an empirical approach to perspective to a rigorously geometric one. The question of whether or not Jan van Eyck had his own ‘perspective geometrical scheme’ has been discussed at length. I refer to the proper broad literature (see for example Veltman 1977 and references therein) for these discussions and conclusions (if any). Here, I want to restate the problem, introducing another element that has never been explicitly taken into consideration while discussing Van Eyck’s perspective schemes, namely the fact that many of his paintings have probably been carried out while looking at (or imagining) the ‘scene’ to be painted as if seen and lighted from different viewpoints. ‘Free’ from the ‘cage’ of a well-defined ‘perspectival scheme’ (which implies typical ‘monocular’ view frames), he faced the problem of the illusion of depth and its inescapable complexity simply by counting on his perception and his ability to ‘eyeball’ what he saw around him. We will see in the next sections how, with this new element in mind, Van Eyck’s perspective ‘constructions’ are

328

renzo leonardi

easily understandable, and represent the best alternative that was achieved, when compared with the mathematical elaborations of the Italians. Let us complete this section while casually bearing in mind the main problematic aspects of perspective and visual perception of depth.1 One of the key visual illusions that led artists to introduce some empirical rules to convey the appearance of depth in their paintings is the ‘receding effect’. This is where objects of the same size in 3D, at different distances from the viewer, appear scaled down in function (foreshortening). A variant of this illusion is the apparent convergence of parallel lines going away from the viewer (parallel to the direction of view or slanting). The empirical rules of perspective must have been gradually developed in close connection with the evolutionary role of painting and its cultural framework, which represented the result of experience that had been accumulated by artists over time. As early as Roman times, it had become clear that in order to reinforce the sense of depth on painted surfaces it was helpful to draw (what in reality were) converging parallel lines (fig. 22.1a). This anticipated what would be later formalized in the fourteenth century. A strong empirical sense of perspective was not enough to fully conquer a rigorous geometrical method for projecting portions of a threedimensional space onto two dimensions (and vice versa!). In our opinion, this has to do with the complex mechanism of vision, as well as the advantages and disadvantages of binocular versus monocular vision. The latter is when we visually explore the space around us, casting our eyes in various directions. Our perception of depth is based on a mixture of signals elaborated by our brain, partly associated with binocular vision (stereopsis), partly elaborated by illusions proper to monocular vision (receding effect, chiaro-scuri, focus versus out-offocus imagery, and so on). The separation between these two parts represented a fundamental step toward the development of a mathematical approach to perspective. The breakthrough that

opened the way to this development was the famous fundamental Brunelleschi experiment and its consequent conceptual elaboration. The success of the experiment was based on carrying out a comparison between a painting, its mirrored image and a view, rigorously within a monocular vision system (one eye looking through a hole). The experiment led very rapidly to the achievement of the so-called ‘one-central-point construction’ and its generalization to multiple point constructions. It has to be kept in mind that the convergence to a single point of the set of lines orthogonal to the picture plane represents only one part of the full story, the other part being that the distance between lines that are parallel to the picture plane must be scaled in a way that is closely coherent with the convergence of the orthogonal lines. To better illustrate the previous remarks, we briefly analyse the simple and symmetrical case of a regular ‘room’ as viewed from a central position inside it and with the main direction of view being orthogonal to the front wall. Let the floor and walls be scanned by square grids, with their lines respectively orthogonal and parallel to the central view direction (fig. 22.1b). When taken from the viewing position (with its ‘monocular’ vision), a shot from a modern camera would perfectly simulate and match the one-point perspectival construction, as Brunelleschi had indirectly demonstrated with his experiment. However, our visual apparatus is more complex than that of a camera. Due to our binocular vision, the way we visually perceive the room is rather different. The geometrical conditions in which our eyes explore the floor and ceiling are different from those in which our eyes detect the lateral vertical walls. Most parts of the ceiling and floor remain in the internal part of our visual cone, where the acuity is maximal, whereas the lateral vertical walls remain mostly in the peripheral part of the cone. The pupil line, in our visual exploration, even if slightly tilted around, remains essentially parallel to the ceiling and floor plane, whereas this line remains essentially orthogonal to the lateral walls. Given our binocular vision, the ‘view’ of

jan van eyck and his ‘stereoscopic’ approach to painting

the floor and ceiling results from the superimposition (in the retina) of two monocular images with specular symmetry, whereas the view of each lateral vertical wall results from the superimposition of two asymmetric monocular images, since one eye is nearer to the lateral wall than the other. Finally, the lateral walls are not viewed from a central position (unless the distance between floor and ceiling is exactly double the distance of our eyes from the floor). It must be added that in ‘real’ rooms (observed from a central-like position), whereas floors and ceilings maintain in large part a geometrical regularity and symmetry through tiles, floorboards, rafters and similar objects, the vertical walls of the room are generally interrupted by windows, curtains, furniture. The ‘converging lines’ related to the objects ranged along the walls, are irregularly distributed at different heights with different extensions and in slightly different vertical planes. In interpreting these lines, our view is more exposed to illusions such as the Poggendorff effect. The symmetries and regularities of floors and ceilings have offered the most skilled painters the best visual conditions for a semi-quantitative perception of convergence and recession of the parallel lines associated with tiles, floorboards, rafters, and so on. They also played a fundamental role in suggesting the introduction of vanishing points or at least small vanishing areas in the practice of reproducing them in two dimensions (fig. 22.1b). On the other hand, the random distribution of lengths and positions of the parallel and orthogonal lines associated with the lateral vertical walls, together with the peculiarity of binocular vision, almost completely masked the possibility of ‘discovering’, through eyeballing, the strong relationship and correlation between the convergence and recession of the lines associated with the lateral vertical walls and those of the roof and floor. Many subjects painted by Jan van Eyck are set in spaces amenable to a ‘room’ as viewed from an inside central position, with a floor paved with tiles or floorboards, a roof spanned by rafters or other

329

similar structures, both aligned with the view direction, and orthogonal to the picture plane. For these paintings, roof and floor are generally drawn with a more or less narrow central convergent point, often slightly distinct, localized at the level where one expects the painter’s horizon line to be, whereas the convergence points of the lines of the lateral walls are almost randomly scattered in a wider area, disregarding correlation with the convergent points of the roof and floor.2 This is exactly what one expects in view of the considerations made while discussing the example of the ‘room’ viewed from a central position. Most of the considerations on Jan Van Eyck’s perspective technique have been carried out analysing the converging effects on the roofs and the floor’s ‘lines’ of the Arnolfini Portrait. In the next section, we shall see that the floor of the gallery of the closed Ghent Altarpiece offers a better situation to explore these techniques because a) on that floor, we can simultaneously analyse both the converging and the receding effects, b) the gallery probably represents a completely imaginary construction and the floor has been drawn by resorting to a pure abstract ‘method’ (possibly based on his visual experience and, perhaps, on his knowledge of some Italian precursor achievements). In the following, I simply refer to some perspective and geometric considerations and ignore any attempt to interpret the symbolism embedded in the paintings, the iconography, the theological constructions and so forth.3 Jan Van Eyck and his ‘Stereoscopic’ Approach to Painting (the Ghent Altarpiece)4 As already anticipated, Van Eyck obtained spatial depth and three-dimensional effects in his paintings not only through his strong empirical sense of perspective and his extremely refined painting techniques (aerial perspective, multiple layering and glazing, light handling, etc.), but also through what I would call a ‘stereoscopic’ approach to his subjects, combining views taken from different positions and angles. The ‘closed’ and the ‘open’

330

renzo leonardi

Fig. 22.2 Van Eyck probably traced the orthogonal floor joins on the four panels simultaneously, taking into account the space that would be occupied by the wooden frame. With minor fluctuations and deviations, he used a central-point construction. However, the lateral rafters’ convergence presents a large deviation from a rigorous respect of the one-point perspective. IRR reveals an underdrawn sequence of elegant trefoil arches.

versions of the Ghent Altarpiece offer an illuminating example of this approach. Closed Altarpiece First we examine the upper row, i.e. the sequence Angel Annunciate – Interior with City View – Interior with Lavabo – Virgin Annunciate. The four subjects are scanned by the physical wooden frames of the panels, which give the illusion of being ‘part’ of the gallery (they slightly shadow the floor of the gallery). As revealed by the infrared reflectograms, in Van Eyck’s original plan, the

upper part of each sequence should have been decorated by an elegant openwork similar to the one painted for the niches of the lower row (Joos Vijd – John the Baptist – John the Evangelist – Elisabeth Borluut) (fig. 22.2); more precisely, decorated with a double trefoil arch for the wider sequences (Angel Annunciate and Virgin Annunciate) and a simple one for the narrower (Interior with City View and Interior with Lavabo). The initial aim was probably to strongly mark the separation of the gallery from the external space. This plan was abandoned in the final painting,

jan van eyck and his ‘stereoscopic’ approach to painting

331

Figs 22.3a-b (a) Van Eyck respected, with surprising precision, the correlation between the receding effect of the parallel joins and the converging effect of the orthogonal joins (alignment of the tiles’ corners as required by the one-point construction: see Fig. 24.1b). The aim of the figure is to show the overall compatibility of the floor’s construction with the one-point construction; (b) Detail illustrating the ‘degree’ of alignment of the tiles’ corners

332

renzo leonardi

Figs 22.4a-b (a) The wall of the open arcade is as wide as the balcony; (b) the surface of the balcony over the city (left) is more ‘tilted’ than the one where the basin is positioned (right)

most likely to open the background of the gallery to a wider and deeper (illusionistic) space. So, in the final and visible version, the sequences are spatially unified in a unique open gallery, seen in perspective from an essentially central position. We do not know how Van Eyck prepared the under-drawing for the four panels. To get some indications on this point I have aligned the four reflectograms next to each other. By using the very accurate under-drawing, I traced: a) the orthogonal joins of the tiles, up to their converging area; b) some diagonals crossing the vertices of the tiles up to their converging point; c) the rasters of the ceiling (figs 22.2, 22.3). From this exercise, the following emerges: 1) Van Eyck probably traced the orthogonal joins on the four panels simultaneously, taking into account the spaces between them that would have been occupied by the wooden frame; 2) he used, with minor fluctuations and deviations, a central-point construction; 3) he respected, with surprising precision, the correlation between the receding effect of the parallel joints and the converging effect; 4) the lateral rasters’ convergence presents a large deviation from the rigorous one point perspective approach.5 In conclusion, the floor tiles have been drawn using some precursory rules of the rigorous one-point construction (figs 22.2, 22.3). 6

As soon as we focus on the sequences individually, however, it becomes evident that for each of them an ‘ad hoc’ approach was used by Van Eyck to achieve his own aims. Let us compare the city view with the lavabo. Firstly, let’s note how the wall of the open arcade is as wide as the balcony (fig. 22.4a). Furthermore, it’s noticeable how the portion of the balcony on the city is viewed from a viewpoint that is definitely nearer and/or higher than the one used for the lavabo. The surface of the balcony over the city is more ‘tilted’ than the flatter one, where the basin is positioned (fig. 22.4b). The effect obtained by this nearer and/or higher viewpoint is to render the view of the courtyards and city streets at ground level, animated with pedestrians and so appears more natural (due to a ‘downward slanting perspective’). Given this choice, Van Eyck must have faced the need to keep the perspective of the whole empirically correct, by maintaining the view from down upward for the superior part of the arches and the capitals, and from a more lateral viewpoint. This has been done by forcing the perspective. For example, in the superior part of the mullioned window, the arches remain strictly circular rather than elliptical (figs 22.5a-d). A final problem to be solved was how to arrange the base of the arcade’s columns that just stand in

jan van eyck and his ‘stereoscopic’ approach to painting

a

c

333

b

d

Figs 22.5a-d In spite of the diagonal view (from bottom right to top left), the arches remain strictly circular rather than elliptical

the middle of the balcony. Van Eyck rotated the bases by 45 degrees. Why did Van Eyck not paint the bases parallel and orthogonal to the border of the balcony? Let us simulate such an arrangement by virtually substituting the bases as painted by Van Eyck for bases with their faces respectively parallel and orthogonal to the balcony (the natural way, as for the capitals). This arrangement would have emphasized: 1. the difficulty to make the line passing through the left corner of the balcony, converging to the correct side, with the line defining the intersection of the bottom of the bases with the surface of the balcony on their visible side; 2. the bases become much narrower than the balcony (fig. 22.6a). The solution adopted to bypass this problem, and baffle our eye, was to rotate the bases by 45 degrees (fig. 22.6b). Would there have been an alternative solution? Yes, there would; that is by

lowering and withdrawing the viewpoint on the city (figs 22.7a-b). However, the effect of a ‘downward slanting perspective’ would have been lost (fig. 22.7b). The tilting of the bases was achieved at the cost of rather improbable architecture (fig. 22.8). The reflectograms help to reveal the efforts of Van Eyck in finding the appropriate solution to the problem (sketching first only one column, then two, and to rotate the bases) (fig. 22.7c). A similar solution using a tilted balcony and rotated bases to effect a downward slanting perspective (see reflectograms) was also adopted on the Virgin Annunciate panel (figs 22.9a-b). The Virgin Annunciate panel offers another interesting detail. A little (3-4 cm high) glass-flask, half filled with water, stands on a small horizontal surface under the balcony of the open window (fig. 22.10). The flask with its light reflections was

334

renzo leonardi

Figs 22.6a-b (a) Virtual substitution of the bases, as painted by Van Eyck, with bases whose edges are respectively parallel and orthogonal to the balcony borders. This arrangement would have introduced some inconsistencies (divergence of some perspectival lines and bases narrower than the balcony, in contradiction with Fig. 24.4a); (b) the solution adopted by Van Eyck was to rotate the bases by 45 degrees

probably modelled from a real one, seen from a very close position, and from a horizon plane that cut the bottle neck, as it results from the perspective arrangement of the circular base, the circular surface of the water and the circular duct of the neck. The flask is painted, in perspective, from a viewpoint that is totally incompatible with the one used for the balcony, (whereas the light hitting the water comes, harmonically, through the window). This suggests that sketches were transferred to the main painting without caring to harmonize the perspective of the whole. For the flask and the chandelier in the niche, on the right upper side of the Virgin, the perspective was handled in a more coherent way. The Archangel and the Virgin themselves are painted in a ‘monumental’ size; they are viewed as if they were rather close to the viewpoint and almost central to it.

Finally, let us analyse the lower row constituted by four niches hosting the donors and the saints. For the trompe-l’oeil ‘statues’ of the Baptist and the Evangelist, one can claim, all in all, that each one of the figures and their plinths are painted as essentially viewed from their own central viewpoint. The perspectival arrangements of the two supplementary niches behind them, however, are realized as viewed, in relation to the figures, from two opposite strong lateral points, helping to better protrude the ‘statues’ from their background (fig. 22.11a). In order to reinforce this illusion, Van Eyck separated the volume of the niches from the ‘external’ space by closing the superior part of the niches with elegant openwork in marble, in the shape of a trefoil arch. The refined use of light and shadow on the ‘statues’, plinths and niches, complete the illusion. The light sources seem to illuminate the saints

jan van eyck and his ‘stereoscopic’ approach to painting

335

a

b

c

Figs 22.7a-c (a) The contradictions shown in the previous figure could have been avoided by lowering and withdrawing the view point over the city; (b) however, the effect of a ‘downward slanting perspective’ would have been lost; (c) the artist’s effort to find the appropriate solution (first by sketching only one column, then two columns and finally rotating the bases) is shown by the IRR

336

renzo leonardi

a

b

c

Fig. 22.8 The tilting of the bases was made at cost of a rather improbable architecture, as illustrated by the cross-sections a, b, c. One sees that the tilted bases remain larger than the balcony.

from a slightly diagonal direction, from upward right to downward left. The position of the shadows caused by the ‘statues’, however, do not seem to be fully coherent with the position of the light sources (fig. 22.11a). Van Eyck chose the most effective solutions to reinforce the three-dimensional effects, even though he renounced the geometrical coherence. As we shall see in the next section, similar considerations, mutatis mutandis, can be expressed, for example, for the Thyssen and the Dresden Annunciations. I will now make some final remarks on the donors’ setting. They are standing on a floor paved with the same type of tiles as the gallery of the

upper row. Furthermore, their spacing converges perspectivally toward a central point area, as in the upper rows, creating an ideal link between the upper and the lower row (fig. 22.11b). The donors are settled on the floor more internally than the saints so that the ‘statues’ appear to spatially protrude more than the donors (figs 22.11a-b). Open Altarpiece We now turn our attention to the ‘open’ polyptych, starting from the higher row. Let us first focus our attention on the floor where the subjects are standing; differently from the upper row of the closed altarpiece, where the floor of the sequences is spa-

jan van eyck and his ‘stereoscopic’ approach to painting

337

Figs 22.9a-b (a) A similar solution (tilted balcony to see-down and rotation of the bases) was also adopted in the Virgin Annunciate panel; (b) IRR testifies to the efforts to find an appropriate solution

tially unified and viewed from the perspective which is essentially a central position, and with the appropriate convergence and recession of the tile joints. The floor of each group of sequences is viewed from rather different quotas and angles with the highest viewpoint for the three central figures. Given the

autonomous role of each group of subjects (Adam and Eve, Musician Angels and Singing Angels, the enthroned Virgin, Deity and Baptist), Van Eyck totally disconnected the space where they are represented by choosing a different horizon line and view angle for each group, and handling the

338

renzo leonardi

Fig. 22.10 Details showing that the flask is painted in perspective from a viewpoint totally incompatible with the one used for the balcony. This suggests that sketches were transferred to the main painting without harmonizing the perspective of the whole

Figs 22.11a-b (a) The saints are viewed by their own central viewpoint. The perspectival arrangements of the two supplementary niches behind them, however, are realized as viewed from two opposite strong lateral points, helping to better project the ‘statues’ from their background. To further reinforce the three-dimensional illusion, Van Eyck separated the volume of the niches from the ‘external’ space with an elegant trefoil arch. The position of the shadows is not fully consistent with the position of the light sources; (b) the donors kneel on a floor paved with tiles similar to those painted in the upper gallery

jan van eyck and his ‘stereoscopic’ approach to painting

floors accordingly (figs 22.12a-c). I will begin by discussing the most external panels, namely the full figure portraits of Adam and Eve (figs 22.12f, 22.12g). For Eve, the floor reduces to a line because the horizon line grazes the floor itself and the sole of her feet (figs 22.12g, 22.12e). For Adam, the floor is slightly lower, hidden by the wooden frame. This last choice allows Van Eyck to create a delightful trompe-l’oeil while playing with Adam’s feet. His left foot is lower and partly hidden by the wooden frame. His right foot is tilted upward, standing on the (hidden) heel bone and emerging over the frame with its metatarsus (figs 22.12d, 22.12f). I think that the purpose for slightly lowering the position of Adam in relation to Eve (her feet graze the frame; see (fig. 22.12e), is to account for the difference in size between male and female, while taking care not to express this disparity in terms of their height. The appropriate proportions between the bodies and their niches are thus maintained. The way Van Eyck handles this problem seems to me an extreme and unsurpassed refinement! The upper parts of the bodies are seen from a higher viewpoint. Now let us compare the Singing Angels and Musician Angels. The horizon line here is higher and more distant than the one for Adam and Eve. The floors of the two panels offer a certain unity of perspective, covered by identical tiles (preciously) decorated, with their joins receding and converging to a (more or less) remote central area. However, the two panels are seen diagonally, from two separate viewpoints (figs 22.12a, 22.12c). The Musician Angels panel is viewed from a more diagonal angle (perhaps to give the observer a better view of the keyboard and the hands of the player) than the Singing Angels panel (see the different bending of the joins of the floor toward the centre); the horizon line lays just over the head of the player; as a result, the upper group of the Musician Angels somewhat towers over the player. On the Singing Angels’ side, the horizon is a bit higher, essentially in between the two series of heads. As a consequence, despite the narrow-

339

ness of the panel, we perceive the group as more horizontally distributed. The central sequences (fig. 22.12b) are composed of arches of three extremely refined niches, which emerge from an arabesqued screen that completely hides the lower part of the niches. The three enthroned figures sit ahead of the screens, on a unified floor of monochromatic tiles and their joins definitely converge toward a central area. The heads of the figures protrude from the niches. The figures are seen frontally with the horizon line at the level of the head of the Deity Enthroned; the floor in front of the screens seems to be viewed from an ‘up-down’ direction, giving the illusion of a wider and deeper floor due to the strong convergence of the joins. The screens separate the (virtual) space inside the niches from the space where the figures sit. The effect of interposing the screen between the niches and the figures, of deepening the floor through a strong convergence of the tile joins, and of enlarging the figures themselves, help to create the illusion of an appropriate three-dimensional space from where the solemn figures protrude. This culminates with the aweinspiring majesty of the Deity Enthroned, who towers over the entire altarpiece. The viewpoint of the (virtual) observer continuously increases from the lateral sequences (Adam and Eve), up to the majesty of the central panels (fig. 22.12b). The richness and the variety of detail that embed these panels, presenting fine and detailed structures almost at any arbitrary small scale, recall the properties of some fractals! Turning now to the lower and most monumental series of panels, let us observe firstly that, unlike the others; the panels are not unified by a (centrally symmetric) architectonic frame, but rather by an immensely wide horizon under which an extremely rich botanical wundercamera is displayed. Essentially, the depth is reached through the foreshortening of the figures of the central panel and the atmosphere of the horizon,7 which includes far mountain peaks partially covered by snow (at

340

renzo leonardi

a

b

d

c

e

Figs 22.12a-g (a-c) Unlike the upper register of the closed altarpiece, here the floor of the sequences is viewed from rather different quotas and angles. Van Eyck totally disconnected the space of the sequences by choosing a different horizon line and view angle for each group. He handled the floors accordingly. E.g. for Eve, the floor’s plane is reduced to a line because the horizon line grazes the floor and Eve’s feet; (d, f) For Adam, the floor is slightly lower, hidden by the wooden frame. His left foot is lowered down, partly hidden by the wooden frame; his right foot is tilted up, standing on the hidden heel bone and emerging over the frame, leading to a delightful trompe-l’oeil; (a, c) The Musician Angels’ panel is viewed from a more diagonal angle than the panel of the Singing Angels (note the different bending of the joins of the floor towards the centre); (b) The figures are seen frontally within the horizon line at the level of the head of the Deity Enthroned. An arabesque screen separates the niches from the figures, which stand in front of the screen, on a unifying floor, which is viewed from an ‘up-down’ direction; the joins of the tiles converge towards a central area. The whole construction contributes to create the illusion of an appropriate three-dimensional space to settle the figures, culminating with the awe-inspiring majesty of the Deity Enthroned, who towers over the entire altarpiece f

g

jan van eyck and his ‘stereoscopic’ approach to painting

the horizon of the Knights of Christ panel). Let us focus on the most important architectonic elements of the panels, namely the central fountain, its column and the altar (fig. 22.13a). At first sight, it seems that the two items have their own proper empirical multi-perspective spaces. The converging point of the two lateral sides of the octagonal basin is very low, and lower than the altar itself; the fountain seems to have been seen from above; the empirical horizon line grazes (or it is just over) the pipes sprinkling water from the column. Over a little bank, the altar has the lateral edges converging well over its superior plan. The altar is ideally seen from a viewpoint much higher than the one used for the fountain, in such a way that one ‘sees’ the superior plan of the altar and the lamb is well rested on it (fig. 22.13a). Why did Van Eyck make this choice? He probably aimed to maintain the same strong emphasis on both the holy fountain and the lamb on the altar, and since he wanted to set them one over the other in the picture plane, he had no other choice. That said, assuming that the painter’s intention was to shape the fountain as a regular octagon, this was rather naively drawn, essentially without a guiding under-drawing. To illustrate this point, I have drawn a (regular) octagonal fountain, a column and an altar within a unique rigorous threepoints perspective and I have schematically superimposed them, as faithfully as possible, on the painted ones (figs 22.13a-c). The following is to be noted: a) the painted fountain presents large deviations from the ‘correct’ perspectival construction; b) one cannot match (i.e. using the same viewpoint) the wide horizontal plane of the painted altar and the plane of the painted fountain. If one matches the painted fountain simultaneously, the horizon line grazes the top of the altar. If one matches the painted altar, the fountain appears as viewed from a much higher viewpoint (figs 22.13a+b, 22.13a+c). Jan van Eyck and his ‘Stereoscopic’ Approach to Painting (Arnolfini Portrait) We will now briefly focus on the Arnolfini Portrait (fig. 22.14a).8 The literature dealing with the

341

perspective properties of this painting is impressive and leads to a variety of different conclusions9. There is, however, a general consensus that the converging ‘areas’ of the lines defining the space beyond the figures (the rafters of the ceiling, and the converging lines associated with some other lines orthogonal to the picture plan) lie essentially on a strip just over the mirror, whereas the converging areas of the floorboards lie at some distance under the mirror. In my opinion, these upper and lower converging areas reflect the empirical handling of perspective from two different positions and the two different horizon lines chosen by the painter (figs 22.14b-c, 22.15a). To my knowledge, this possibility has never been explored. Notice that the ‘eyes’ of the viewers (reflected in the mirror) are essentially positioned on the horizontal median line of the mirror (fig. 22.14b). This line identifies, by definition, the horizon line of the viewers. Corresponding to the lower (and nearer) viewpoint (of the painter), the other horizon line, must obviously be lower (fig. 22.14c). The interesting effect obtained by the painter through these choices is an enhancing of the depth of the backspace in relation to the forefront. To some extent, it is like seeing the background from a reversed binocular. The price to pay for this choice is a parallax effect (on top of the foreshortening) in the picture plane; the figures appear to be taller and wider, so that the mirror and the chandelier seem to be too low to be in line with the height of the sitters10 ; the chest seems to be partly hidden and the gown pushes against the bed (figs 22.14a-c, 22.15a-b). The virtual image in the mirror (the view from the mirror) offers an exciting counterproof of our hypothesis of a double viewpoint. In the following, I want to show that the virtual image as represented by Van Eyck in the mirror is perfectly consistent with this hypothesis. Once Van Eyck had painted the virtual image of the two viewers (probably one of the viewers is the standing painter himself) from his standing position, he completed the virtual image as viewed from the lower and nearer (with

342

renzo leonardi

a

a+b

b

a+c

c

Figs 22.13a-c (a) The fountain and the altar are painted as if viewed from two different viewpoints; (b) Reconstruction, within a rigorous 3-point perspective, of an octagonal fountain and an altar as viewed from the lower viewpoint, forcing the rear- and back-front of the painted fountain to match; (c) Same as above, seen from the higher viewpoint: the altar was matched to the painted one; (a)+(b) Illustration of the differences between painted fountain, altar (a) and the rigorous 3-point reconstruction (b); (a)+(c) Same as above, seen from the higher viewpoint

jan van eyck and his ‘stereoscopic’ approach to painting

respect to the sitters) point of view; i.e. he used the same viewpoint adopted to paint the figures in the main picture (fig. 22.15c). One has to observe that the largest portion of the room that we see in the full painting is essentially behind the figures (fig. 22.14a), whereas the largest portion of the room we see in the mirror is the one in front of the figures (figs 22.15c-d); these two parts partially overlap around the hanging chandelier, the window and the figures, which all appear in both the main painting and the virtual image of the mirror.

343

From the (virtual) viewpoint of the mirror, the picture ‘plane’ lies opposite to the mirror (on the same side of the two viewers) and the foreshortening affects the mirrored ‘scene’ in relation to the distance from the virtual viewpoint of the mirror (fig. 22.15c). Whatever is behind the figures is nearer to the viewpoint and whatever is in front of them is more distant. This is why, for example, the two viewers appear smaller than the figures, and this is obviously why the lines perpendicular to the ‘picture’ plane (such as the roof rafters) continue toward a converging area on the side of the viewers

Fig. 22.14 (a) The literature dealing with the perspective properties of this painting leads to a variety of different conclusions. There is, however, a general consensus that some converging areas of the lines ‘perpendicular’ to the picture plan lie essentially on a strip just over the mirror, whereas others lie at some distance under the mirror. The upper and lower converging areas probably reflect the empirical handling of perspective from two different view positions of the painter; (b) Van Eyck has painted the room behind the figures from a standing position (he can probably be identified as one of the viewers). He also painted the viewers in the mirror. The ‘eyes’ of the viewers appear essentially in the horizontal median of the mirror (the horizon line of the standing painter). Note the apparent position of the chandelier; (c) Van Eyck has painted the figures and the space in front of them from a lower and nearer viewpoint. The figures appear in the picture plane taller and wider both for parallax effects and ‘smaller’ foreshortening (nearer viewpoint!). Note the apparent position of the top of the hat of the figure.

344

renzo leonardi

c

a

b

d

Fig. 22.15 (a, b) The superposition of the previous two views (Fig. 22.14b, c) explains why, in the picture, the bottom line of the chandelier is lower than the top of the hat; (c, d) Van Eyck completed the mirrored image from a seated position. The grand-angular ‘visual’ cone of the mirror is slightly tilted upward. The apparent position of the chandelier is higher than the apparent position of the figures’ heads. The figures’ heads are higher than the viewers. The room in front of the figures is completely reflected in the mirror; (e) In the final stage of the painting the windows extend up to the roof (Fig. 22.14a). IRR (Fig. 22.15e) shows that in the initial stage of the painting, the upper part of the window and the shutters were drawn much lower, and a portion of the vertical wall was just standing over the window, consistent with the view in the mirror (Fig. 22.15d). e

(fig. 22.15d). By the way, the converging point (area) of the roof rafters in the mirror seems to be located lower than the viewers’ heads, as one expects, the painter’s viewpoint being lower than the viewers. This conclusion, however, might be weakened by the uncertainties in ‘extrapolating’

the rafter lines. Furthermore, since the viewpoint of the painter is lower than the one of the viewers, the heads of the viewers are seen by the painter (and painted) in the mirror at a higher level than those of the viewers (figs 22.15c-d). If Van Eyck had painted the figures from a standing position,

jan van eyck and his ‘stereoscopic’ approach to painting

the four heads (eyes) would have been essentially at the same (horizon) level (I have assumed here that the real height of the four persons is essentially the same and that there are no steps on the floor.11 A further consequence of painting the virtual image in the mirror from the lower viewpoint, is that the “visual cone” of the mirror is somewhat tilted upward so that the roof (essentially the part of the roof in front of the figures all the way up to the vertical wall that closes the room) is completely mirrored. By the way, this reveals an interestingly constructed roof detail, where the rafters are sustained by a transversal truss. The part of the roof on the back side of the figures, from the chandelier onward, does not reflect in the mirror (figs 22.15c-d). Moreover, whereas in the full picture the chandelier seems to graze the head of the figures (as an effect of foreshortening and parallax) (fig. 22.15b) in the virtual image, parallax and foreshortening effects work the other way round. In fact, positioning the chandelier nearer to the mirror than to the sitters, when projected on the virtual picture surface (seen through the mirror), it appears much higher than the heads of the figures and relatively wider than the main room chandelier (figs 22.15c-d). Finally, the figures appear confined in a narrower space in relation to the lateral (backward) sides (fig. 22.15d). At this stage, I will discuss what has been considered an irreducible contradiction between the windows, as seen by the painter, and as reflected in the mirror.12 I refer to the fact that above the windows, as reflected in the mirror, there is clearly a section of vertical wall, whereas it is not there in the window of the main picture (the upper part of window extends up to the roof, whereas in the mirror it is lower). This fact has been considered hardly compatible with the hypothesis that, what Jan Van Eyck has painted in the mirror, is the result of some sort of unmanipulated tableau vivant.13 In fact, we can come to an opposite conclusion if we couple (as one should always do) the indications derived from the reflectography of the painting14 with the ones that appear on the surface of the

345

painting (figs 22.14a, 22.15e). This comparison reveals an extremely interesting detail. In the initial stage of the painting, the upper part of the window was drawn much lower (and so were the shutters), and there was a section of the vertical wall just above the window. An interesting conclusion can be reached. The mirror ‘has seen’ (and Jan Van Eyck has registered) the window as it was probably really shaped in that room. In a first version, Jan Van Eyck had painted the window faithfully (fig. 22.15e). In a second step, probably in the studio, Jan Van Eyck, for his own reasons, raised the window up to the roof (fig. 22.14a). It would have been risky to artificially adjust, in the mirror, the consequences of such an artificial modification on the window and Jan Van Eyck preferred to remain, with the virtual image in the mirror, in the tableau vivant-like view. On the other hand, one has to observe that in the mirror there are no traces of the carpet behind the sitters. The situation and the position of the chandelier need a more detailed comment. Previously, we have assumed that the chandelier in the main picture has been painted from the higher viewpoint. I want to consider the possibility that it could have been painted from the lower viewpoint, as for the sitters. Also in this case, the chandelier is subject to a parallax effect and, in the picture plane, it results slightly higher and wider than if it were viewed from the higher viewpoint. However, because the chandelier is confined within a space behind the sitters, the parallax effect is more likely to affect the figures than the chandelier; the relative apparent quotas between the head of the figures and the chandelier still decreases and the apparent position of the chandelier would, in any case, result lower than the real one. A last remark may be made on the mirror’s curvature. In the recent past it has been assumed that for the purpose of computerassisted operations on the mirror, this was a portion of a sphere.15 If this hypothesis is plausible for the central part of the surface, it is untenable for the borders of the mirror. Only a very small radius of curvature at the borders would allow the reflection

346

renzo leonardi

of the red fabric at the bottom of the mirror (unless the mirror was almost a hemisphere; if so, we would run into other issues) (fig. 22.15d). It is highly probable that if Van Eyck painted this red fabric at the bottom of the mirror, he must have seen it in the mirror (this does not exclude the possibility that Van Eyck had arbitrarily corrected some lines at the borders of the mirror like the one defining the flank of the chest)!16 The qualitative effects and behaviours just described above can be easily reproduced by playing with a convex mirror: a room with the appropriate localization of some items (viewers, figures, etc.) and, finally, by choosing different viewpoints. A more quantitative analysis of all these effects would demand the precise knowledge of too many unknown parameters (distance between various items, heights, real shape and dimension of the mirror, etc.) and it would rule out any realistic way of quantitatively modelling all that. Other examples: the Thyssen and Dresden Annunciations Unlike the two saints of the Ghent Altarpiece, the ‘statues’ on the Thyssen and Dresden Annunciations, are tilted over their plinth, toward the centre, to somewhat face each other (figs 22.16a-b). The way in which the two Annunciations are handled deserves a dedicated comment. Both stand on similar plinths. Both are rendered in trompel’oeil. However, the first one is given the illusion of being inserted deeply in a pair of niches in the form of a parallelepiped (fig. 22.16b); the other appears almost well outside the picture plane (fig. 22.16a). How has Jan Van Eyck obtained these two entirely different and opposite effects starting from almost identical plinths, each viewed frontally, by a similar central viewpoint and illuminated by a similar lateral ideal light source? Let us examine the subtle expedients deployed to obtain these two opposite illusions. The two opposite illusions are already masterly prepared at the level of the frames and the plinths. In the Thyssen Annunciation, Jan Van Eyck plays with a painted

marble frame in order to connect the painted one with the real frame. The plinths of the ‘statues’ appear to stand on the (painted) bases supporting the marble frames (fig. 22.16c). These bases appear to be the upper surface of the physical wooden frame surrounding the pictures. To reinforce this effect, the (painted) plinths slightly invade the external carved frames (they ‘come out a little bit’, so to speak.17 Furthermore, there are no illusionistic niches behind the statues; on the contrary, the painter closes for his background a flat, black stone surface in which the statues are slightly reflected. To reinforce the illusion of flatness (not niches!), Jan Van Eyck mimics the reflected images of the statues by contouring the profiles of the two ‘statues’ and the dove with a whitish halo, shifted on their right side (figs 22.16a, 22.16c). It is obvious that, within a real three-dimensional situation, an imaginary observer can ‘see’ this halo on the right side only if he observes the statues from a strongly lateral viewpoint (viewing both the Virgin and the Angel from right to left; whereas the statues are viewed from their own central point. The final and powerful expedient put in place to reinforce the effect of an illusionistic protrusion of the statues from the picture plane toward the observer, is the appropriate use of light and shadow. The statues seem to be illuminated diagonally by a diffuse source of light from a (top) right to (bottom) left direction. Jan Van Eyck partly shadowed the left side of the painted marble frame; the shadow is caused by the imaginary interposition of the statues between the ideal light source on the right and the marble frame. Once more, the shadow is positioned on the marble frame in such a way as to maximize the three-dimensional effect, rather than in full coherence with the position of the (diffused) light source. In other words, the shadow, the statues and the light source are partly misaligned. In the Dresden Annunciation, the plinths are painted in order to appear deeply inside the frames lying on the bases of two niches (figs 22.16b, 22.16d). The two illusionistic deep niches are

jan van eyck and his ‘stereoscopic’ approach to painting

347

a

b

c

d

Fig. 22.16 (a, b) The two Annunciations are painted in trompe-l’oeil. The first one (a) appears almost well outside the picture plane, the other (b) is illusionistically inserted in a pair of parallelepiped niches. The two opposite illusions are masterfully prepared already at the level of the frames and the plinths. In the Thyssen Annunciation (a, c), Van Eyck inserted a painted marble frame to connect the painted frame with the real one and the plinths of the ‘statues’ appear to stand on the (painted) bases supporting the marble frames and slightly invade the external carved frames (c). The background is closed by a black stone (no niches!); a whitish halo mimics the reflected images. Both halos are viewed from a strong lateral view point, from right to left. In the Dresden Annunciation (b, d), the plinths appear deeply inside the frames lying on the bases of the two niches. The niches are perspectivally seen and suffused by a diffuse light entering diagonally from left to right, from two lateral viewpoints opposed to each other. Each (tilted) ‘statue’ and their plinths are essentially viewed frontally. There is no full coherence among the position of the light sources, the statues and the position of the shadows.

carefully painted, each one ‘seen’ in perspective from two lateral viewpoints opposed to each other. The Angel niche is ‘seen’ from right to left, and the Virgin’s niche is ‘seen’ from left to right (fig. 22.16b) (whereas the shift of the whitish halo contouring the Thyssen ‘statues’ demands the same lateral view for both, as I have just shown above). Remember that in both Annunciations,

each (tilted) statue and its plinth are essentially viewed frontally. In contrast to the Thyssen black background, the niches and the ‘statues’ are pervaded by a diffuse light entering diagonally from left to right. Here and there the statues and the dove cast their shadows on the niches. The shadows are located on the wall of the niches in such a way as to reinforce the illusion of two three-dimensional

348

renzo leonardi

‘statues’ positioned well inside the niches. Once more, there is no full coherence between the positioning of the light sources, the statues and the position of the shadows. The way the ‘statues’ are diagonally illuminated is pretty much the same; however, the shadow of the Virgin is essentially on the right lateral wall of the niche, whereas the shadow of the Angel and that of the dove above the Virgin are on the central wall of the niches. However, the illusion of depth is simply astonishing. In Conclusion I hope I have shown convincingly that Van Eyck, without resorting to a rigorous perspectival scheme, has been able to reinforce the illusion of depth, superimposing views from different angles and different distances in his paintings. In several cases the results are even more spectacular than those obtained by a strict use of rigorous rules of perspective. This technique, thanks to the extreme care with which Van Eyck handles every little detail in each square millimetre of his paintings, playing with colours, light and shadow and any sort of optical effect such as reflections, refractions and diffusions18 to render the three-dimensional illusion at any scale of observation, reproduces what today one would define as a hyper-realistic effect rather than the celebrated ‘realism’ of his paintings. Let us recall in summary that hyper-realism ‘is the simulation of something which never really existed’ (Jean Baudrillard).

N OTES 1 The reader should refer to the wide literature on these points. For perspective, some hints may be found in: http://www.handprint. com/HP/WCL/perspectX.htlm (2012) with X standing for numbers 1 to 5. On illusion and perception, see for example, Luckiesh 1922; Schiffman 2000. 2 Carelton 1982; Collier, Carelton 1983; Doehlemann 1905; Elkins 1991; Kern 1912; Ward 1975; Ward 1983. 3 On these and other aspects I refer, for example, to Harbison 1992, Harbison 1995; Pächt 1989; Van der Velden 2011; Hicks 2012, and to the vast literature quoted therein. 4 In this section, the name Van Eyck refers to both Hubert and Jan. In order to fully appreciate the present analysis, I suggest implementing some of our figures with the high quality imagery from the splendid documentation on http://closertovaneyck.kikirpa.be. 5 As noticed in the previous section, handling correctly the convergence of the lateral lines orthogonal to the picture plane demands an advanced knowledge of perspectival rules 6 The convergence of the alignment lines of the tiles’ corners is obviously affected by the small fluctuations in the alignment of the corners and by the uncertainties in extrapolating the lines up to the convergence area. All in all, however, the consistency with a central geometrical approach is amazing. 7 Aerial perspective; as a reminder, see, for example, Leonardi 2010. 8 To better follow the present analysis, it is suggested to implement some of our figures, for example the mirror, with larger and more detailed reproductions of them, easily available as public domain on the internet. 9 Carelton 1982; Collier, Carelton 1983; Criminisi, Kemp, Kang. 2004; Doehlemann 1905; Elkins 1991; Kern 1912; Veltman 1977; Ward 1975; Ward 1983. 10 Harbison 1995, p. 31. 11 In the mirror, there are traces of a couple of transverse floorboards between the viewers and the sitters, which are difficult to interpret. By the way, note that the sitters are barefoot. 12 Ward 1983, p. 684. 13 Ward 1983, p. 684. 14 Billinge, Campbell 1995. 15 Criminisi, Kemp, Kang 2004. 16 The mirror could have been made by squeezing a sphere of glass at the moment of its creation (glassblowing), through the appropriate manoeuvring by the master glassmaker. This procedure probably created the high curvature at the borders. The reflecting layer could have been injected at the moment of the glassblowing or later, through an appropriate opening on the back. However, other techniques could have led to a similar high curvature at the borders. 17 A similar remark can be found in Kemperdick, Lammertse 2012. 18 See for example: De Mey, Martens, Stroo 2012.

Fig. 23.1 Jan van Eyck, Dresden Triptych, 1437, oil on panel, 33.1 x 27.5 cm (centre panel), 33.1 x 13.6 cm (wings) Dresden, Gemäldegalerie Alte Meister, Staatliche Kunstsammlungen Dresden (gallery no. 799)

23

The Development Process of the Dresden Triptych. News and Questions Uta Neidhardt

ABSTRACT: In 1844, a large area of damage in the drapery of the Virgin’s robe in Jan van Eycks Dresden Triptych was restored, with the old overpainting in the Virgin’s red mantle being completely removed and the area being extensively reconstructed by the Dresden academic painter Eduard Bendemann (1811-1889). By using an image stacking application it became clear that for the Madonna of the Dresden Triptych Jan van Eyck obviously selected a figure type in which the form of the mantle featured several bowl-shaped folds between the Virgin’s knees or at least a slightly recessed area. In addition to the already known changes to the head of the Virgin and the right foot of the Christ Child, significant changes to the head of the Child were recently discovered. Moreover, the infrared reflectography image shows that in an initial version the Child’s head was turned to the right and his posture and gaze were directed towards his mother. In the case of the kneeling donor on the left wing-panel, recent investigation by IRR has revealed that this still unidentified man dressed in the fashion of the Burgundian court originally wore a dagger.

—o—

The Attribution The history of scholarship relating to the Dresden Triptych by Jan van Eyck (fig. 23.1) is nowhere near as wide-ranging and intensive as that concerning the Ghent Altarpiece, which is understandable in view of the differing status of the two altarpieces as regards their origins and function. However, a brief review of the investigations carried out on the only surviving triptych by Van Eyck demonstrates how

research conducted over several generations has brought about an astonishing accumulation of knowledge on this outstanding miniature work. Scholarly interest was at first primarily focused on fundamental questions such as attribution, dating and provenance, before issues like the structural composition of the image and the painting technique started to be investigated in parallel with the development of scientific methods of examination from the mid-twentieth century onwards. In a number of respects, the history of the investigation of the Dresden Triptych ran parallel with the investigation of other major works by Jan van Eyck. Seventeen years after the start of our own investigations on the Dresden Triptych, twelve years after the detailed presentation of the work at the Jan van Eyck Symposium in London1 and seven years after the exhibition in Dresden entitled Das Geheimnis des Jan van Eyck (The Secret of Jan van Eyck),2 it is now time to present new information about this altarpiece, which will help to fully solve the Dresden ‘Van Eyck puzzle’. A finding made in the course of a research project on the Dresden art commentator, collector and patron of the arts, Johann Gottlob von Quandt (1787-1859),3 has turned out to be of great significance for the history of the attribution of the Dresden Triptych. As a member of the Dresden Gallery Commission and a pioneering advocate of the rearrangement of the gallery’s works according

352

uta neidhardt

to art-historical schools, he was evidently the first person to have regarded Jan van Eyck as the author of the Dresden Triptych, it having previously been attributed to Albrecht Dürer. He had begun studying Old German and Early Netherlandish painting several years before moving to Dresden in 1824. In an article published as early as 1816 ‘On the current state of the Gemäldegallerie in Dresden’, he attributed the work to Jan van Eyck more or less in passing: Hence, the small altarpiece with the beautiful, knightly, guardian angel who has guided the pilgrim to the sacred destination of his pilgrimage, with Catherine on the other side and with the Blessed Virgin reverentially positioned on a throne in the middle of a magnificent cathedral and dressed in royal robes, […] – this painting, which was previously attributed to A. Dürer, but which is more probably a work of Joh. von Eyk, – would only develop its full impact if it were displayed in a small, separate room where everything from that period of pious love and zeal were exhibited together.4

gallery catalogues until the 1840s.5 However, it was not the reattribution but rather the evidently very poor state of preservation of this small altarpiece that first caused intensive attention to be focused on it in Dresden in 1837. Even in the first surviving condition report, drawn up in that year, there is mention of the problem that still today presents us with an almost unsolvable task – that of reconstructing the original Eyckian drapery in the Virgin’s robe on the centre panel. In the handwritten note concerning the altarpiece it is mentioned that: The red robe of the Virgin has, unfortunately, suffered much damage and […] has been repaired, but the overpainting is so crude and inexpert that the beautiful appearance of the opulent and pleasing drapery, which was probably below the retouched layer, is completely concealed.6

It is therefore clearly not the Berlin art scholar, Aloys Hirt, as was previously thought, but rather Johann Gottlob von Quandt – a figure who is very closely associated with the Dresden collection – who is to be credited with having been the first to attribute the altarpiece to Van Eyck. It is remarkable that Jan van Eyck was recognised as the creator of the Dresden Triptych at such an early date. This was the period when the Boisserée brothers were displaying their collection of Old German and Old Netherlandish paintings to prominent society figures. It was not until six years later that the first monographs about Jan and Hubert van Eyck were published by Gustav Friedrich Waagen and Johanna Schopenhauer.

The consolidation of the paint layer that was carried out in response to this in 18387 had, however, to be redone just six years later. In 1844 the large area of damage on the centre panel was restored, with the old overpaint in the Virgin’s red mantle being completely removed and the area being extensively reconstructed by Eduard Bendemann. This restoration work has frequently been cited and has been praised and criticised in equal measure. In view of the still moot question of the extent to which Van Eyck’s original underdrawing and painting were destroyed, it is worth taking a closer look at this restoration effort. Following a decision by the Dresden Gallery Commission, the academic painter Eduard Bendemann (1811-1889), who had been employed at the Dresden Art Academy since 1838, was commissioned with the complex task of reconstructing the Virgin’s robe. Writing from memory nearly 40 years later, Bendemann described his approach as follows:

Condition and Restoration This early, pioneering attribution of the Dresden Triptych did not appear in the official Dresden

[…] I did nothing but fill in the red mantle where it had completely peeled away from the ground so that in some places there was nothing

the development process of the dresden triptych

left of the paint or the lines of the folds. In order to do this, I drew a study […] after nature and adapted it where possible to the existing remains of the original painting. The main lines from the knees downwards are new. […] A similarly [damaged] area is on the side of the right arm, and so it appears as if the paint had peeled off there, too […] – I did not produce any of the bejewelled ornamentation – that is done in such a way that my skills would not have been anywhere near adequate […]. Bendemann concluded his comments with the selfcritical remark: When I saw the picture during my last visit to Dresden, I was immediately struck by the fact that – as had already been clear to me from the start – my folds were not sufficiently in the style of v. Eyck.8 We can only agree with this judgement; at the same time we are faced with the task of determining as precisely as possible where Jan van Eyck’s original painting has indeed been lost and the extent of Bendemann’s retouching within the Virgin’s mantle. The History of Research since 1945-1956 In the hope of casting light on these and other questions, various radio-diagnostic and technical investigations have been conducted on the Dresden Triptych since the middle of the last century. The first X-radiograph of the centre panel was made in the early 1950s, when the triptych was in the Soviet Union, having been seized as spoils of war along with all the other Dresden art treasures. In 1957 an international committee of experts, which was established after the return of the Dresden Gemäldegalerie in order to assess the main works of the collection from a conservation point of view, prompted the first infrared photographs of the altarpiece and examinations using ultraviolet light. For the first time, this provided information about the underdrawing. After the famous signature in the inner

353

moulding of the lower side of the centre frame, which was retouched at an unknown early date, was uncovered in 1959,9 there was increasing interest in the Dresden Triptych among international Van Eyck researchers. The first investigation of the altarpiece using an infrared image converter was conducted in Dresden by J.R.J. van Asperen de Boer as early as 1982 – three years after the ground-breaking publication of the infrared reflectograms of the Ghent Altarpiece. After that, a number of examinations using infrared reflectography were conducted in the 1990s by Molly Faries (1993) and Peter van den Brink (1996), among others, as well as – from 1995 – the staff of the Dresden Gemäldegalerie and its restoration workshop. Thanks to constantly improving technology, nearly every one of these investigations provided new findings concerning the material structure, the painting technique and the condition of the altarpiece, as well as Jan van Eyck’s original artistic intentions. In particular, the investigation of the centre panel of the Dresden Triptych by means of infrared reflectography revealed a complex structure which is still extremely difficult to construe, so that various, sometimes contradictory, interpretations have been made in the past. In her article published in 1999,10 which took account of our test results obtained in 1998, Molly Faries confirmed the difficulty of distinguishing Jan van Eyck’s underdrawing from that of the restorer and of differentiating the clearly destroyed sections of the Virgin’s mantle from intact areas. Our earlier investigations carried out on the triptych and published in 2000 and 200511 served primarily to determine the extent of destruction of the original painting in the mantle of the Dresden Madonna, which was previously largely unknown. At the same time, the attempt was made to interpret the appearance of the underdrawing visible on the infrared reflectogram as being the result of vestiges from Van Eyck’s original drawing having been overlaid by the reconstructive drawing of Bendemann, the restorer (fig. 23.2). It turned out, however, to be extremely difficult to differentiate between the drawings, which merged into one

354

uta neidhardt

Fig. 23.2 Jan van Eyck, Dresden Triptych, centre panel, IRR

the development process of the dresden triptych

another and were additionally obscured by the painted retouching on top of them. In an interesting article published in 2002 about the relationship between the work known as the ‘Leipzig Drawing’ and the centre panel of the Dresden Triptych,12 Georg Zeman interpreted the silverpoint drawing in the style of Van Eyck held in Leipzig as being a copy of Jan van Eyck’s lost preliminary drawing for the Dresden Virgin and Child group. He proceeded on the assumption that apart from a few areas on the right-hand edge, the hem of the mantle and below the Child’s right foot, no remains of the original drawing in Van Eyck’s own hand had been preserved in the surviving paint layer and in the concealed drawing of the Virgin’s mantle. Owing to the continued uncertainty regarding the extent of the reconstruction of the underdrawing and the retouching, we decided to conduct further investigations on the Dresden altarpiece in 2011. Furthermore, our earlier infrared images had already revealed very interesting indications of changes in the conception of the painting during the creative process, which are evident not at the level of the underdrawing but in the paint layer, and we intended to verify these. We also hoped that improved radio-diagnostic techniques might provide new findings regarding the question of the identity of the donor depicted on the left wing, which has puzzled researchers for decades. Bendemann’s Restoration Before venturing to reassess Eduard Bendemann’s retouching of the Dresden Triptych it is necessary to consider the extent of the damage and the methods employed during other known restoration procedures in the area of the Virgin’s mantle. This is only possible to a limited extent because few restoration documents have been preserved. In a note about the restoration procedure undertaken in 1844, the restorer Carl Martin Schirmer stated that: On the main picture, the overpainting on the red [mantle?] of the Madonna has been removed and almost completely repainted […].13

355

And when consolidation of the paint layer again became necessary in 1857, it was stated in a report by the Gallery Commission afterwards that: […] it appears as if the wooden panel, some parts of which were removed during Matthäi’s period of office [Gallery Inspector from 1823 to 1845], though against his will, has since begun to shift again and has shrunk a little.14 This second statement, in particular, is confirmed by the X-radiograph of the said area in the centre panel of the altarpiece (fig. 23.3). The large, welldefined light grey to white areas represent those parts of the picture in which the damage to the original painting evidently extended right down to the support. The lead-containing putty applied here may have been used to repair damaged (or ‘removed’) parts of the original ground. It is difficult to assess the other parts, which are visible on the X-radiograph as darker, similarly sharply contoured grey areas covering almost the entire area of the Virgin’s mantle except for the lower right seam and some of the bundles of drapery below the Christ Child. In the case of these areas, we somewhat cautiously surmise that the original ground has survived along with some vestiges of Jan van Eyck’s underdrawing. We are well aware that the existing X-radiograph represents a conglomeration of various superimposed versions of the painting and successive restoration efforts, all of which have gradually formed the surface of the picture over a long period. Thus, our initial presumption is that in the middle of the Virgin’s mantle below the Child, in the lower right-hand area of the mantle and in the upper torso of the Virgin there are areas in which at least some of the original ground and underdrawing have survived (fig. 23.4). Hence, the underdrawings visible on the infrared reflectogram of these areas of the mantle may be considered, to an indefinable extent, to be drawings in the hand of Jan van Eyck. However, the interpretation of these ‘original’ areas of drawing is difficult for two reasons. As the relatively uniform

356

uta neidhardt

Fig. 23.3 Jan van Eyck, Dresden Triptych, centre panel, X-radiograph

the development process of the dresden triptych

Fig. 23.4 Jan van Eyck, Dresden Triptych, centre panel, IRR showing losses, digital assembly

357

358

uta neidhardt

appearance of the underdrawing indicates, Bendemann, on the basis of a drapery study drawn from nature, which unfortunately is no longer extant, possibly reworked the remains of Van Eyck’s drawing in order to adapt it to the new ‘overall impression’ of his much softer drapery drawing. The appearance of what are presumably the remains of Van Eyck’s own drawing is also obscured by the fine black lines Bendemann used to indicate the depth of the folds in his retouching of the painting. Therefore, the appearance of the remains of the original underdrawing in the Virgin’s mantle gives very little indication of Van Eyck’s characteristic drawing style, with closely set bundles of lines and cross-hatching, such as is found, for example, in the underdrawing of the figure of St Catherine on the right wing-panel. The forms of the painted Virgin’s mantle have a generally uniform appearance, albeit completely un-Eyckian in style. The fabric falling gently from the Virgin’s knees and lying on the throne dais in rounded, tightly curved folds is much closer in style to the works of the history painter Eduard Bendemann than to those of Jan van Eyck. This has frequently been remarked upon ever since the retouching was completed in the mid-nineteenth century, and it is immediately obvious when the altarpiece is compared with an original painting by Bendemann – the oil sketch, held in Dresden, which he made for the 1833 painting Two Girls at the Well (fig. 23.5). In carrying out the red retouching, for which Bendemann used oil or resin paints, he endeavoured to produce a uniform effect on the surface by also varnishing over the surviving vestiges of original paint. Such areas include, for example, the area of red above the bejewelled hem of the mantle in front of the Virgin’s left leg. Reconstruction of the Original Drapery In view of the extensive damage to the Virgin’s mantle, the question remains as to its original appearance. If we follow Georg Zeman’s 2002 suggestion that the Leipzig drawing is a copy of Van Eyck’s preliminary drawing or underdrawing for the

mantle, the result is an image that assumes the total loss of Van Eyck’s underdrawing (fig. 23.6). The Madonna type in which the garment forms a box shape in the area between the Virgin’s legs is similar to that of the Frankfurt Lucca Madonna, which was painted at around the same time.15 In this instance we wanted to be absolutely precise and commissioned an image stacking application (fig. 23.7). Whilst there are similarities between the cascading folds descending from the Christ Child on the left-hand side of the figure and the remains of the original underdrawing in the upper part of the area beneath the infant’s right foot, the smooth box form in the middle and also the deep V-shaped folds on the right in the Leipzig drawing diverge significantly from the drawing visible in the infrared reflectogram of the Dresden Triptych, which is presumably of Eyckian conception. The visible folds in the undamaged areas, such as the folded white cloth hanging down from the Virgin’s lap and the tightly folded sections of fabric on the right-hand edge, indicate an entirely different form of mantle in these areas. It appears as if the folds descending from the Virgin’s left knee derive from Van Eyck’s original design, whereas neither the Leipzig drawing nor Bendemann’s retouching in the central region of her body and the right hand side of the mantle are anything like the original form. It is feasible that for the Madonna of the Dresden Triptych Jan van Eyck selected a figure type in which the form of the mantle featured several bowl-shaped folds between the Virgin’s knees or at least a slightly recessed area. Such features are found in other works by Jan van Eyck, such as the enthroned figures in the upper register of the Ghent Altarpiece and the early Eyckian Fountain of Life panel in Madrid.16 Even greater similarity to the Eyckian preliminary drawing of the Dresden Virgin’s mantle may possibly be found in the drawing entitled The Mystical Marriage of St Catherine held at the Germanisches Nationalmuseum in Nuremberg – a copy dating from the second half of the fifteenth century – which incorporates an Eyckian invention17 (fig. 23.8). The form of the drapery in

the development process of the dresden triptych

359

Fig. 23.5 Eduard Bendemann, Two Girls at the Well, 1833, oil on canvas on board, 25 x 34.5 cm Dresden, Gemäldegalerie Neue Meister, Staatliche Kunstsammlungen Dresden (gallery no. 2652)

the Virgin’s robe shows the irregular deep bowlshaped folds between the legs, which are undisputedly evident in the Dresden Madonna, as well as the small cascade of folds on the left-hand edge of her garment. In addition, it gives an idea of the form of the folds on the right-hand side of the Madonna: the cloak hanging down on the right covers her left arm more loosely than Bendemann would have thought when filling in this gap in the Dresden original and lower down it falls in peculiarly undirected smaller folds which extend right down to the pedestal. The small group of known drawings made in connection with the Van Eyck workshop or as

later copies also includes other examples of Madonna types whose originals have been lost, which may have been similar to the Dresden Madonna. In our opinion, this group of later copies after a type of Madonna that corresponded closely to the Dresden Madonna includes the Leipzig drawing, which probably dates from the beginning of the sixteenth century. A further variant of the ‘Dresden type’ has been preserved in the work known as the Simpson Carson Madonna, dating from around 1470-1480, in which the drapery folds of the Virgin’s mantle are possibly more similar to those of the Dresden Madonna than is the case with the Leipzig drawing.18

360

uta neidhardt

Fig. 23.6 Jan van Eyck, Dresden Triptych, IRR showing losses and Leipzig drawing, centre panel, digital assembly

the development process of the dresden triptych

361

Fig. 23.7 Jan van Eyck, Dresden Triptych, IRR with Leipzig drawing, detail of the centre panel, image stacking application

Changes in the Paint Layer (Centre Panel) The most interesting discoveries made using infrared reflectography, however, were not connected with the underdrawing on the Dresden Triptych but were found instead in the paint layer. Jan van Eyck clearly made a number of fundamental changes to the conception of the image during the painting process, for which there is no evidence in either the preliminary drawing, where such exists, or in

the underdrawing of this small altarpiece. Thanks to the use of a new Osiris infrared camera, we were able to make another spectacular find in 2011. In addition to the changes to the head of the Virgin and the right foot of the Christ Child, which have already been described elsewhere,19 significant changes to the head of the Child were discovered recently. It can be seen in the infrared reflectography image that in an initial version the head was

362

uta neidhardt

Fig. 23.8 Low Countries, after a model influenced by Jan van Eyck, Virgin and Child with Saints, second half of the 15th century, pen and brown ink, brush and black ink, wash and watercolour on paper, 20.1 x 29.4 cm, lower right: ‘J.C.R’ (collector’s monogram) Nuremberg, Germanisches Nationalmuseum, Graphische Sammlung

turned to the right and the Child’s posture and gaze were directed towards his mother (fig. 23.9). The head, which was in three-quarter profile and slightly elevated, was already at an advanced stage of execution and would have required the body to be depicted quite differently to that in the final version. However, an earlier version of the body of the Child is not evident in the painting of the Dresden Triptych that is visible today. Obviously, Jan van Eyck had initially planned a Madonna image in which the Christ Child was turned towards the breast and face of his mother. The posture of the Child must have been similar to that found on the Lucca Madonna (c.1437) or the Madonna in the Church held in Berlin (c.1438). The intensity and intimacy of the mother-child relationship, as expressed, for example, in the

breastfeeding image of the Lucca Madonna or in the caressing of the Child in the Antwerp Virgin by a Fountain (1439), does not seem to have been what was intended in the first version of the Dresden Madonna. Instead, the Child seems to be looking up at the mother in a rather distant way; one might even go so far as to say that the infrared reflectogram gives the impression of a serious facial expression. The original rightward tilt of the head of the Madonna in the Dresden Triptych was immediately reminiscent of the Virgin and Child with Canon Joris van der Paele (1436). However, in the case of the Dresden Madonna it is unclear whether the respective initial versions of the head of the Virgin and the head of the Child were produced at the same stage in the creation of the work. If that were the

the development process of the dresden triptych

Fig. 23.9 Jan van Eyck, Dresden Triptych, centre panel, detail, IRR

case, the image would have been one in which the Virgin’s gaze was directed slightly downwards and to the right, but with the Child looking up at her. Unlike in the case of the Van der Paele Madonna, these two different gaze directions would not have made any sense within the present form of the composition and the positioning of the figures on the triptych. It is possible that the altering of the positions and bearing of the two main protagonists in the holy scene took place on account of a specification or change in the conception of the images on the wing-panels. Such considerations always raise the question as to the identity of the donor depicted in the left wing-panel of the altarpiece. To return to the new finding, it can be stated that during the painting process Van Eyck decided to turn both the head of the Virgin and the Christ Child himself slightly towards the left. In an allusion to the altar, the Child on the Virgin’s lap is presented to an observer located outside the sacred zone. The group now assumes a demonstrative air, with the Virgin presenting Christ to the believer who, immersed in prayer, receives a vision of her in precisely this representative form.

363

Changes in the Paint Layer (Left Wing-panel) Another extremely interesting find was made on the left wing-panel of the altarpiece in the figure of the kneeling donor. As described earlier, Jan van Eyck made a number of alterations to the posture of the kneeling worshipper during the process of the picture’s creation, both in the underdrawing and in the paint layer.20 These relate to the position of his hands and feet. More recent investigation has revealed that this still unidentified man dressed in the fashion of the Burgundian court originally wore a dagger (fig. 23.10). It was positioned at the lower edge of the donor’s broad right sleeve and was held in place by a belt. Unlike the changes to the figures on the centre panel, Van Eyck had begun this interesting detail in the underdrawing and executed it in the initial painting stage. The preliminary drawing in the area of the dagger grip and the difference between this and the version executed in colour are clearly visible. The weapon was evidently a rondel dagger or dague à rouelles, a type of weapon that had originated in Italy and spread throughout Europe between 1350 and 1500. A fifteenth-century German rondel dagger with a boxwood grip from the Dresden Armoury (Rüstkammer) has a very similar form to that of the donor on the Van Eyck painting (fig. 23.11). The original dagger of the Dresden donor with its clearly recognisable flat pommel cap would have been about thirty to forty centimetres long, an important indication of the proportions of the kneeling figure represented in the picture. Extending downwards from the guard is a light-coloured strip, clearly visible on the infrared reflectography image and also already included in the underdrawing. This was undoubtedly the hanging leather strap of the belt on which the man wore his dagger, which was otherwise covered by the hanging sleeve of his houppelande. The motif of the kneeling donor with a belt is not unique among the works of Van Eyck; it is also found on the figures of Joos Vijd in the Ghent Altarpiece and the chancellor in the Virgin and Child with Chancellor Rolin. Both those figures, however, wear or wore a large purse on their belt. On the Merode

364

uta neidhardt

Fig. 23.10 Jan van Eyck, Dresden Triptych, left wing, detail, IRR

Altarpiece by Robert Campin and Rogier van der Weyden, which was begun in 1425 and was changed several times, the donor, Peter Engelbrecht, is portrayed wearing both his purse and a ballock dagger on his belt. The Question of the Donor’s Identity These interesting details, which have previously been hidden from the viewer under the final design of the donor’s green cloak, provide important new information in precisely the area where it is most urgently needed. This leads us directly to what has for decades been the most frequently discussed question as regards the Dresden Triptych: the identity of the person who commissioned it. Right up to very recently, various scholars have put forward an abundance of proposals, ranging from Michel de Lannoy and Michel de Ligne, via Michiel Adornes and members of the Rapondi-Burlamacchi and Giustiniani families, to Jan van Eyck himself. Through the reconstruction and publication of the indistinct original coats of arms on the inner sides of the wing-panel frames, we have now ascertained that the Giustiniani family was definitely involved in the creation of the triptych. Evidence that the altarpiece was at least temporarily in the possession

of this Genoese family is the existence of a retable from the second half of the fifteenth century in Pontremoli near Genoa, the centre panel of which can be seen as deriving from the centre panel of the Dresden Triptych. Since then, research conducted by Peter Heath21 and Noelle Streeton22 has resulted in further proposals being put forward. Heath suggested that the Dresden donor might have been a member of the Flemish judiciary, pointing in particular to iconographical references within the image and connections with the depiction of the Just Judges on the Ghent Altarpiece. Streeton, on the other hand, took up the strong argument in favour of the Giustiniani connection and used the ledgers of the Milanese Borromei Bank of 1438 to report on the financial activities of a certain Raffaello Giustiniani, who was commercially active in Bruges on behalf of the entire family, and possibly also on behalf of Michele Giustiniani who by that time had long returned to Genoa. There are still few points of certainty among the mix of sometimes contradictory proposals for solving the question of the identity of the Dresden donor. Since no archival solution has yet been produced on the basis of sources, we can for now only reflect on the altarpiece itself, a work that was commissioned from Jan van Eyck by a private individual in the Giustiniani family. We have recently established that the painter initially intended to portray the praying man in the picture with a belt and dagger at his side. The rondel dagger, which in the fourteenth century had been the preserve of aristocratic circles, soon became a prestigious item of apparel among secular and bourgeois wearers, and even members of the lower classes were not prohibited from carrying it. Unfortunately, therefore, the possession of such a dagger by a man wearing Burgundian style garments gives little clue as to his social status, his nationality or even his occupation. The decision of the painter (or the person represented?) to go without this accessory in the context of the triptych, is remarkable, however. In our attempt to identify the historical figure depicted here, this late alteration of a pictorial element that had been

the development process of the dresden triptych

365

Fig. 23.11 Dagger, 15th-16th century; handle: German, early 15th century, boxwood, carved; blade with marking punch, 34 cm (blade: 18 cm) Dresden, Rüstkammer, Staatliche Kunstsammlungen Dresden

intended from the outset points us in a direction that we suggested as early as 1998. Having originally been equipped with a prestigious defensive status symbol and bearing a rather restrained expression, the portrait of the donor was gradually altered by Van Eyck, presumably at the donor’s own request. The removal of the dagger and the opening of the hands raised in prayer give greater emphasis to the profound religious emotion of the individual. NOTES * This article has been written in close collaboration with my colleagues Christoph Schölzel, restorer in the paintings restoration workshop of the Staatliche Kunstsammlungen Dresden, and Konstanze Krüger, researcher at the Gemäldegalerie Alte Meister of the Staatliche Kunstsammlungen Dresden. Their comprehensive investigations concerning technical questions and the restoration history of the Dresden Triptych, along with numerous inspiring discussions, formed the basis of and were a necessary prerequisite for this study; I am greatly obliged to both of them. I am also very grateful to Mr Bertram Lorenz, student on the Art Technology and Conservation of Works of Art course at the Hochschule für Bildende Künste Dresden, for kindly providing me with a series of digital photomontages created in preparation for a practical

study concerning the artistic techniques used in the Dresden Triptych as part of the ‘Historical Painting Techniques’ module, which helped to clarify the extent of paint losses in the area of the Virgin’s robe. It was also exceptionally beneficial to be able to collaborate with Jörg Wittenberg, whose image stacking application developed in the sphere of cartography made possible the precise mapping of the painting, infrared reflectogram, X-radiograph and the ‘Leipzig Drawing’. I should like to take this opportunity to express my sincere thanks to him as well. 1 Neidhardt, Schölzel 2000. 2 Dresden 2005. 3 For this reference to Quandt’s early attribution, I am grateful to Andreas Rüfenacht, Berne, whose thesis entitled ‘Johann Gottlob von Quandt (1787-1859). Kunstschriftsteller, Sammler und Förderer der Künste in Dresden’ included discussion of Quandt’s assessment of works in the Dresden Gallery. 4 von Quandt 1816. 5 Matthäi 1846. 6 Matthäi 1835. 7 See n. 6, handwritten note dated 22 June 1837 in a copy in the library of the Gemäldegalerie Alte Meister (between pp. 26 and 27), evidently added later: ‘In 1838 a few areas were consolidated and the whole painting was revarnished.’ 8 Eduard Bendemann, letter to Hugo Bürkner dated 3 April 1888. Archive of the Staatliche Kunstsammlungen Dresden 01/GG 8, vol. 9 „Tätigkeit der Galerie=Commission und die Restauration der Gemälde” 1883-1891, fols 258-259. 9 Neidhardt, Schölzel in: Foister, Jones, Cool 2000, p. 34. 10 Faries 1999. 11 See nn. 1, 2.

366

uta neidhardt

12 Zeman 2002. 13 Handwritten restoration note by the restorer Carl Martin Schirmer on a piece of paper pasted on the back of the centre panel of the triptych, J[une] 1844. 14 Archive of the Staatliche Kunstsammlungen Dresden 01/GG 8, vol. 3 ‘Acten directorielle Aufzeichnungen über von der Galerie=Commission gepflogene Verhandlungen und gefaßte Beschlüsse enthaltend’ 1855-1869, fols 37, 38. 15 Sander 1993.

16 17 18 19 20 21 22

Herzner 2011b. Kemperdick, Lammertse 2012. Weale 1909 Neidhardt, Schölzel in: Foister, Jones, Cool 2000, p. 30-31. Neidhardt, Schölzel in: Foister, Jones, Cool 2000, p. 31. Heath, 2008. Streeton, 2011b.

24

Voyager dans les tableaux de Van Eyck Jacques Paviot

ABSTRACT: Travelling in Van Eyck’s Paintings Ever since research into Jan van Eyck’s paintings began, attempts have been made to identify the places, buildings and landscapes he represented. In reconsidering the question, I first tackle towns and buildings, then geographical settings, vegetation and stylistic borrowings. Bude or Prague can be put forward for the site of the town in the Virgin and Child with Chancellor Rolin, with details that could have been taken from Antwerp. The towers of Utrecht Cathedral or Old St Paul’s Cathedral in London might have been the inspiration for some of his buildings. The mountains could have been derived from the Pyrenees but the Alps are more likely. Experts have identified the rocks but it is difficult to give precise source locations. The vegetation is rich with the types of plant that grow in a Mediterranean climate, in the Iberian and Italian peninsulas. The artistic borrowings from Florence, Pisa, Siena and Assisi presuppose a sojourn in Italy. The caveat is that Van Eyck, in his authenticated works, never copied a motif in its entirety, but rather used and assembled details.

—o— À la suite de W.  H. James Weale,1 un certain nombre d’auteurs ont essayé d’identifier des monuments ou des parties de monuments, des villes, des lieux, des phénomènes physiques dans les tableaux des Van Eyck. En tant qu’historien, mon but, en voyageant dans les peintures des Van Eyck est d’essayer d’apporter des données sur la vie et les œuvres – avec toutes les précautions nécessaires, parce qu’il n’y aura pas de preuve historique, seulement des preuves formelles, et les résultats seront seulement des suppositions. J’espère ne faire montre d’aucun préjugé en faveur de telle ou telle solution. Un

autre point : ici, j’entends Van Eyck comme un atelier, un groupe de personnes, qu’ils soient Jan, Hubert, Lambert, un assistant, c’est donc un nom générique, ainsi les données que l’on voit dans les tableaux pouvaient se trouver dans l’atelier « Van Eyck ». Je suis conscient que l’art de Van Eyck se veut réaliste ou plutôt illusionniste, ce qui n’implique pas que les représentations soient vraies. Cependant, je pense que l’on peut tenter l’exercice, d’abord avec les vues de villes et de bâtiments, puis avec les phénomènes géographiques et la végétation, enfin avec les emprunts à d’autres œuvres d’art. Villes et bâtiments Beaucoup d’encre a coulé au sujet de la ville à l’arrière-plan de la Vierge à l’Enfant avec le chancelier Rolin. En faisant abstraction de l’arrière-plan montagneux, Genève, Lyon, Prague, Liège, Maastricht, Autun, Utrecht, ont été proposés,2 les champions de Liège étant Jean Lejeune3 et Joseph Philippe.4 Personnellement, je pense que, s’il y avait une ville réelle derrière l’idée générale du paysage du tableau, on ne peut présenter que deux possibilités : Bude et Prague.5 Aujourd’hui, la vue à partir du bastion des Pêcheurs,6 dans le château, sur le Danube et Pest est surprenante. Mais il n’y avait pas de pont au Moyen Âge. D’un autre côté, le pont Charles à Prague a été terminé au début du XVe siècle, et la vue de la tour de la cathédrale Saint-Guy, achevée durant le XVe siècle,7 est aussi impressionnante. Plus importante est la structure du pont, avec sa tour

368

jacques paviot

sur la rivière Vlatva. Une comparaison avec une représentation sur bois du XVIIe siècle (dans le chœur de la cathédrale Saint-Guy8) est intéressante. On y voit aussi une île au milieu du fleuve, ici l’île de Strelecky, qui aurait pu être la source de celle peinte par Van Eyck. Cependant il faut reconnaître que le pont peint par Van Eyck n’est pas le pont Charles. Dans l’état de nos connaissances, ce pont – copié dans la Vierge avec sainte Barbe, sainte Élisabeth et un chartreux de la collection Frick – ne semble pas être le « portrait » d’un pont réel. Dans le même tableau, il n’y a pas de doute au sujet de la tour figurant au-dessus de la tête de l’Enfant Jésus : c’est le « portrait » des deux étages supérieurs de la tour de la cathédrale Saint-Martin d’Utrecht :9 une lanterne octogonale sur une structure carrée ; elle a été construite entre 1321 et 1382. La même tour a été ajoutée dans le panneau de l’Adoration de l’Agneau au XVIe siècle.10 Cependant le peintre a changé de trois à deux le nombre des lancettes sur chaque côté de la tour. D’autres tours ont été proposées : de Saint-Jean à Maastricht (mais de la seconde moitié du XVe siècle, sur le modèle de celle d’Utrecht), de Notre-Dame à Amersfoort (mais édifiée entre 1444 et vers 1470),11 et on peut ajouter celle de l’abbaye Saint-Michel à Anvers, détruite à la Révolution française mais qui est connue par la gravure Antverpia Mercatorum Emporium, du début du XVIe siècle. Du même panorama, on peut relever un autre détail qui a pu être utilisé par Van Eyck : les murs bâtis sur la rive de l’Escaut et les navires et bateaux amarrés devant eux. Dans d’autres peintures, on peut reconnaître deux fois la tour de Notre-Dame de Bruges dans l’Agneau Mystique : à l’arrière-plan de l’Adoration de l’Agneau, entre d’autres tours, et dans le panneau des Pèlerins, surplombant une ville minuscule. Certains auteurs veulent voir le chœur de la cathédrale de Cologne dans une autre église du panneau de l’Adoration de l’Agneau, aux tours encore à construire,12 mais il est difficile de reconnaître l’édifice : par exemple, les arcades sont rondes dans le tableau, en ogive dans la réalité. Il est aussi difficile de reconnaître la tour sud de la cathédrale de

Cologne derrière Saint Barbe :13 la tour peinte est hexagonale alors qu’elle est en réalité carrée. Cependant, dans les deux tableaux, l’idée a pu venir de la cathédrale de Cologne. Dans la Vierge à l’Enfant avec sainte Barbe, sainte Élisabeth et un chartreux, et aussi dans le frontispice de la Cité de Dieu de Jean Chevrot14 d’un atelier eyckien, on voit un « portrait » de la cathédrale Saint-Paul de Londres, avec aussi bien la nef que la tour. Cette cathédrale a disparu dans le Grand Incendie de 1666, mais on la reconnaît d’après les vues de Londres réalisées par Antoon van den Wijngaerde, vers 1544.15 Phénomènes géographiques À l’arrière-plan du panneau des Chevaliers du Christ de l’Agneau Mystique, de la Vierge à l’Enfant avec le chancelier Rolin, de la Stigmatisation de saint François, de la Vierge à l’Enfant avec sainte Barbe, sainte Élisabeth et un chartreux, des Trois Marie au Tombeau, des Crucifixions de New York et de Berlin, Van Eyck a peint des montagnes de manière réaliste, certaines recouvertes de neige. Ce dernier détail a été étudié par Charles Sterling, qui y a vu les Alpes,16 mais sans aucune identification. Cela pourrait-il être les Pyrénées ? Jan van Eyck a pu être un membre de l’ambassade envoyée en Aragon durant l’été et l’automne 1426, en compagnie de Lourdin de Saligny, André de Toulongeon et des maîtres Jean de Terrant et Jean Hibert, pour traiter d’un projet de mariage entre le duc de Bourgogne Philippe le Bon et Isabelle d’Urgel.17 Nous ne connaissons pas l’itinéraire suivi. L’Itinéraire de Bruges, du début du XVe siècle, indique deux routes en direction de Montpellier : par Paris et Le Puy, et par Reims, Troyes, Chalon-sur-Saône, Lyon, Valence, Avignon ; à partir de Montpellier, l’itinéraire était le suivant : Béziers, Narbonne, Perpignan, Le Boulou, Le Perthus, La Jonquera, Figueras, Gérone, Moncada, Barcelone et, au-delà Tarragone, Tortose et Valence.18 D’un autre côté, Jan van Eyck n’a pas participé à la deuxième ambassade d’Aragon en 1427 (qui a emprunté un autre itinéraire19) et l’ambassade au Portugal, en 1428-1429,20

voyager dans les tableaux de van eyck

à laquelle il a participé, s’y est rendue par mer. Aucune autre source, d’archives ou des tableaux eux-mêmes, ne suggère un autre voyage en péninsule Ibérique. Si Van Eyck se trouvait dans l’ambassade de 1426 et si celle-ci s’est rendue en Aragon par voie de terre, il a pu utiliser le motif des Pyrénées ; cependant il y a plus de chance qu’il a utilisé celui des Alpes. Dans les tableaux mentionnés, sauf la Crucifixion de New York, des châteaux s’accrochent sur les pentes des montagnes ou les surplombent. De tels châteaux peuvent être vus sur des collines partout, par exemple – pour s’en tenir aux itinéraires possibles de Van Eyck – dans la vallée du Rhône le long de la route vers Avignon, et plus particulièrement sur les pentes italiennes des Alpes, dans le Piémont, en Lombardie, dans le Trentin et le HautAdige, en Vénétie. Au-delà des premiers massifs, les neiges éternelles couvrent les sommets les plus élevés. Dans la Crucifixion de Berlin, les chaînes couvertes de neige courent d’un bout à l’autre du panneau. Dans les Chevaliers du Christ, les massifs sont plus proches et on distingue moins les montagnes couvertes de neige. Celles-ci sont indistinctes dans les Trois Marie au Tombeau et la Stigmatisation de saint François. Elles sont plus détaillées dans la Vierge à l’Enfant avec le chancelier Rolin et la Crucifixion de Berlin. De plus, le grand massif couvert de neige sur la droite de la Crucifixion de New York semble être le même, inversé, que celui sur la gauche de la Vierge à l’Enfant avec le chancelier Rolin.21 Dans la Crucifixion de New York, Renzo Leonardi a reconnu dans ces montagnes le massif du Mont-Blanc, tel qu’on le voit du col de la Faucille, donc du nordest.22 Cependant, une telle identification ne peut être répétée pour toutes les montagnes peintes, qui semblent le fruit de l’imagination. Si nous pouvons en voir des parties réelles, nous sommes obligés de reconnaître que les montagnes de Van Eyck sont recomposées : par exemple, les contreforts peuvent être vus d’Italie, les sommets de France. Quand on voulait aller de Flandre en Italie, il y avait deux routes principales : à travers la Bourgogne et la

369

Savoie ou en remontant la vallée du Rhin et à travers les cantons suisses et le Tyrol.23 Il faut relever qu’à part les Alpes, il n’y a pas de souvenir géographique de l’Italie. Deux tableaux sont intéressants pour leur représentation de pierres et de roches, la Stigmatisation de saint François et la Crucifixion de New York. Dans la partie droite de la Stigmatisation, Charles Sterling a vu une représentation « fidèle » des rochers du Sasso delle Stimmate à La Verna, ce qui suggère que le peintre les a vus lui-même.24 Cependant, il n’y avait pas de raison à ce qu’un peintre ou même une personne du XVe siècle se rende à cet endroit : en effet, ce lieu somme toute reculé ne se trouvait pas sur une route soit de pèlerins, soit de marchands. À la droite de la peinture de Philadelphie (qui est plus détaillée que celle de Turin), les rochers sont soit a sequence of thicker limestone beds, 0.5– 1.2 ft thick, interlayered with more thinly-bedded limestones and shales,25 tels que l’on peut en voir dans la région de Dinant en Belgique ou en Espagne,26 soit two kinds of interlayered sedimentary lithologies, sandstone and shale,27 tels qu’on en trouve dans le Bassin parisien (jusqu’à la Flandre méridionale).28 Les blocs qui se sont écroulés au bas de la falaise sont de calcaire, et des fossiles y sont représentés.29 À la gauche de saint François, on peut voir la représentation d’une roche métamorphique selon Scott L. Montgomery,30 mais, pour Kenneh Bé, the dark grey color indicates a basalt, a dark, intrusive rock of magmatic origin.31 Pourtant il pourrait aussi s’agir de calcaire, comme on peut le voir d’après les joints (les fissures dans la pierre) et la végétation herbacée au-dessus.32 Derrière la tête de saint François, selon Scott L. Montgomery, there appear the somewhat exaggerated, but otherwise realistic formations of heavily weathered, relatively massive limestone outcrops.33 Cela nous fait penser aux Dolomites dans l’Italie du nord-est, mais au Moyen Âge il n’y avait pas d’alpinistes. Ces roches ressemblent étrangement aux roches de grès que l’on peut voir dans la Suisse saxonne et en Bohême, plus particulièrement à Bastei. Dans la

370

jacques paviot

même région, la vue à partir de Königstein supporte la comparaison avec le paysage à l’arrière-plan de la même peinture. Dans la Crucifixion de New York, il y a une représentation intéressante d’érosion en nid d’abeille (honeycomb weathering) au premier plan de ce qui est probablement du grès (mais qui pourrait être du granite ou du calcaire).34 Végétation Si nos regards se portent sur la végétation – un sujet déjà traité dans L’Agneau Mystique au laboratoire in 195335  —, plusieurs arbres « méditerranéens » apparaissent dans l’Agneau Mystique. Dans l’Adoration de l’Agneau, les Ermites et les Pèlerins, nous pouvons voir des cyprès et des dattiers. Le cyprès peint par Van Eyck est un cupressus sempervirens, var. Stricta, connu sous le nom de cyprès italien, toscan ou des cimetières. Il provient de Phénicie et est commun en Méditerranée orientale et en Italie ; il peut atteindre 35 m de haut. Le dattier (phoenix dactylifera) est répandu dans toute l’Afrique du Nord et le Proche-Orient. Il a été introduit dans la péninsule Ibérique durant l’Antiquité par les Carthaginois (un souvenir en est la palmeraie d’Elche à Valence) et en Sicile au haut Moyen Âge par les Arabes. Il peut atteindre 30 m de haut. Dans les Ermites, nous voyons un pin parasol adulte (pinus pinea). On le trouve autour de la mer Méditerranée et au Portugal. Sa hauteur peut atteindre 25  m. Dans l’Adoration de l’Agneau, Van Eyck a peint un figuier commun (ficus carica), un grenadier commun (punica granatum), et aussi, dans les Ermites et les Pèlerins des orangers (citrus aurantium). Le figuier, jusqu’au XVe siècle, était cultivé du Proche-Orient au Portugal ; sa hauteur peut atteindre 10 m. De même, le grenadier était cultivé du Proche-Orient au monde méditerranéen et son fruit a donné son nom à la ville de Grenade. Venant de l’Extrême-Orient via la Perse, l’oranger a été introduit en Italie au XIe siècle, ensuite dans la péninsule Ibérique qui envoyait ses fruits en Europe du nord à la fin du XIVe siècle. L’arbre a une hauteur moyenne de 10 m.36

Dans l’Agneau Mystique, il est un fruit que l’on voit deux fois : tenu par Ève dans son panneau, et, dans l’Adoration de l’Agneau, par le « prophète » vêtu de blanc et couronné de lauriers, celui-ci tenant d’ailleurs également une branche. Mais de quel fruit s’agit-il ? James Snyder a proposé la pomme d’Adam (citrus lumia, var. pomum Adami), qui était cultivé dans les péninsules Italique et Ibérique selon des descriptions et des gravures de la période moderne.37 Ronald Van Belle y a vu le combava (angl. kaffir lime ; lat. citrus hystrix), qui pousse dans l’Asie du sud-est,38 et plus récemment l’etrog utilisé lors de la fête juive des Huttes (Souccot).39 Cependant le citrus hystrix possède des doubles feuilles distinctives qui ne correspondent pas celle du rameau tenu par le « prophète » ; de plus l’etrog n’est pas un citrus hystrix, mais un cédrat (citrus medica).40 Ce dernier fruit, commun autour de la mer Méditerranée, peut être celui peint par Van Eyck comme il a fait des références aux Juifs dans ses peintures ;41 de plus ses feuilles sont les mêmes que celles qu’a peintes Van Eyck. Emprunts artistiques Ils semblent venir uniquement d’Italie : Florence, Pise Dans n’importe quelle étude sur l’Agneau Mystique, l’auteur compare la position de l’Adam eyckien avec celle de l’Ève, qui a les bras inversés, dans la fresque de Masaccio l’Expulsion du jardin d’Éden, dans la chapelle Brancacci, dans l’église Santa Maria del Carmine à Florence, datée généralement des années 1426-1427. Le prototype en est l’Aphrodite de Lysippe, connue par des copies romaines comme celle qui a été trouvée près de Sienne avant 1348,42 et dont on pouvait voir de nouvelles copies à Pise, au début du XIVe siècle, dans les œuvres de Giovanni Pisano ou de son cercle, telles que la tombe de l’empereur Henri VII, le groupe Donna Pisa, et la chaire de la cathédrale avec la statue de la Prudence. Dans son Ève, Van Eyck aurait repris le balancement de la Prudence, mais inversé.

voyager dans les tableaux de van eyck

Mais en parlant de la Prudence, il faut rester prudent puisque les frères de Limbourg, dans les Très Riches Heures du duc de Berry, ont donné des positions similaires à Adam et Ève : quand ils sont expulsés du jardin d’Éden ou représentés sous la forme de statues dans la Flagellation du Christ.43 On pourrait même remonter plus haut dans le temps : le torse de l’Ève eyckienne est semblable à celui de l’Adam sculpté que l’on pouvait voir sur la façade du transept méridional de Notre-Dame de Paris et qui est daté vers 1260.44 Assise Si nous pouvons considérer que Van Eyck ne s’est pas rendu à La Verna, il y a de fortes présomptions pour penser qu’il a séjourné à Assise. Au moins, pour un peintre, il y avait les fresques de Giotto à admirer, et il semble bien que Van Eyck les a vues, ce que signalent plusieurs emprunts. Il y a la position de frère Élie dans la Stigmatisation de saint François : c’est la même, mais inversée pour les bras, que celle d’un des frères qui se trouve au premier plan, sur la gauche, dans Saint François prêchant devant le pape Honorius III.45 La position de saint François a pu être reprise de celle du Miracle du Crucifix, de la Vision des trônes, du Miracle de la source. Van Eyck a choisi de peindre saint François les deux genoux à terre, aussi devait-il représenter ses deux pieds, comme Giotto l’a fait pour le médecin de la Vérification des stigmates. Enfin, l’idée de la cordelière dont l’extrémité repose sur le sol peut venir de la Stigmatisation de Giotto. Sienne On observe au moins un emprunt d’une peinture siennoise : la fontaine de l’Adoration de l’Agneau est copiée des Ermites à la fontaine d’Élie par Pietro Lorenzetti, la prédelle du retable des Carmes, peinte en 1329.46 Les deux bassins sont de marbre, et les conduits centraux ont de grandes ressemblances. Le dessin est le même, même si le conduit est de pierre chez Lorenzetti, métallique chez Van Eyck. Le bassin est hexagonal chez Lorenzetti, mais asymétrique chez Van Eyck. Le peintre du Nord a

371

commencé un bassin en hexagone à l’avant et l’a terminé en décagone à l’arrière, probablement pour lui donner plus de profondeur, mais au prix de la perspective. Pour conclure, je présente quelques propositions, selon la chronologie, sur les voyages de Van Eyck, sans prendre en compte les anciens Pays-Bas : r je suggère, mais ne le jurerais pas, que Van Eyck a pu aller en Bohême ; les seuls éléments sont le pont dans la Vierge à l’Enfant avec le chancelier Rolin (et la Vierge à l’Enfant avec des saints et un chartreux) et les rochers dans la Stigmatisation de saint François. En venant d’Allemagne, on suit la route de Dresde à Prague à travers la Suisse saxonne, en remontant la vallée de l’Elbe. Karel Kmet propose la date de 1422, quand Jan van Eyck était au service de Jean de Bavière ; r il n’y aucun indice dans les peintures du « loingtain voiage secret » de 1425-1426 ; r au contraire, nous pouvons voir plusieurs souvenirs de l’ambassade au Portugal de 1428-1429 : la sibylle de Cumes de l’Agneau Mystique qui serait une reprise du portrait perdu de l’infante Isabelle ; les différents types d’arbres : pin parasol, palmier, oranger, figuier, grenadier ; il aurait pu aller à Londres, lors du voyage de retour ; r je suggère que Van Eyck est allé en Italie dans les années 1430-1432, la fontaine ne se trouvant pas dans la composition originale de l’Adoration de l’Agneau et ayant été ajoutée plus tard ; on peut raisonnablement penser qu’il est passé par les Alpes, puis s’est rendu à Florence, Pise, Sienne, Assise (et non à La Verna) ; le cyprès peut constituer un autre souvenir ; r il n’y aucun indice dans les peintures des « certains voiaiges loingtains et estranges marches où mondit seigneur l’a envoié pour aucunes matieres secretes » en 1436.47 En ce qui concerne la réalisation des peintures, Van Eyck a composé un mélange des paysages, sites, villes, bâtiments réels qu’il a vus et dessinés : par exemple, dans la Stigmatisation de saint François,

372

jacques paviot

il place trois sortes de roches les unes à côté des autres ; dans la Vierge à l’Enfant avec le chancelier Rolin, il est impossible de reconnaître une ville réelle : le seul bâtiment identifiable est la tour de la cathédrale d’Utrecht, mais seulement dans ses deux niveaux supérieurs. Aussi, il semble curieux que soit entièrement représentée la cathédrale SaintPaul de Londres dans la Vierge à l’Enfant avec des saints et un chartreux. Enfin il est curieux que pour des voyages secrets à caractère diplomatique dans des régions qui semblent être en marge ou au-delà de l’Europe, Van Eyck semble ne pas s’être muni d’un carnet de dessins.

NOTES 1 Weale 1908. 2 Liste donnée sans références par Dhanens 1980, p. 273. 3 Lejeune 1956, p. 79-159. 4 Philippe 1969, p. 59-100. 5 Notebaert 1939, p. 2-13 (je n’ai pu consulter cet article), rejeté par Panofsky 1953, vol. I, p. 430-431 ; Kmet 2003-2007. 6 Mais construit entre 1896 et 1902. 7 Par exemple Benesovská, Hlobil, 1999, p. 125-126. 8 Une série de panneaux de bois sculptés, fixés contre la paroi interne du déambulatoire, représentent la fuite de l’Électeur palatin Frédéric, « roi d’un hiver » de Bohême, et sa femme Élisabeth Stuart, à la suite de la bataille de la Montagne blanche en 1620. 9 Lejeune 1956, p. 137. 10 Van Asperen de Boer 1979, p. 199. 11 Lejeune 1956, p. 137. 12 Lauer 1992, p. 309-314. 13 Dhanens 1980, p. 262-265. 14 Vers 1445, Bruxelles, KBR, MS 9015, fol. 1. 15 Colvil, Foister 1996. 16 Sterling 1976, p. 29.

17 Dans mon article de 1990 (Paviot 1990b), suivant les données des archives bourguignonnes, j’ai pensé que cette ambassade a eu lieu en 1425. Le mandement de paiement a été émis le 31 juillet 1426, en fait avant que l’ambassade ne soit partie, et le sauf-conduit du roi Alphonse V d’Aragon a été délivré le 28 septembre 1426. Voir Blasco Vallés, Reche Ontillera 2007. 18 Itinéraire de Bruges (XVe siècle) 1908, p. 171-187. 19 Paviot 1990a ; Spitzbarth 2013, p. 616-621. 20 Paviot 1990b ; Paviot 1995, p. 205-218. 21 Je remercie Renzo Leonardi de cette remarque. 22 Leonardi 2007. 23 Voir les pèlerinages contemporains du Tournaisien Coppart de Velaines : Paviot 2007 ; voir aussi l’Itinéraire de Bruges. 24 Sterling 1976, p. 29-31. 25 Montgomery 1996, p. 10. 26 Montgomery 1996, p. 9. 27 Bé 1997, p.  88-95, ici p.  89 ; il s’agirait plus exactement d’« a flysch or turbidite sequence » (p. 90). 28 Bé 1997, p. 90. 29 Montgomery 1996, p. 10 ; Bé 1997, p. 88. 30 Montgomery 1996, p. 10. 31 Bé 1997, p. 91. 32 Je remercie mon collègue Stéphane Cordier, du département de Géographie à l’Université Paris Est Créteil, de son aide pour l’identification des roches. 33 Montgomery 1996, p. 10. 34 Mes remerciements réitérés à Stéphane Cordier. 35 Coremans 1953, p. 123-125. 36 Coremans 1953, p. 123-125. 37 Snyder 1976. 38 Van Belle 2002. 39 Van Belle 2012 ; Dequeker 2011. 40 Voir entre autres, The Jewish Encyclopedia 1901-1906, s. v. etrog. 41 Paviot 2006. 42 Von Schlosser 1904, p. 141-150 ; Mahler 1905. 43 Fol. 25v° et 144r° ; voir Châtelet 2011, p. 128-129. 44 Paris, Musée national du Moyen Âge (de Cluny). Orignellement, il tenait une pomme dans sa main droite. 45 Nous pouvons aussi relever que ce frère assis se retrouve, mais vu de trois quarts et de dos, dans un bas de page des Heures de TurinMilan (Trois ermites près d’une forêt, Paris, Musée du Louvre, RF 2025). 46 Sienne, Pinacoteca nazionale. Voir Dhanens 1989. 47 Albert Châtelet avance l’idée que Jan van Eyck serait allé en Terre sainte ; voir Châtelet 2011, p. 234-235.

Fig. 25.1 Jan van Eyck, St Francis Receiving the Stigmata, 1430s, oil on panel, 29.2 x 33.4 cm Turin, Galleria Sabauda (cat. no. 187) Fig. 25.2 Jan van Eyck, St Francis Receiving the Stigmata, 1430s, oil on vellum on panel, 12.4 x 14.6 cm [image size does not include later addition of border] Philadelphia, Philadelphia Museum of Art, John G. Johnson Collection (cat. no. 314)

25

Replications of Exemplary Form. New Evidence on Jan van Eyck’s St Francis Receiving the Stigmata Jamie L. Smith

ABSTRACT: Prior technical studies on St Francis Receiving the Stigmata indicated that Jan van Eyck painted the panel, now in Turin, which his workshop replicated on parchment to create one of the earliest copies of an oil painting, now in Philadelphia. This paper examines manuscript illuminations, archival documents, devotional texts, and geological and architectural evidence – none previously assessed in connection with van Eyck’s works – to discover the meaning and function of this pair of paintings. Van Eyck endowed his picture with the authority of an exemplar by codifying patterns of divine image production and reproduction in its imagery. His colourful depictions of Christ’s wounds figure the detailed descriptions of the crucifixion in Netherlandish Passion literature. The duplication of Christ’s wounds in St Francis’s stigmatization signals that Christ is a type for formal replication. Thus the picture’s realism and subject endorsed its copying. Van Eyck’s composition echoes specific liturgical images. This correspondence suggests that the recapitulation of the Passion in the Eucharist is a theme of the picture. Physical evidence located in the Adornes family’s Jerusalem Chapel and archival records of its chaplaincy provide the bases for new theories concerning the commission and purpose of the Turin panel and elucidate the significance of its replication.

—o— Jan van Eyck’s two paintings of St Francis receiving the stigmata, generally thought to date from the 1430s, represent one of the earliest known instances of the replication of an oil painting.1 Much debate has focused on the origins, meaning and function

of the unsigned pictures. This article examines evidence from illuminated manuscripts, archival documents, devotional texts, architectural fixtures, and geological specimens, none of which previously has been assessed in connection with Van Eyck’s works. Formal and iconographic analyses of the imagery provide new insights into the purpose of its reproduction and open an alternative line of inquiry into the patronage of the paintings. Extensive technical studies performed between 1983 and 1997 indicate that Jan van Eyck painted the panel (fig. 25.1), which is now in Turin, and that he or his workshop created the smaller replica on parchment (fig. 25.2), which is now in Philadelphia. Infrared reflectography revealed extensive underdrawing in the Turin picture but virtually no underdrawing in the Philadelphia one.2 Marigene Butler and Rudolf van Asperen de Boer therefore concluded that the Turin panel is an original work and the Philadelphia picture is its copy. Van Asperen de Boer reiterated the attribution of the Turin painting to Jan van Eyck, noting similarities between its underdrawing and those of the Ghent Altarpiece, and the Virgin and Child with Chancellor Rolin.3 Butler observed, with stereomicroscopy and projection, that the Philadelphia painting replicates the detailed Turin imagery nearly exactly, with a few spatial differences.4 She asserted the

376

jamie l. smith

brushwork is so similar that the copy was painted by the same master, or by a close assistant.5 The Philadelphia parchment support is affixed to an oak panel. Through dendrochronological analysis, Peter Klein found that this panel was cut from the same tree as van Eyck’s portraits of Giovanni Arnolfini and Baudouin de Lannoy, both now in Berlin. Klein concluded that the Philadelphia picture was made in Van Eyck’s workshop.6 The imagery portrays a well-known episode in St Francis’s life: his miraculous vision and stigmatization at Mount La Verna in Italy. In a landscape, the saint kneels, posed almost in right profile, with arms bent and hands extended in parallel. He faces a seraph-winged figure of Christ, nailed to a Cross that hovers over the rocks. The Cross is in the form of a Tau, which was the personal emblem of St Francis.7 Bloody wounds on his hands and feet resemble those of Christ.8 To his right, a companion, Brother Leo, sleeps on the ground. As many have observed, Van Eyck’s portrayal deviates from precedents.9 The saint’s pose and Christ’s figure do not conform to the Italian tradition, wherein this subject is common. Images of St Francis’s stigmatization are rare in northern European art before Van Eyck. The precedent noted by scholars is an illumination painted by an unknown master, around 1410, in a Book of Hours commissioned by Jean  II le Meingre de Boucicaut, Marshal of France (1365-1421).10 In the Boucicaut Master’s picture, Brother Leo leans on a text and stares downward, unlike the sleeping figure in Van Eyck’s version. One feature in common with Van Eyck’s composition is the Tau cross. But a seraph, not Christ, appears on the Boucicaut Master’s cross. The pose of Van Eyck’s saint does not match the pose of the Boucicaut Master’s figure, which recalls earlier Italian depictions, with Francis’s hands spread wide apart, and his wounds connected by lines to the seraph. Whereas Van Eyck’s St Francis was not patterned after the Boucicaut depiction, or after earlier Italian precedents, I contend that contemporary models for the pose were readily available

to the Flemish artist. Consider the illustration (fig. 25.3) at the beginning of the Canon of the Mass, on the Te Igitur page of a missal made for a church in Bruges in the fifteenth century.11 The historiated ‘T’ that begins the Eucharist rite shows the consecration of the Host. The celebrant kneels in right profile and elevates the Host with hands parallel. He faces an altar with an altarpiece featuring a cross. A similar example is found in a missal made between 1478 and 1492 for Ter Doest abbey near Bruges.12 This type of Te Igitur image was common in missals produced in Burgundy and the Netherlands during the fourteenth and fifteenth centuries.13 I hypothesize that Van Eyck modelled his composition for St Francis Receiving the Stigmata on this typical form of Te Igitur illustration. St Francis’s pose, in relation to the rocks and the Cross, resembles a priest celebrating mass at an altar surmounted by a crucifix. It is likely that Van Eyck saw illustrations in Flemish missals because, as many art historians have argued, he may have worked as a manuscript painter early in his career.14 Moreover, detailed depictions of open pages in multiple texts on the interior panels of the Ghent Altarpiece demonstrate Van Eyck’s close familiarity with liturgical books.15 In Flemish missals, Te Igitur pages often faced a full page crucifixion with a Tau cross, as in the missal from Bruges. Such crucifixion paintings were commonly the artistic highpoint among illuminations in a missal.16 Alluding to this convention in liturgical art, Van Eyck asserted his skill and signalled that artistic creation and Eucharistic conversion are themes of his picture. Van Eyck’s imagery evokes the words spoken to convert bread and wine into Christ and recalls St Francis’s writings on Christ’s presence in the Eucharist.17 Theologians cited Christ’s appearance to St Francis at his stigmatization as proof of the divinity’s concurrent location in heaven and on diverse altars.18 Francis’s thirteenth-century biographer, Bonaventure, and the influential Dominican writer, Thomas Aquinas, agreed that presence through Eucharistic conversion does not require

replications of exemplary form

377

Fig. 25.3 Missal from a church in Bruges, detail, 15th century, parchment, overall 31.1 x 22.3 cm Bruges, Openbare Bibliotheek Brugge, MS. 314, fol. 88r

bodily change in Christ.19 Aquinas explained that the place the bread occupies is ordered to the body of Christ through the remaining dimensions of the bread.20 Van Eyck may have shown Christ’s body, instead of the traditional seraph, to demonstrate this principle. St Francis’s physical dimensions – his size and features – remain constant while he receives the wounds. The celebrant’s pose conveys that the saint’s body is ordered to Christ’s body. The Dominican theologian, Dirc van Delf, wrote that Christ’s Passion may be understood as ‘a picture’ and ‘an example’ for the performance of the sacrament.21 This concept agrees with the assertion, in an anonymous fourteenth-century Netherlandish text on the symbolism of the mass, that the position of the celebrant’s hands recalls the

position of Christ’s hands on the Cross.22 We may therefore conclude that Van Eyck represented St Francis in the pose of a celebrant at an altar to evoke the symbolic re-enactment of the Passion in the rite of consecration. Van Eyck’s painting identifies the replication of Christ’s body and blood through conversion in the Eucharist with the replication of Christ through stigmatization in St Francis. Francis’s imitation of Christ’s exemplary life was thought to have merited his conversion into Christ.23 The dynamics of imitation and replication expounded in Van Eyck’s imagery endorse his picture as an exemplar to be emulated. The artist apparently succeeded in promoting the painting’s exemplary status because its imagery was indeed emulated, first in its copying,

378

jamie l. smith

at least once in his workshop, and over the next several decades, in its imitation by artists working in Flanders, Italy, Spain and Germany.24 Van Eyck’s construal of Christ as a generative source agrees with the devotional trope of Christ as an artistic creator. An anonymous sermon written in Bruges, around 1487, exclaimed hoe groot een constenare (‘how great an artist’) man’s Creator is, describing him as Een abel scildere (‘a skilful painter’) who paints his self-portrait in the soul.25 The Flemish spiritual leader Jan van Ruusbroec asserted, around 1362, in The Little Book of Enlightenment, that Christ ‘remade man’ through his death.26 In the eleventh-century treatise Why God Became Man, the Flemish-born theologian Anselm of Bec argued that visible evidence of Christ’s suffering demonstrated his humanity and death. He explained, ‘[a] picture of the body, which proves that it was fitting for God to condescend to [the crucifixion], must be shown so the very body of the truth may shine more brightly.’27 Van Eyck depicted the crucified Christ, in St Francis Receiving the Stigmata and in the Crucifixion, in New York, with a pale body, red bleeding wounds and a dark face.28 His imagery recalls the focus in Netherlandish worship on Christ’s bleeding, laceration and bruising. Such emphasis is demonstrated in these verses from an early-fifteenthcentury West Flemish prayer book: Dat bloet liep over dinen mont zo smart Dat dijn ansichte besmet wart (‘Blood ran over your mouth so sore, Thus your face was stained’) and recount Mary’s sight of Haer lieve kint versceeden ant cruce, en[de] zijn lechame wart al zwart (‘her beloved child torn on the Cross and his body was all dark’).29 Though Van Eyck’s naturalistic style was rooted in the late-medieval Netherlandish painting tradition, devotions focused on the tearing, staining, and darkening of Christ’s flesh may have informed his realism, prompting him to present the wounded Christ as a model for realist picturing.30 Van Eyck’s impact is evident in Christ as the Man of Sorrows (fig. 25.4), painted around 1450 by Petrus Christus.31 Christus’s picture corresponds to

Fig. 25.4 Petrus Christus, Christ as the Man of Sorrows, oil on wood, 11.2 x 8.5 cm, c.1450 Birmingham, Birmingham Art Gallery (acc. no. 1935P306)

St Francis Receiving the Stigmata in its liturgical theme as well as in its detailed description of Christ’s wounded flesh. The stylistic and iconographic construct in Christ as the Man of Sorrows facilitates our understanding of the symbolism in Van Eyck’s picture of St Francis. In Christus’s painting, Christ’s face is dark. His body is pale, discoloured by blue-grey bruises, and stained with blood. He is not a passive sufferer, but an active creator. He squeezes the spear wound, colouring his side red with flowing blood, a gesture that figures the application of paint on the panel.32 Angels holding the Lily of Mercy and the Sword of Justice reveal Christ by drawing aside green curtains. Curtains are itemized in inventories of chapels in Bruges.33 The charter for the chapel of the Confraternity of the Dry Tree lists Petrus Christus as a member and itemizes cortinin (curtains) in its inventory.34 Christus’s green

replications of exemplary form

curtains are similar to the green curtains depicted around the crucifixion altarpieces in illustrations on Te Igitur pages (fig. 25.3) in the missals discussed above.35 Christus’s drapery mimics curtains surrounding pictures of Christ’s wounded body on altars. This presentation casts Christ in a Eucharistic context as both image and image maker. Christus’s picture, in portrait format, suggests that, as he endured fatal wounds, Christ remade himself into a new image, a self-portrait, coloured by his sacramental blood. St Francis Receiving the Stigmata presents this concept in a typological argument. An element of this typology is recognizable in Van Eyck’s portrayal of Brother Leo, who sleeps instead of pondering a text. James Snyder argued that Leo refers to the disciples who slept as Christ prayed on the Mount of Olives (Lk 22: 41-45), noting that the chalice Christ asked to forego and the

379

blood he shed with his sweat foreshadowed the Eucharist.36 In addition to these associations, I assert that Leo figures Adam, who slept on the ground as God removed his rib to form Eve (Gen 2: 21-22). This subject is represented in a historiated initial (fig. 25.5) in a book of hours made in Utrecht between 1455 and 1460.37 Leo’s pose, though more vertical and compact, is comparable to that of Adam, who sleeps on the earth with his head propped on his hand and his legs bent. Van Eyck connected Leo’s side visually to St Francis’s side with cord belts. The saint, posed as a Eucharistic celebrant, corresponds to Eve, who prefigured the sacraments that issued from Christ’s wounded side, which was replicated in Francis. Augustine explained how the birth of the sacrament from Christ’s spear wound was prefigured by Eve’s birth from the side of the sleeping Adam: ‘This second Adam bowed His head and fell asleep on the Cross,

Fig. 25.5 Book of hours made in Utrecht,detail, c.1455-1460, parchment, overall 19 x 14 cm The Hague, Koninklijke Bibliotheek, 135 E 40, fol. 27v

380

jamie l. smith

that a spouse might be formed for Him from that which flowed from the sleeper’s side.’38 The spring pouring from the rock beside Leo recalls water flowing from the rock that Moses struck to quench the thirst of the Israelites as they crossed the desert after escaping slavery in Egypt (Ex 17: 1-7). The image evokes St Paul’s identification of the rock with Christ and, thus, elaborates on the meaning of the sacraments flowing from him (1 Cor 10: 4). Van Eyck’s arrangement of these figures and objects is consistent with a typological grouping found in illustrated texts of the Biblia pauperum made in the Netherlands and Germany during the fifteenth century.39 Such books display the creation of Eve from Adam’s rib on the left of a page, Christ receiving the blow of the lance in the centre, and Moses striking water from the rock at the right. The images in the Biblia pauperum expound the theological precept that Christ’s side wound, which was prefigured by the creation of Eve and by Moses striking the rock, generated the sacraments of the Church. This meaning also attaches to the images of Leo, St Francis, Christ, and the spring flowing from the rock in St Francis Receiving the Stigmata. The pose derived from Adam identifies Leo as an embodiment of the innate, but latent, image of the Creator in the first man. Van Eyck’s imagery contextualizes Christ as the second Adam, whose death produced the sacraments and remade the image of the Creator into a new image of the wounded saviour, one that was later reproduced in St Francis’s stigmata. The setting Van Eyck painted for the stigmatization is dominated by an outcropping of rocks. The artist articulated these stones (fig. 25.6) so exactly that geologist Kenneth Bé determined them to be highly accurate renderings of fossils of marine invertebrates in white sandstone, as are found specifically in France and south Flanders.40 The prominence and specificity of the fossils indicates that they have a significant and particular meaning in the picture.41 In a thirteenth-century lapidary, Jacob van Maerlant described fossils as dondersteenen (‘thunderstones’) formed in clouds by

Fig. 25.6 Jan van Eyck, St Francis Receiving the Stigmata, detail, 1430s, oil on vellum on panel, overall 12.4 x 14.6 cm Philadelphia, Philadelphia Museum of Art, John G. Johnson Collection (cat. no. 314)

lightning.42 Van Maerlant’s description is consistent with the writings of ancient authors, including Pliny the Elder, who claimed that such stones were diipetes ‘thrown by the gods.’43 Therefore, Van Eyck’s fossils may be understood to signify divine creation, thereby amplifying Christ’s miraculous creation of his image in Francis. The rocks ‘thrown by the gods’ symbolically align Van Eyck’s painting with acheiropoieta, a class of objects believed to have been created through divine agency. Legendary icons, including the Veil of Veronica and the Mandylion of Edessa, were images of Christ’s face purportedly not made by human hands.44 According to tradition, these miracleworking pictures were created by Christ and had the power to reproduce themselves in replica images.45

replications of exemplary form

The implied association between the wondrous stones and these icons reinforces the motif of selfgeneration that unfolds in Van Eyck’s two pictures of St Francis’s stigmatization. Leo’s identification with Adam elaborates the theme of divine creation, both in its association with the creation of Eve, and in its implicit reference to Adam as a picture of God. Dirc van Delf asserted, Adam was naden va beelde Gods ghescapen, Cristus die was eygentlick dat beeld Gods selve, dat alle ander beelden conde maken (‘Adam was made after the image of God, yet Christ alone was the picture of God himself which could make all other pictures’).46 Van Maerlant and others attributed magic properties to fossils, including the power to promote restful sleep.47 As we see, Brother Leo slumbers under their spell. Yet, his identification with Adam suggests that he might awaken in accord with the Apostle Paul’s entreaty in Ephesians (5:14): ‘Awake thou that sleepest, and arise from the dead, and Christ shall give thee light.’ St Jerome reiterated this appeal to explain how Christ’s blood ran from the Cross and washed away Adam’s sins.48 Leo’s placement, among the rocks beneath a cross, elicits the legendary location of Adam’s tomb at Golgotha. The reference to Golgotha in Van Eyck’s picture prompted me to consider the possibility of a connection between St Francis Receiving the Stigmata and the Kalvarieberg Altar (fig. 25.7) in the Jerusalem Chapel, built by the Adornes family as their private chapel in Bruges.49 In the conclusion of this essay, I will offer preliminary observations and speculative ideas about associations between the altar and Van Eyck’s paintings and propose a potential new line of inquiry into the patronage of the paintings. The unusual Kalvarieberg Altar shares several features with Van Eyck’s painting. The Kalvarieberg is surmounted by three crosses. Adam’s skull and bones are carved below them in relief on the face of the altar. This arrangement corresponds to Van Eyck’s grouping of Christ on the Cross and Leo, Adam’s embodiment, seated amid the rocks below. Instruments of the Passion, which created

381

the wounds that Christ replicated in St Francis, are carved over the altar’s surface. The altar, bearing traces of polychrome, is made of white sandstone, like the rocks Van Eyck painted to symbolize an altar. At either side of the Kalvarieberg Altar, a staircase ascends to the choir. The stairs are carved from stones containing marine invertebrate fossils (fig. 25.8) which resemble the fossils in the painting. Fossils were ascribed powers comparable to miracle-working relics.50 The altar served as a reliquary for a fragment of the Holy Sepulchre brought from Jerusalem by the Adornes brothers, Jacob and Pieter, who, in 1427, received papal permission to build a chapel dedicated to the Holy Sepulchre. The Jerusalem Chapel, completed around 1483, was modelled on the Church of the Holy Sepulchre in Jerusalem and contains a replica of Christ’s tomb.51 Pieter’s son, Anselm (1424-1483), laid the foundation stone when he was a child.52 On 10 February 1470, prior to a pilgrimage to the Holy Land, Anselm made a will wherein he bequeathed two paintings by Jan van Eyck ‘of St Francis in portraiture’ to his daughters Marguerite and Louise. James Weale and others identified the Turin and Philadelphia pictures with this bequest.53 I agree with this identification and, moreover, wonder whether the Turin panel may have originally been commissioned to be placed in the Jerusalem Chapel, perhaps on or near the Kalvarieberg Altar, or, alternatively, whether the picture may have been conceived for private domestic use to commemorate or honour the chapel. Though Luc Devliegher and Valentin Vermeersch dated the carved altar 1430-1435, Jean-Pierre Esther placed its date around 1485.54 Information presently available about the various stages of the chapel’s construction and decoration does not permit a conclusive finding. If the earlier date is accurate, Van Eyck’s imagery may have been either conceived with the altar, or derived from it. If the later date is correct, then the painting may reflect plans for, or be the inspiration for, the Kalvarieberg design. The painting’s presence in the chapel would have invoked

382

jamie l. smith

Fig. 25.7 Anonymous, Kalvarieberg Altar, 15th century, sandstone, c.375 x 319 cm (without crosses) Bruges, Jerusalem Chapel

replications of exemplary form

383

Fig. 25.8 Pavement of staircase, detail, 15th century, stone, c.12 x 10 cm Bruges, Jerusalem Chapel

St Francis’s guardianship of the replica shrine because the Franciscan Order was the appointed Guardian of Christ’s tomb.55 Many scholars have argued that Van Eyck’s depiction of St Francis is a portrait because his head and face appear to display individualized features.56 If it is a portrait, the liturgical pose indicates that the sitter was a priest. In my mind the most likely candidate is Jan Beert, magister artium, priest, and chaplain of the Jerusalem Chapel. Documents indicate that Beert was a close friend of the Adornes family and had a personal devotion to the Jerusalem Chapel.57 Beert is first mentioned as chaplain of the Jerusalem Chapel, along with a second chaplain, Johannes van den Coutere, in the archives of the Jerusalem Foundation in 1435.

A bequest of books he made to the chapel library included a text of writings by Thomas Aquinas. Beert attempted to donate farm land in Kortrijk to the Adornes foundation. Though the donation was prevented by his brother, Beert’s effort demonstrates his devotion to the Jerusalem Chapel. He, along with Jacob and Anselm, was co-executor of Pieter Adornes’s will, written between 22 February and 8 April 1452. The chaplain died between 24 November 1452, when he is last mentioned in the foundation archives, and 15 October 1454, the date of a document founding a daily mass in the Jerusalem Chapel for his salvation. If Beert were the sitter, he may have posed for van Eyck as early as 1435, when he is first mentioned, or as late as 1441, the year the artist died.

384

jamie l. smith

Van Eyck’s portrayal resembles a man around forty years of age. If it is Beert, then we may estimate that the chaplain was somewhere between the ages of fifty-one and fifty-nine at his death. Either the chaplain or his employer and friend, Pieter Adornes, may have commissioned the pair of paintings from Van Eyck. Or one of these two men may have commissioned the Turin panel, and the other commissioned its copy. The picture’s theme of selfreplication indicates that the copy was planned along with the original. Pieter Adornes may have commissioned the Turin panel from Van Eyck with the instruction that St Francis was to resemble his friend Beert. However, I think it is more likely that Beert commissioned the painting himself. He may have given this painting as a gift to Pieter. Or he could have donated it to the Jerusalem Chapel, where he later established a foundation for mass to be said daily for his salvation. These circumstances of the later scenario are similar to the commission of the Virgin and Child with Canon Joris van der Paele, which contained a donor portrait of a churchman and was donated to a chapel where the donor established a foundation for daily masses in his memory. If Beert commissioned the Turin panel for the chapel, he may also have commissioned the Philadelphia copy as a gift to his employer, Pieter. It is also conceivable that Pieter commissioned the replica for his personal possession, or as a gift for a close relative, such as his son Anselm, whose eventual ownership of both works is indicated in the latter’s will. The replication of the exemplar of Christ in St Francis that is explicated in Van Eyck’s painting parallels the simulation of Christ’s tomb in the Jerusalem Chapel. The picture’s liturgical composition may honour the chapel’s function as a place for the repeated performance of the Eucharist. There the chaplain Jan Beert symbolically re-enacted the Passion and enunciated the words Te Igitur to reproduce Christ’s body in the Host, which was administered to the members of Adornes family at Holy Communion. Thomas à Kempis suggested in the Imitation of Christ, written around 1428, that

communicants who imitate Christ’s suffering could be transformed into Christ by his Eucharistic presence.58 Anselm Adornes may have wished to signify his part in the chain of divine replications. He directed in his will that portraits of himself and his wife Margareta be painted and attached as wings to Van Eyck’s picture.59 These additions would have memorialized the familiar assembly of husband and wife, participating in the mass, performed by the chaplain in the Jerusalem Chapel. NOTES 1 St Francis Receiving the Stigmata, 1430s, oil on panel, 29.2 ≈ 33.4 cm, Galleria Sabauda, Turin, no. 187. St Francis Receiving the Stigmata, 1430s, oil on vellum on panel, 12.4 ≈ 14.6 cm (not including later border), Philadelphia Museum of Art, John G. Johnson Collection, no. 314. Art historians have assessed the attribution and dating of the works, suggesting dates ranging from 1427 to 1438. See Mather 1906, pp. 358-359; Ricketts 1906, p. 426; Friedländer 1924, pp. 101-102; Aru and de Geradon 1952, pp. 5-13; Friedländer 1967, pp. 62-63; Faggin 1968, no. 5; Sterling 1976, pp. 7-81; Stroo, Smeyers 1994, pp. 291-292; and Watkins 1997, passim. The provenance of the Turin panel has been traced to an unidentified former nun in Casale Monferrato, in the province of Alessandria, from whom Luigi Fascio, mayor of Feletto Canavese, in the province of Turin, acquired it in the early nineteenth century. The work was acquired from him in 1866 by Massimo D’Azeglio, director of the Turin museum. See Hymans, 1888, pp. 78-83. The Philadelphia picture was first recorded in the possession of a Scotsman, William à Court (1779-1860), who transported the work from Lisbon to London in 1827. In an inventory he stated that he obtained the work, misattributed to Albrecht Dürer, from a physician in Lisbon. See Rishel 1997, pp. 3-4. St Francis Receiving the Stigmata is not the only painting by Van Eyck that was copied under his direction. Art historians have argued persuasively that a replica of Van Eyck’s Virgin by a Fountain, dated 1439, was reproduced in a workshop copy completed around 1440, which is now in the Robert Noortman collection, Maastricht, and was formerly owned by Margaret of Austria. See Ghent/ Amsterdam 2002, exh. cat., nos 26 and 27, pp. 235, and 236. Also see New Haven/London 2004, exh. cat., no. 353, pp. 589 - 590. Copies of two versions of the Holy Face by Van Eyck also survive. The Holy Face in the Groeningemuseum, Bruges, was copied around 1635 after an Eyckian model dated 30 January 1440. Christ’s collar is decorated with gemstones. Another version, in the Staatliche Museen, Gemäldegalerie, Berlin, was copied during the sixteenth century after an Eyckian model dated 31 January 1438. Christ is depicted in a robe with an embroidered collar. Other copies are in the Alte Pinakothek, Munich and in a private collection in England. See Panofsky 1971, pp. 187, 430, n. 187-1; Ghent/Amsterdam 2002, exh. cat, no. 37, p. 239; and Dhanens 1980, pp. 292-294, figs 292, 294. Regarding additional copies of the Holy Face after Van Eyck, see Smeyers 1994, pp. 195-224. Van Eyck also painted two portraits of Isabel of Portugal, one of which was sent to Philip the Good by land and the other by sea, to ensure at least one picture would reach him safely. Luber 1998, pp. 24-37, esp. p. 31. 2 Butler 1997, pp. 32-42. 3 Van Asperen de Boer 1997, p. 52. Also see Sterling 1976, pp. 14-15. 4 Butler 1997, pp. 34, 38, 39, 42. 5 Van Eyck may have manipulated his brushstrokes in the Philadelphia painting to achieve similar optical effects as seen in the Turin

replications of exemplary form

panel in a different scale. See Luber 1998, pp. 32-33. See also Streeton in this volume. 6 The earliest felling date for the tree from which the two sections of the panel were cut is 1392. Klein estimated the most plausible felling date statistically to be between 1396 and 1402. Klein 1997, pp. 47-49. 7 Van Maerlant 1954, ll. 5095, 5390, 7295-7298, 8069-8078. 8 X-radiographs of the Turin panel reveal that Van Eyck repainted Francis’s feet, which he had originally depicted as covered, to reveal the wounds replicated in them. See Butler 1997, pp. 23-25. 9 I introduce for comparison two thirteenth-century psalter illustrations made in France in which Francis receives the stigmata standing or kneeling in front of an altar: Carpentras, Bibliothèque municipale (Bibliothèque Inguimbertine) MS 0077, fol. 180v; and Paris, Bibliothèque de L’Arsenal, MS 280, Archives Charmet, unpaginated. Though there are formal similarities between the positions of St Francis and his hands in relation to the crucifix in these illuminations and Van Eyck’s composition, I am uncertain whether Van Eyck had exposure to such illustrations. Luber argued that van Eyck based his representation on textual accounts of St Francis’s life, or depended on the verbal instructions of his patrons rather than visual sources. Luber 1998, p. 31. The representation of St Francis that Van Eyck’s figure is most often contrasted with is in the fresco by Giotto in the Bardi Chapel, Santa Croce, Florence. See Snyder 1997, pp. 78- 84. Lack of formal agreement between the pictures by Van Eyck and Giotto is consistent with an absence of corroborating historical circumstances. Disparate settings, styles and poses indicate that there was no correspondence between these two works. An Italian picture closer to Van Eyck’s composition is the fresco of Francis praying before the cross of San Damiano, c.1300, Upper Church, San Francesco, Assisi, Italy. Though van Eyck may have travelled through Italy in his service to Duke Philip the Good, there is no documentation of his itinerary and no evidence that he saw either Giotto’s fresco or the San Damiano image. See Derbes, Neff 2004, pp. 448-451, fig. 14.1. On the representation of St Francis in Italian painting see Van Os 1974, pp. 115-132. For a discussion of Van Eyck’s travels see Paviot in this volume. 10 Boucicaut Master, St Francis Receiving the Stigmata, Paris, Musée Jacquemart-André, MS 2 Boucicaut Hours, fol. 37v. See Snyder 1997, p. 79, 86 and n. 16. 11 Te Igitur page (detail), missal from a church in Bruges, fifteenth century. Bruges, Openbare Bibliotheek Brugge, MS 314, fol. 88r. Parchment, overall 31.1 ≈ 22.3 cm. De Poorter 1934, pp. 351352 and Brussels 1959 exh. cat. no. 99, p. 104. 12 Te Igitur page (detail), missal of Ter Doest Abbey, 1478-1492. Bruges, Grootseminarie te Brugge MS 49-18, fol. 201r. 13 Earlier examples of similar Te Igitur folio illuminations include: The Hague, Koninklijke Bibliotheek MS 78D 40, fol. 64r (Festal Missal, made in Amiens in 1323, Petrus de Raimbaucourt, illuminator); The Hague, Koninklijke Bibliotheek MS 76D 14, fol. 156r (Missal, use of Utrecht, made in Utrecht, c.1425-1450). In both paintings, the capital ‘T’ of the Te Igitur text contains a scene of a priest celebrating mass standing before an altar, positioned in right profile, holding his hands parallel before him and looking upward. Celebrants kneeling before altars also appear in contemporary pictures of saints, such as the murder of St Thomas Becket illustrated in a full page miniature in a book of hours painted in Bruges, c.1450: Cambridge, Cambridge University Library, MS Ii.6.14, fol. 22v. See Cambridge 1993; cat. no. 39, 126-127, figs 108, 127. A prayer book made in Bruges, c.1390-1400, now in a private collection in England, features Becket similarly in a kneeling posture. See Smeyers et al 1993, pp. 90-92, fig. p. 91. Fifteenth-century Netherlandish illuminators also depicted Gregory the Great kneeling before an altar in pictures of the Mass of St Gregory, as is evident in a full-page miniature in a book of hours made in Utrecht, c.1455-1460. The Hague, Koninklijke Bibliotheek, MS 135 E 40, fol. 110v. The pope is similarly posed in a miniature of the Mass of St Gregory witnessed by Philip the Good.

385

The illustration was added at the behest of Philip the Good to the Hours of Philip the Bold Duke of Burgundy, c.1370 (Paris). Cambridge, University of Cambridge, Fitzwilliam Museum, MS 3-1954, fol. 253v. Textual evidence suggests that miniatures of kneeling celebrants reflect contemporary liturgical practice, as priests genuflected before and after raising the consecrated Host. In Western Europe the custom of kneeling during the mass developed during the Middle Ages and was formalized at the beginning of the sixteenth century in missal texts specifying genuflections by the priest before and after raising the consecrated Host and chalice. See Thurston 1916, pp. 441, 546. Francis’s kneeling posture in Van Eyck’s picture would have elicited in many contemporary viewers visual and proprioceptive associations with the Eucharistic rite. Kneeling was also practiced by the laity. An anonymous Middle Dutch tract on the symbolism of the liturgy instructs worshipers to kneel during the Collect, Die collecten leest hi dan; So soude vrouwe ende man Alle vallen op hare knien (‘He [the priest] then reads the collect; Thus women and men all should fall upon their knees’). Oudemans 1852, ll. 360-361. Worshipers kneeling before the altar in the Ghent Altarpiece may reflect this custom. The convention is also suggested in Jacob van Maerlant’s account of St Francis’s pet lamb kneeling at mass when the Host was consecrated. Van Maerlant 1954, ll. 4393-4402. 14 See Van Buren, Marrow, Pettenati 1994-1996, passim; and Smeyers 1997, pp. 64-74. 15 See Stirnemann, Chavannes-Mazel and Dwarswaard in this volume, which elaborates on the detailed nature of these text images. 16 See for example, Bruges, Openbare Bibliotheek Brugge, MS 314, fols 87v and 88r; Bruges, Grootseminarie te Brugge, MS 48-3, fols 174v and 175r. Fifteenth-century Flemish missals recall Ottonian sacramentaries displaying an orant priest at the foot of a Tau cross, such as those made at Fulda and Tours. See Calkins 1986, pp. 17- 23; McKinnon 1978, pp. 21-52; and McLachlan 1978, pp. 27-35. 17 For a discussion of St Francis’s particular devotion to the presence of Christ in the Eucharist and his personal relationship to the performance of the mass see Cunningham 2006, pp. 56-71. 18 The Franciscan theologian John Pecham (d. 1292), who taught at Paris and Oxford and served as archbishop of Canterbury, wrote that Christ was locally present in two places simultaneously during miraculous appearances. Burr 1984, p. 44. The Mass of St Gregory is another instance of Christ’s miraculous appearance after his ascension. As noted above, Van Eyck’s composition also recalls full page illuminations of the Mass of St Gregory in Flemish books of hours. 19 Bonaventure postulated that Christ’s simultaneous presence in heaven and at diverse altars occurs by miracle in accordance with the truth of the sacrament. He also asserted that the process of conversion results in the alteration of the Eucharistic species, not in Christ. Burr 1984, p. 9. Smeyers argued that ‘Master I’ based the crucifix in a miniature of St Thomas Aquinas in his study on Van Eyck’s St Francis receiving the stigmata. Smeyers 1997, pp. 64-74. 20 Burr 1984, p. 10. 21 Van Delf 1937, LV, 436. 22 Oudemans 1852, ll. 844-85. The Benedictine, Pashcasius Radbertus, asserted in an influential ninth-century treatise that consecration symbolically re-enacts the crucifixion. See Smalley 1952, p. 91. 23 Snyder 1997 pp. 77-78; Van Os 1974, pp. 115-132. 24 Derivations and copies of Van Eyck’s St Francis Receiving the Stigmata are listed below by region of origin. Germany: a fifteenthcentury version by an anonymous artist is in the Historisches Museum, Bamberg. See Madrid/Valencia 2001, pp. 259, fig. 28 B. Italy: Sandro Botticelli, Adoration of the Magi (1470-1475) London: National Gallery. See Panhans 1974, pp. 188-198; Spantigati 1997, pp. 21-22; Butler 1997, p. 39; Jolly 1998, pp. 369-394. Similarities with the landscape have also been noted between paintings by Filippino Lippi and Domenico del Ghirlandaio, and drawings by Leonardo da Vinci. See Gombrich 1976, pp. 33-34, fig. 77; and Bruges 2002, exh. cat., no. 28, p. 236. The Netherlands: attributed to the Master of Hoogstraeten (active c.1485-1500), The Stigmatization of St Francis, Museo del Prado,

386

jamie l. smith

Madrid. Bruges 2002, exh. cat., no. 45, p. 242. A copy after Van Eyck by an anonymous painter in Brussels, c.1500, was last recorded in 1928 at an Amsterdam sale of the collection of Amédée Prouvost of Robaix. Rishel 1997, pp. 8-9. A right wing of an altarpiece attributed to Jacques Daret (Tournai c.1403-active until 1468) showing St Francis receiving the Stigmata sold at Koller Auctions, Old Master Paintings Sale, 21 September 2012, Zurich, Switzerland. I am grateful to Til-Holger Borchert for kindly bringing this auction record to my attention. Spain: Master of the Porziuncola, Stigmatization of St Francis, c.1470, Capuchin convent, Castellón. The 1448 will of the painter Juan Reixach itemizes an oil painting by ‘Johannes’ depicting St Francis receiving the stigmata. Bruges 2002, exh. cat., no. 121, p. 268. Reixach borrowed landscape elements from Van Eyck’s painting in his St Matthew for the predella of an altarpiece documented in 1446 in the hospital of València. Strehlke 2001, pp. 590-593. 25 Bruges, Openbare Bibliotheek Brugge, MS 408, fols 94v-95r. 26 Van Ruusbroec 1989, p. 140, ll. 327-329. 27 Anselm of Canterbury [Bec] 1956, p. 105. The Dunes library in Lissewege, near Bruges contained a thirteenth-century copy of this text in a codex of Anselm’s various writings (Bruges, Openbare Bibliotheek Brugge, MS 104, fols 1r-139r) De Poorter 1934, pp. 136, 137; Isaac 1984, pp. 67- 68. 28 Jan van Eyck, Crucifixion, Oil on canvas transferred from wood, c.1430, 56.5 ≈ 19.7 cm, New York, Metropolitan Museum of Art. The painting appears to be the left panel associated with the right panel showing the Last Judgement. On new evidence regarding these works and inscriptions they bear, see Ainsworth in this volume. 29 Bruges, Grootseminarie ter Brugge, MS 72-175, fol. 41v, ll. 260-261 and fol. 81r, ll. 3-4. 30 In previous papers, I argued that Van Eyck represented the wounded Christ as an exemplar for realist oil painting: Smith 2003, unpaginated; Smith 2008, pp 39-69, 154-177, 323-334; and Smith 2009, unpaginated. I presented evidence that Van Eyck modelled his artistic identity after Christ in the orthography of his motto and in the symbolism of his painted imagery. Smith 2012, pp. 273-312; and Smith 2014, pp 60-67. 31 Petrus Christus, Christ as the Man of Sorrows, oil on wood, 11.2 ≈ 8.5 cm, Birmingham, Birmingham Art Gallery, acc. no. 1935P306. See Ainsworth 1995, exh. cat., no. 9, pp. 112-116; and Upton 1990, pp. 55-56, fig. 53. 32 Camille interpreted Meister Francke’s Man of Sorrows as a self-portrait of the artist. Camille 1998, pp. 183-210. On Christ as an active Eucharistic celebrant see Lane 1984, p. 313. 33 The 1504 Tanners’ Guild chapel inventory contains three entries for green curtains: ij gronenne gourdinen (‘2 green curtains’); nocht ij groenen gourdinen met een rabbart voir onse vrauwe (‘2 more green curtains with a valance in front of [an image of] Our Lady’); nocht ij gronen gourdinen, voir Saincte Bavve (‘two more green curtains in front of [an image of] St Bavo’). Bruges, Rijksarchief te Brugge, Bruges Tanners’ Guild chapel register, 1504, fol. 108r. The 1479 Bruges Tanners’ Guild inventory lists j scoen tafel met den Cronemente van onser Vrauwen, der voren hanghende twe voe cordinen (‘one beautiful panel painting with the Coronation of the Virgin, before which hang two silk curtains’); tafel…ghedect met eenen rode cleede’ (‘panel…decked with a red cloth’); and tafelkin ghedect met ij blau cordinkins (‘small panel decked with 2 small blue curtains’). Bruges, Rijksarchief te Brugge, Bruges Tanners’ Guild chapel register, 1479, fol. 8r. 34 Bruges, Stadsarchief Brugge, Gilde Droogenboom, no. 505, portfolio 2, Ledenlijst van de gilde, fol. 1r. 35 Bruges, Grootseminarie te Brugge, MS 49-18, fol. 201r; Bruges, Grootseminarie te Brugge, MS 48-3, fol. 175r; and Bruges, Openbare Bibliotheek Brugge, MS 314, fol. 88r. 36 Snyder 1997, pp. 84-85. All biblical passages referred to in this paper are from the Douay-Rheims version of the Bible. 37 The Hague, Koninklijke Bibliotheek, 135 E 40, Book of Hours, Utrecht, c.1455-1460, fol. 27v.

38 Augustine 1888, p. 2. 39 The Hague, Koninklijke Bibliotheek, RMMW, 10 A 15 fol. 32v. Biblia pauperum. Hesdin or Amiens, Rambures Master (illuminator); c.1470; and Oxford, The Bodleian Library, University of Oxford, Arch. G c.14; Blockbook, Germany, c.1470. 40 Bé concluded, ‘It is possible that the strata belong to the area that geologists call the Paris Basin, which covers most of France and southern Flanders. It should be emphasized that the outcrop is surely not representative of the Sasso di Stimmate at La Verna, the traditional site of St Francis’s stigmatization.’ Bé 1997, pp. 88-95, citation p. 90. Bé refuted Sterling’s claim that Van Eyck depicted the limestone cliffs of La Verna. Sterling 1976, p. 29. Bé’s observations discredit Weale’s assertion that the artist painted both pictures on his diplomatic mission in Spain. Weale argued this position on the basis of Francis’s brown or tan-gray habit, the exotic palmetto plant behind him, and the background landscape which the author claimed displays Montserrat and the Pyrenees. Weale 1886a, p. 7; Weale 1886b, p. 269; and Weale 1903, p. 15. Friedländer rejected Weale’s hypothesis of an Iberian origin arguing that the Adornes will proves both paintings were in Bruges in 1470. Friedländer 1924, p. 64. Snyder also refuted Weale, citing evidence that Franciscan friars wore various coloured garments charitably donated to them and noting brown was not the Order’s official colour until the late seventeenth century. Snyder 1997, p. 87, n. 22. 41 The significance of the stones is reinforced in the Turin panel where traces of a now illegible inscription are visible on the rock to the right of the seraph-Christ. Spantigati 1997, p. 19, fig. 18. 42 Van Maerlant identified cerauneus as a donresteen and stated that this type of stone fell down with lightning from the sky. Van Maerlant 1981, XII, 417. Van Maerlant stated that these stones are found in Germany, Spain and the Netherlands. Interestingly, it is these regions that Van Eyck made pictorial references to in his paintings of St Francis in the form of Alpine mountains, Flemish architecture and Iberian foliage. It was also in these regions that the pictures were later copied and imitated. Van Maerlant also used the term donresteen in the Spiegel historiael to describe what seems to have been a meteor that fell in Rome during the reign of Pope John XII. Van Maerlant 1863, XVII, 1-15. Also see Verwijs, Verdam 1882-1952, entry: Dondersteen. 43 Pliny the Elder 1855, Book II, chapters 59, 58. 44 Like several of Van Eyck’s panels, the reverse of the Turin panel is painted in a marbleized pattern that simulates stone. This simulation resonates with the rocks ‘thrown by the gods’ depicted on the face of the panel. The parallel may have been conceived to align Van Eyck’s painting with a class of objects believed to have been created through divine agency. On miraculously created images of Christ see Kessler 2002, pp. 11-12; and Kessler 2000, pp. 15-17, 70-76. The Holy Face of Laon is a famous icon painting of the Mandylion of Edessa located during the late Middle Ages in the Cistercian convent of Mantreuil-en-Thiérache, near the border of France and the Netherlands. See New York 2004, exh. cat., no. 95, pp.174-175. The relic may have had an audience and following around Bruges. A fifteenthcentury chronicler of the Dunes Abbey recorded in Supplementum Cronice Abbatum de Dunis that an image on cloth of the face of our Lord Jesus Christ was brought to the abbey for the inauguration of the institution on 3 October 1262. Ainsworth has identified this icon with The Holy Face of Laon. Ainsworth 2004, pp. 560-561. Dhanens observed that a neckless Holy Face is depicted on a singing angel’s vestments in the Ghent Altarpiece. Dhanens 1980, p. 110, fig. 70. This image suggests that Van Eyck was familiar with the Laon relic. 45 Runciman 1931, p. 238-252. Works by Van Eyck copied in his workshop during his lifetime, St Francis Receiving the Stigmata, Holy Face and the Virgin by a Fountain, all contain images of Christ. The Holy Face and the Virgin by a Fountain are both derived from formats of legendary acheiropoieta, the Mandylion and the portrait of the Virgin painted by St Luke. See Kessler, Zacharias, 2000, pp. 60-63, 90-93, fig. 55. 46 Van Delf 1938, VI, 87.

replications of exemplary form

47 See Van Maerlant 1981, lines 15,459-15,471. Redonensis 1893, XXVIII, cols. 1756B - 1757A. 48 Jerome 1892, p. 61. 49 The Adornes were a family of prominent merchants of Genoese and Flemish descent who had lived in Bruges since the late fourteenth century. On the history of the family’s private chapel, the Jerusalem Chapel, its archives and inventory, see Bruges 1983, exh. cat., passim; Geirnaert 1987, pp. 5-12, 112-172. Geirnaert 1989a, pp. 3-11, 26, 28-30, 32-39, 49, 50, 53, 54, 59, 63, 66, 74, 82, 103, 105, 106, 127, 128, 130, 132, 138, 140 and 145. On the Kalverieberg Altar see Devliegher 1965, p. 47; Vermeersch, 1981, p. 142; and Bruges 1983, exh. cat., no. 12, pp. 72, 104-105. 50 The impact of the wondrous properties ascribed to fossils may be indicated, in part, by the occurrence of fossil images in Netherlandish painting. Whereas I do not know of any representations of such stones before c.1425-1430, fossils appear in a number of works painted thereafter. Fossils appear in two works contemporaneous with Van Eyck’s paintings: Robert Campin’s Nativity, Musée des Beaux-Arts, Dijon and the left wing of The Annunciation, The Metropolitan Museum of Art, The Cloisters Collection, New York. A later work containing fossils is Petrus Christus’s Nativity, National Gallery of Art, Washington. See Bé 1997, pp. 88-95. 51 I am grateful to Noël Geirnaert for verifying the dates of the papal documents pertaining to the foundation of the Jerusalem Chapel. The first papal bull concerning the chapel was granted by Martin V on 12 May 1427. The text of this bull is preserved in two copies. Geirnaert 1989a, p. 26. The bull of 13 July 1435 granted by Pope Eugene  IV concerning the chapel is the earliest original bull regarding the chapel kept in the Stadsarchief Brugge. Bruges 1983, exh. cat., no. 6, p. 100. Rishel confused the date of this papal bull with the date of one issued by Pope Eugene IV on 13 July 1435 concerning the building of almshouses near the chapel and, therefore, erroneously reported the date construction began on the chapel as 1435 instead of 1427-1428. See Rishel 1997, pp. 7-8. The date of the bull granting permission to build the chapel was also mistakenly reported as 1435 in New York 2004, exh. cat., p. 548. On the Adornes’s relic of Christ’s tomb see Esther 1983, pp. 68-73. 52 Geirnaert 1983, pp. 15-16. 53 The contents of the original document of Anselm’s will, now lost, are known from a sixteenth-century copy. Though currently missing, the contents of the copy are known from transcriptions and a

387

photographic reproduction. Aru, de Geradon 1952, p. XIX, Rishel 1997, p. 7, fig. 6. On the will and its mention of two paintings of St Francis see Pinchart 1860, p. 267; Hymans 1883, pp. 108-116; Weale 1886, p. 73; Mayer 1926, p. 200; De Poorter 1931, pp. 225-239; Weiss 1956, p. 6; Faggin 1968, pp. 88-89; Panofsky 1971, pp. 192, nn. 1, 300, 312 and 432; Geirnaert 1987, no. 728-730, p. 114; Geirnaert 1989a, no.  195, pp. 74-75; Rishel 1997, pp. 4-11, n. 10; Van Asperen de Boer 1997, p. 56; Geirnaert 1998, pp. 40-45; and Geirnaert 2000, pp. 165-170. 54 Esther argued that the arms of Anselm Adornes’s wife, Maria van der Banck, which were polychromed on shields carved in relief, are contemporaneous with the altar’s creation. Esther 1983, pp. 104-105. Devliegher 1965, p. 47; and Vermeersch 1981, p. 142. 55 Guardianship was granted by Pope Clement VI in 1312. Rishel noted that Anselm Adornes requested that his pallbearers be Franciscan friars. Rishel 1997, p. 8. 56 Various identities have been suggested for the sitter, including Anselm Adornes, his father, Pieter Adornes, an unidentified Spaniard and Nicolas Rolin. See Weale 1886, p. 269; Mayer 1926, p. 200; Goldschmidt 1930, no. 132, p. 50; Denis 1964, p. 43; Spantigati 1997, p. 22; Snyder 1997, p. 86, n. 7. On the question of whether St Francis is shown in Van Eyck’s pictures with a tonsure, Spantigati asserted ‘… the restoration has removed the repainting that had given St Francis a tonsure nonexistent at the time.’ Spantigati 1997, p, 22. But, referring to the Philadelphia painting, Van Asperen de Boer wrote, ‘The hair seen in the reflectograms as a few thickish brush strokes leaves no doubt about the intended tonsure. Before the restoration in 1982 this was also visible on the painting, but it is less obvious now, as the present appearance of the top of the head is a little confusing.’ Van Asperen de Boer 1997, pp. 53-54. 57 I am especially appreciative to Noël Geirnaert for assisting me in accessing documentation of Jan Beert’s life and his relationship with the Adornes family and the Jerusalem Chapel. See Derolez 1966, pp. 1-5; Geirnaert 1989a, no. 62, 65, 75, 111, 115 and 128. pp. 33, 35, 38, 48, 49 and 53; and Geirnaert 1989b, pp. 313-321. 58 ‘Let Your presence wholly inflame me, consume and transform me into Yourself, that I may become one spirit with You by the grace of inward union.’ Kempis 1949, IV, 16. 59 Rishel 1997, pp. 4-11, n. 10; Van Asperen de Boer 1997, p. 56; Geirnaert 1998, pp. 40-45; and Geirnaert 2000, pp. 165-170.

Fig. 26.1 Jan van Eyck, Virgin and Child (Lucca Madonna), c.1435, oil on oak panel, 65.7 × 49.6 cm, Frankfurt, Städel Museum (inv. no. 944)

26

Questioning the Technical Paradigm of the Ghent Altarpiece Noëlle L.W. Streeton

ABSTRACT: The quantity and quality of analytical data from the Ghent Altarpiece, and especially information derived from paint cross sections, is difficult to compare with much more limited results from other, mostly smaller and later works. This imbalance of evidence has had implications for understanding the results of other studies, primarily because deviations from the technical standard of the Ghent Altarpiece have not been adequately acknowledged. A recent study has addressed this imbalance through the investigation of blue, green, red and grisaille passages in a range of paintings attributed to Jan van Eyck and his contemporaries. In total, the results suggest that Eyckian methods varied more than can be divined from paintings thought to pre-date 1435. This paper considers current descriptions of an ‘Eyckian’ technique, as well as a few case studies that challenge these descriptions. The article concludes with proposals for further enquiries that will help to clarify the significance of differences between the Ghent panels and later paintings.

—o— Introduction Few works by any painter have been examined so comprehensively and sampled as extensively as the Ghent Altarpiece. Since August 1945, when the dismantled polyptych was returned to Belgium from the Austrian mines at Altaussee, it has been the subject of at least seven investigative campaigns, the majority based in Brussels.1 The studies led by Paul Coremans (1940s-1950s), J.R.J van Asperen de Boer (1960s-1970s) and Pim Brinkman (1980s) have become significant points of reference for conservators and art historians wishing to

describe the techniques identified in the Ghent Altarpiece. The contributing researchers have arguably come closer than others to an elusive definition of techniques that could be considered as typical for Jan van Eyck. Perhaps for this reason – and despite the thorny issues that continue to surround this particular corpus of works – few have questioned whether their understanding of Eyckian technique, gained through the Ghent Altarpiece results, should be deployed to describe the techniques found in other Eyckian paintings. This paper considers the extent to which the techniques and materials found in the interior panels of the Ghent Altarpiece can be considered to be typical or paradigmatic for Van Eyck. After a brief introduction to the ways in which an Eyckian technique has come to be described, the focus turns to a limited number of results from a study that has questioned this technical paradigm. The aim is to characterize the observable features in the works thought to have been produced in Van Eyck’s Bruges workshop and ultimately to consider the significance of their differences.2 Describing an Eyckian Technique Details of past and ongoing scientific studies are presented elsewhere in this volume.3 However, descriptions of the complexity of Van Eyck’s technique are relevant here and for this reason a few points will be reiterated.

390

noëlle l.w. streeton

For those engaged in studies of the Van Eyck corpus, the 1950s research of Paul Coremans, John Gettens and Jean Thissen is legendary. Their most acclaimed finding was of course the characterization of an Eyckian painting medium, which they described as ‘a base of drying oil’ with the addition of ‘x’, a component thought to be a resin that was added selectively to green and red glazes to heighten colour saturation yet further.4 This declaration has shaped the course of paintings research for more than sixty years, inspiring many to pursue related enquiries that continue to the present day. However, what captured the imaginations of so many, and what still ensures that L’Agneau Mystique au laboratoire is required reading for anyone interested in this field, is that Coremans, Gettens and Thissen went so much further. They described in painstaking detail the complex stratigraphy of jewel-tone colours, both in the Ghent panels and in Dieric Bouts’s Altarpiece of the Holy Sacrament. With this they offered a way to understand a paint structure behind Van Eyck’s extraordinary verisimilitude. They provided exhaustive descriptions, both visual and verbal, of the structures behind his translucent, or indeed transparent, optical effects; the structures behind his motifs for rendering jewelencrusted brocades and matt woollen fabrics; his methods for rendering reflections on water, metallic surfaces and heavenly figures in leafy landscapes. In short, they described what has come to be understood as Eyckian painting technique.5 Since the 1950s, the information on Van Eyck’s painting materials and methods has been significantly implemented, especially by Leopold Kockaert, Van Asperen de Boer, Brinkman, and those working in Brinkman’s team.6 All have emphasized, like Coremans, Gettens and Thissen, how Van Eyck created the Ghent images in multiple, thin and systematic layers, to achieve finely structured, smooth and densely coloured surfaces. Still today, the transliterated term ‘Eyckian technique’ refers to the systematic construction of exquisite colour through the build-up of superimposed layers of paint, beginning with an opaque base and adding

ever darker, more saturated and more translucent colour of the same hue. In this way, Van Eyck and his similarly-trained contemporaries achieved translucent or transparent optical effects in blue, green and red passages. This description is a very satisfying and eloquent way to qualify, and even quantify, a paint structure behind the imagery in works associated with Van Eyck. It has offered a way to understand what evidently transfixed Van Eyck’s patrons, as well as other painters, collectors, curators and far more recently, conservators and conservation scientists. The description also provides a model for understanding the translucent and transparent optical effects within Van Eyck’s arsenal of motifs. Arguably for these reasons – and because the quantity and quality of data for the Ghent panels is so extraordinary – it is the layering technique identified with the Ghent interior panels that has become the bedrock, standard or paradigm, for comparisons with earlier, contemporary and later works. However, it is important to note that the interior panels have been consistently discussed in much greater detail than the grisaille passages on the exterior wings, with an understandable emphasis on exploring Van Eyck’s manipulation of oilbased jewel tones.7 Thus, it might be said that the analytical results for the altarpiece’s interior have been privileged. Conversely, very little emphasis has been placed on single layers of paint, which appear to have been limited to opaque passages, such as strokes of black and those based on lead white, which characterize the grisaille images and the greyish/sandstone-coloured paint used to depict interior scenes. The physicality of these passages rarely features in discussions of techniques that were typical for Van Eyck, and for other early Netherlandish painters. While there have been numerous advantages to past and ongoing approaches to the often complex stratigraphy of Van Eyck’s rich colours, the tenacious idea of a technical standard has complicated or even impeded efforts to characterize works that deviate or differ from it. The grisaille images on

391

questioning the technical paradigm of the ghent altarpiece

the exterior wings of the Ghent Altarpiece are in many ways distinctive from those found on the opposite faces and cannot be described in ways that are compatible with the now classic descriptions for landscapes and draperies formed of multiple layers of paint. A tight definition of Eyckian technique is also difficult to apply to the Crucifixion and Last Judgement in New York, the Virgin and Child in Frankfurt (known as the Lucca Madonna), the Virgin and Child with Saints Catherine and Michael and a Donor (also referred to as the Dresden Triptych or Giustiniani Triptych), the Berlin Virgin and Child in a Church, both versions of the Virgin by a Fountain (Antwerp and private collection), both surviving versions of St Francis Receiving the Stigmata (Turin and Philadelphia), as well as the arguably unfin-

ished St Barbara in Antwerp (see Table 1). Given the breadth of the study from which this paper is drawn, it is not possible or desirable to offer details from each of the aforementioned works. Instead this paper will plough a narrow path, focusing on the Ghent Annunciation panels and the two grisaille St John panels (see plates A and B), the Lucca Madonna (fig. 26.1) and the Philadelphia version of St Francis Receiving the Stigmata (fig. 26.2). Methods and Challenges With regard to methods, perhaps this is stating the obvious. The currently available data from, among others, the Lucca Madonna and the two surviving St Francis paintings – and especially those data drawn from paint cross sections – bear no resem-

Table 1 Paintings attributed to van Eyck and his workshop Jan and Hubert van Eyck, Ghent Altarpiece, dedicated 6 May 1432 Exterior: Annunciation and St John panels

Cross sections 11 (of c.300)

Jan van Eyck and assistant, Crucifixion and Last Judgement, c.1425–1441, New York Jan van Eyck, Virgin and Child (Lucca Madonna), c.1435, Frankfurt Jan van Eyck, Virgin and Child with Saints Catherine, Michael and a Donor, Dresden

Dendro

IR/ IRR

X-ray

Other *

x

x

x

x

transfer

x

x

x

x

x

x

x

1 (frame)

Jan van Eyck, St Barbara, 1437, Antwerp Jan van Eyck, Virgin and Child in a Church, c.1437-38, Berlin

x 2

Jan van Eyck, St Francis Receiving the Stigmata, c.1440, Turin Jan van Eyck and/or workshop, St Francis Receiving the Stigmata, c.1440-41, Philadelphia

x

x

x

x

x

x

x

x

2 (frame)

x

x

x

x

c.13

x

x

5

x

x

x

x

Jan van Eyck, Virgin by a Fountain, 1439, Antwerp Jan van Eyck and/or workshop, Virgin by a Fountain, c.1439-40, private collection

x

* ‘Other’ refers to other analytical techniques (e.g. XRF, UV-Vis, SEM-EDX, GC/MS and FTIR)

392

noëlle l.w. streeton

Fig. 26.2 Jan van Eyck and/or workshop, St Francis Receiving the Stigmata, c.1440-1441, oil on vellum mounted on oak panel, 12.4 × 14.6 cm, Philadelphia, Philadelphia Museum of Art, John G. Johnson Collection (acc. no. 314)

blance to the quantity and quality of data for the Ghent Altarpiece. My investigations therefore necessarily relied on the paper trail left by conservators in the conservation dossiers for these works, as well as on limited new analytical materials and direct observation of the paintings themselves. As outlined in Table 1, the conservation dossiers include a small number of cross sections. In addition, each of the supports has been the subject of dendrochronological dating. All have been examined with infrared reflectography and X-radiography, and all have been studied in other ways too, such as with

X-ray fluorescence (xrf). The painted surfaces were examined systematically under magnification, both before and after studying the available conservation and scientific data. Paint reconstructions were a complementary tool for understanding better, among other things, the build-up of densely layered structures and the impact of ageing and/or discolouration. Finally, the dossiers and paintings were discussed with the curators and conservators currently responsible for these works. Predictably, though, as I assimilated the information from dossiers, examinations and recon-

questioning the technical paradigm of the ghent altarpiece

structions, I came up against a problem. As one would expect, disparate archives contain diverse and often incomparable sets of information. Thus, the sparse information from smaller and mostly later works could not be reconciled with the far larger and far more comprehensive archive for the Ghent Altarpiece. The unparalleled campaigns of imaging, sampling and analysis have provided unique access to the physical make-up of this masterpiece, and while a number of museums that hold other paintings have commissioned a range of independent studies (most since 1990) the analytical information from these varies widely and currently does not offer bases for direct comparisons – particularly among works thought to post-date 1435. Despite these methodological challenges, the information that was available coupled with my own examinations could provide a point of entry for considering the order and method of paint application, as well as the nature of the paint itself. This way of looking also led me to consider something that is fundamental: that those paintings produced on a far smaller scale than the Ghent panels should logically embody techniques appropriate to their size and function. In light of this study, which remains decidedly a work in progress, it is clear that there are differences, both minor and significant, between the Ghent panels and other paintings produced over the course of Van Eyck’s career. Developing Techniques Van Eyck’s period of known activity is far shorter than later artists such as Gerard David (c.14601523), Lucas Cranach the Elder (c.1472-1553), Peter Paul Rubens (1577-1640) and Rembrandt (1606-1699). But even so, it is permissible to envisage that Van Eyck, with the assistance of his workshop, altered his practices in the course of a decade. This is not to say that his practices ‘evolved’. Given that the surviving works are but a fraction of the workshop’s original output, issues of an evolving workshop practice cannot be brought to bear in discussions of Van Eyck. Nevertheless, it

393

seems that Van Eyck developed methods that were appropriate to his changing iconographic programmes. These are not novel thoughts. Alan Burroughs, in Art Criticism from a Laboratory, suggested in the 1930s that Van Eyck’s glazing technique had not been fully developed or ‘perfected’ (with all that this implies) by 1432 when the Ghent Altarpiece was dedicated.8 Burroughs’ hypothesis was, however, overshadowed by the release of the results of Coremans and his team in 1952, and since then, those following in Coremans’ footsteps have not focused on the probability that Van Eyck’s technique was far from static. This is interesting if only because roughly one third of the paintings attributed to Van Eyck and his Bruges workshop appear to deviate, to varying degrees, from the technical standard. While taking stock of deviations is perhaps not vitally important for dating or attribution, it is important for interpreting the potential motivations of the painter and the ways in which the Bruges workshop functioned, especially in the latter part of the 1430s. The remainder of this paper will focus on a limited number of results from a study that commenced in 2005, a part of which involved the registration of observable features in a range of Early Netherlandish paintings.9 Grey Tones The Annunciation and St John panels fit few of the criteria for the complexity that is so strongly associated with rich colour, textures and glazes. Instead, the greyish/sandstone-coloured paint is in many areas formed of a single layer, often using the underdrawing to enhance and deepen the tone. Of the eleven cross sections from these images, nine were taken during Coremans’s first sampling campaign from a highlight, shadow or mid-tone in areas that represent figures, their chamber, or stone.10 The cross sections show that these passages are consistently based on reasonably uniform mixtures of lead white, with varying amounts of carbon black and finely ground red and yellow ochres. These

394

noëlle l.w. streeton

sampled passages are technically similar to those, for example, found in the Madrid Annunciation.11 They are also closely comparable to passages depicting domestic chambers and church interiors, which characterize Van Eyck’s paintings produced throughout the 1430s. The chamber in the Lucca Madonna, the chapel in the Dresden Triptych and on a far larger scale, the setting for the Virgin and Child with Canon Joris van der Paele in Bruges, have architectural features apparently formed from a simple mixture, based on lead white, and judging from those that have been sampled, often laid on in a single layer in midtones and shadows. Given the preoccupation with jewel-tones in Eyckian paintings, it is not difficult to argue that the usefulness of lead white to Van Eyck’s workshop has been eclipsed by a greater appreciation for exotic and more expensive colours. But in this regard, too, there were uses of dearer colours beyond those that fit neatly into the existing technical model for Van Eyck. Ultramarine as a Finishing Pigment Discussions of colour in relation to Van Eyck often begin with his uses of ultramarine. I will do the same, but only to make one small point. For smallscale paintings, such as the two surviving versions of St Francis Receiving the Stigmata, the two versions of the Virgin by a Fountain and the Dresden Triptych, surface effects that required ultramarine were evidently tailored to images intended for intimate viewing. Like Vermeer, Van Eyck employed ultramarine of varying coarseness for a variety of surface effects in paintings designed for close inspection.12 For example, details that define floor tiles and the baldacchino of the Lucca Madonna were produced with ultramarine mixed with lead white, apparently in a single layer. This stands in contrast to the complex structures identified with the blue robe of the Ghent Virgin Enthroned, or the sky in the Adoration of the Lamb. In the Lucca Madonna, and many other Early Netherlandish works, such details provided a tonal quality that could only be achieved with this

particular pigment, as well as a level of finish to engage the viewer with an image that was optically pleasing, intrinsically valuable and undoubtedly calculated to heighten the experience of the divine. Similarly, stippled shadows in the face and fingertips of St Francis in both the Philadelphia and Turin versions of St Francis Receiving the Stigmata offer examples of the ways that Van Eyck and/or his workshop used discreet touches of ultramarine (fig. 26.3). In the Philadelphia painting, under low magnification, it is apparent that reasonably coarsely ground ultramarine, mixed with a flesh tone, was stippled to define eye sockets and the fingertips of St Francis and Brother Leo. And on the upper lip and chin of St Francis, it seems that particles (perhaps not bound in medium) were applied with the tip of a fine brush. A paint reconstruction of this passage indicated that particles could be picked up on a fine sable brush and adhered to the surface while the flesh tone was wet to produce the stippled effect of a ‘five o’clock shadow’. This way of cooling the skin tone and creating shadows across the faces and fingers must have met or exceeded the patron’s expectations for an image that was precious both inherently and spiritually. If the Philadelphia image was not painted by Van Eyck but instead by a member of the workshop, it attests to the master’s ability to convey his grasp of the handling properties of mineral blues and especially ultramarine to those working in his immediate artistic circle. A Small-Scale Landscape Staying with St Francis but turning to passages based on verdigris, it is noteworthy that the two surviving versions in Turin and Philadelphia are most often discussed in terms of their landscape imagery. The elements depicting grasses and a recession into the distance in both paintings are, technically speaking, nearly identical to each other, although on differing scales and on different supports as well, with one on panel and the other on parchment adhered to panel. Nevertheless, the green passages have discoloured severely, to the extent that the original visual

questioning the technical paradigm of the ghent altarpiece

395

Figs 26.3a-d St Francis Receiving the Stigmata, Philadelphia. Microphotographs showing coarsely ground particles of ultramarine, stippled and dragged across surface or mixed with lead white and vermilion in skin tones: (a) ×50 blue in white of eye and shadows, St Francis; (b) ×25 whiskers, St Francis; (c) ×25 finger tips, St Francis; (d) ×18 finger tips, Brother Leo

impression is severely altered. According to available evidence for both versions,13 along with current thinking on the behaviour of verdigris-containing glazes,14 it appears that these passages did not retain their colour because they were, like the areas of sky, very thinly applied. The green passages were not layered thickly enough over each other, or over a mixture heavy in lead-tin yellow and/or lead white to render them less vulnerable to change. Even though isolated particles have retained their colour (figs 26.4a-b), with few if any layers beneath the transformation to brown has correspondingly transformed their appearances far more than paintings produced with the alternative systematic layering system (figs 26.4c-d).

Systematic layering, with more than one layer of copper green, over layers heavy in lead white and/or lead-tin yellow, is the hallmark of the landscape elements in the Ghent panels and the Rotterdam Three Marys at the Tomb.15 The green mantle in the Ghent St John Enthroned also owes the character of its colour to this ordered system. Given the prevalence of this method in Eyckian, contemporary and later paintings, it appears that it was a recognized means to preserve the colour and tame what was acknowledged to be the unpredictable behaviour of a copper green.16 Furthermore, and according to the recent studies of copper greens carried out by Margriet van Eikema Hommes and Klaas Jan van den Berg,17 a range of factors (such as

396

noëlle l.w. streeton

Figs 26.4a-d St Francis Receiving the Stigmata, Philadelphia, discoloured copper green of foliage. Microphotographs show the extent of discolouration, as well as isolated patches where green colour has been retained: (a) ×25 grass on rocky outcrop, left of the seraph; (b) ×25 tree left of St Francis; (c) ×25 foliage above the feet of the saint; (d) ×50 blade of grass with isolated green particles of verdigris

variable relative humidity and perhaps a reaction with the varnish) in combination with thin layers of copper-green glaze, meant that these passages were more vulnerable to browning. It is perhaps tempting to propose that one technique for handling copper green was traded for another over the course of the 1430s. However, such an assertion would not be accurate. The green robe of Micah in the niche above the Ghent Virgin Annunciate is also comprised of a thin coppergreen glaze, which has discoloured considerably. The green of the sleeve is painted over a much thicker layer of lead white with some black.18 A similar transformation to brown was identified in

the green glaze layer (20-40 μ) over a lead white and copper-green mixture in the Angel Annunciate’s wing. Both examples contrast considerably with the deep green multilayered vegetation in the Knights of Christ panel.19 With regard to the surviving St Francis paintings, though, the differences between their supports also raise a number of questions, most notably whether the Philadelphia version on parchment was initially intended as a study for Philip the Good’s famous terrestrial globe.20 In the late 1430s, the Van Eyck workshop was engaged in a project that married concerns for the behaviour of oil paint and the measurement and representation of

questioning the technical paradigm of the ghent altarpiece

distance on a semi-flexible material, namely parchment. It is probable that the combination of these particular materials (oil paint and parchment) to produce specific atmospheric effects was new to the painter.21 The project would therefore have required some experimentation before arriving at a solution, which thereafter was taken to completion (after December 1440) by the astronomer Guillaume Hobit, who was commissioned to produce the three-dimensional globe itself.22 For such a project Van Eyck would naturally have employed his detailed knowledge of the possibilities of oil-based paints to represent a visible landscape. But a new approach would have been required to imagine the landscape of the entire inhabited world, both on a microscopic scale and as it would appear on a sphere. The experimental design for a terrestrial globe and process of rationalizing vast spaces on the semi-flexible material of parchment are arguably reflected in the format, materials and paint structure of the tiny Philadelphia St Francis. A Scarlet Robe Like the study of the behaviour of copper greens there has been an equally intense focus on red passages in Eyckian paintings, especially in the characterization of glaze layers formed of red lakes. Numerous studies have demonstrated that to depict red matt woollen fabrics (those produced in Bruges were called scarlets) it was common to layer with mixtures of vermilion and lead white in untreated oil, followed by one or more glaze layers based on a madder and/or kermes lake in heat-bodied oil.23 This system was commonly employed to lend nuanced tonalities to red glazes. The intense scarlet woollen gown in the Lucca Madonna appears, on initial inspection, to be closely comparable to or the same as those found in red draperies in the Ghent Altarpiece and many other contemporary pictures. However, examinations of the Lucca painting in 2007 suggest that this Virgin’s scarlet robe should perhaps not be characterized in this way. The painting’s surface was examined first with reference to an X-ray, which shows the sweeping

397

brush marks in the first underlayer – possibly a layer of lead white in untreated oil. A second examination was later undertaken in coordination with Doris Oltrogge who took measurements with ultravioletvisible spectroscopy (UV-Vis).24 Both the clarity and mottled appearance of certain passages led us to question the role of opaque layers in this image – that is, whether this red mantle lacked the build-up of opaque, vermilion- and lead white-containing underlayers (fig. 26.5). For this reason, Oltrogge took a proportionally large number of measurements in and around a deep red fold of the Virgin’s robe, but none of the spectra registered any mercury.25 Less surprising was that the spectra were consistent with her database spectra for kermes and madder lakes. Vermilion and lead white evidently served both a visual and practical function and building underlayers with mixtures that included these pigments was undoubtedly more common than a paint film formed of untempered red glazes, one on top of another. The latter is more impractical in the absence of a metallic drier (e.g. lead white) primarily because medium-rich, untempered lake paints solidify layer-by-layer very slowly. Of course the UV-Vis results – with no indication of the presence of vermilion – might be put down to the depth of the glaze layers, and/or the shortcomings of UVVis spectroscopy. But if accurate, an intense red mantle based on a series of true glazes is not without precedent. A series of four red-lake glaze layers was identified in a thin section from the scarlet robe of Joos Vijd, published by Brinkmann and his team in 1988/1989.26 To these, carbon black (seen as feathery particles suspended in the glaze) was added to accentuate shadows, but unlike deep folds of red fabrics in the Arnolfini Portrait, the Washington Annunciation or the Portrait of Margaret van Eyck, no ultramarine has been found. The clarity of the paint film, as well as the mottled appearance of some passages point to the absence of both vermilion and lead white, and while it might be that each subsequent layer is based on a mixture of red lake pigment and a pre-polymerized oil, perhaps thinned with a

398

noëlle l.w. streeton

Fig. 26.5 Lucca Madonna, unframed in the Städel conservation department, November 2007. The microphotograph shows both the vibrancy and selective discoloration of the red lake glaze

solvent – it remains to be established whether the drying properties of metallic pigments were displaced with a material like a powdered glass to hasten the drying process. This would be similar to what Marika Spring has found in the red bedhanging in the Arnolfini Portrait and many other works.27 Finely ground glass used in a paint reconstruction prepared for this study showed that this additive aids the settling of the heat-bodied oil paint, hastens the drying process and decreases the likelihood of wrinkling and irregularities. This additive also allows for maximum transparency and vibrancy.

Lead White and Regimentation In stark contrast to the vibrancy of the Lucca Madonna’s robe, the walls of the chamber are formed of thin, opaque, reduced-colour tones, laid side-by-side in single layers. As mentioned earlier, this is consistent with other such chambers and grisaille images in the Van Eyck corpus. This effect (if not the structure) was clearly appreciated by later painters for whom workshop organization was a key to commercial success. For example, Bernard van Orley created a similar grisaille chamber in the Cambridge Annunciation (Cambridge, Fitzwilliam Museum); and Karel van Mander relied on a primuersel – or a coloured priming layer – to produce the backdrops in a number of paintings, such as for the grotto in his Adoration of the Shepherds (Haarlem, Frans Hals Museum).28 This was a technique that van Mander valued because it expedited the painting process. Significantly, too, Eyckian interiors placed limits on, or eliminated, the uncertainties associated with copper greens. Whether or not reduced-colour elements were underdrawn by the master or traced and modified from reproducible workshop drawings, the grey-tone elements themselves point to a formula and format that could have been applied consistently by a number of hands. This feasibly enabled Van Eyck to meet the demands of his role as servant to the duke, while still completing commissions for merchants and courtiers. This proposal runs counter to previous suggestions that Van Eyck had no commercial incentive to delegate tasks and therefore painted without the help of the assistants who are known to have been active in his workshop.29 However, to be clear, this proposal does not imply that members of the workshop would only possess the requisite skill to lay in a backdrop. There seems to be general agreement that at least one member of the workshop also produced a small grisaille diptych, which is now in Paris. This proposal also does not imply that a formula was simple, or that formulaic practices were readily established. The probable pentimenti in the back wall of the Washington Annunciation, as has

questioning the technical paradigm of the ghent altarpiece

been discussed by Melanie Gifford, point to experimentation before arriving at a viable solution.30 What is suggested here, though, is this: given the nature of Van Eyck’s interior scenes, there is no reason to insist that Van Eyck alone was capable of executing all elements of paintings that are attributed to him – especially passages that behave most predictably. Here we might also consider the contributions of others to underlayers in landscape and some red and blue passages, which were also built up in regimented ways. Where From Here? By questioning a technique that has been considered by many to be typical for Van Eyck, it has been possible to recognize ways in which a wide variety of methods were used in Van Eyck’s Bruges workshop. These are difficult to categorize in terms of the classic model for Eyckian technique, even though the treatment of greens and reds potentially identified in later paintings made their first appearance on the exterior wings of the Ghent Altarpiece. Further analyses are necessary to clarify these observations and to this end, I hope that the programme of analyses that has been suggested by Professors Anne van Grevenstein and Ron Spronk will eventually take place. This campaign would ideally be designed around the same mobile analytical instruments, so as to produce directly comparable results. Importantly too, such research should help to establish a better context for the Ghent Altarpiece in Van Eyck’s corpus, as well as a more balanced picture of Van Eyck’s techniques in the final years of his life. This line of enquiry, especially with regard to expedience and concerns for the drying of fresh paint, has been especially valuable for considering the potential activities in Van Eyck’s Bruges workshop in the second half of the 1430s. This was a period when intense volatility in the North Sea trading network changed the operational landscape for many craftsmen – and perhaps also for Van Eyck. At the very least, this study can help a range of future researchers to ask better questions and

399

while a socio-economic framework for these results remains underdeveloped, this study could eventually be seen in the context of the political volatility of the period when Van Eyck’s Bruges workshop was active. Commercial disruptions are the common thread running through historical accounts written primarily by economic historians, all of whom have recognized sharp transitions in the political economy of the Burgundian Netherlands after 1429. All also recognized the significance of Philip the Good’s quarrels with the English crown during the 1430s – transitions and tensions that have not been absorbed into the literature on Van Eyck’s painting technique. Against this backdrop, assumptions that the methods and materials identified in the monumental Ghent Altarpiece could be broadly representative of this painter’s working methods over time appear less valid. It is our good fortune that interactions between historians, curators, conservators and scientists have flourished since the work of Coremans commenced. In fact, the interdisciplinary study of paintings is now considered the norm. As the new Ghent Altarpiece project progresses, perhaps new, parallel, cross-disciplinary examinations will help to shed greater light on this magnificent altarpiece’s place alongside Van Eyck’s other surviving paintings. N OTES * In addition to the organizers of the Van Eyck Studies colloquium, thanks are due to those who have contributed to the development of this paper. These include Libby Sheldon and Clifford Price (University College London), James Bolton (Queen Mary, University of London), Jana Sanyova (KIK-IRPA), Jochen Sander (Frankfurt, Städel Museum,), Mark Tucker (Philadelphia Museum of Art), the Arts and Humanities Research Council of Great Britain and the Faculty of the Humanities, University of Oslo. 1 Coremans, Gittens, Thissen 1952; Coremans 1953; Coremans 1954; Kockaert 1973/74; Kockaert and Verrier 1978-1979; Van Asperen de Boer 1979; Brinkman et al. 1984-1985 and 1988-1989; Verougstraete, Van Schoute 1987; Vynckier 1999-2000. New investigations began in the spring of 2010 based in Ghent, initially as part of the project ‘Lasting Support’ (see http://closertovaneyck.kikirpa.be/; Van Grevenstein, Spronk 2011). ‘Lasting Support’ has been followed by a five-year campaign of conservation-restoration and research by the KIK-IRPA in Brussels, which commenced in September 2012. 2 This theme has been treated comprehensively in Streeton 2013. 3 See for example the papers by Spring and Morrison, and Gifford et al., in this volume.

400

noëlle l.w. streeton

4 Coremans, Gittens, Thissen 1952, p. 2 and in abbreviated form in Coremans 1953, p. 76. Their ‘x’ was an additional component that could not be securely identified with contemporary instruments. See White, Pilc 1993, pp. 88-89; White 1995, pp. 132-135 for more recent identifications of pine resin. 5 Coremans 1953, p. 76: La technique picturale eyckienne est basée sur la translucidité des couleurs: rare sont les couches picturales réellement opaques, très nombreuse sont celles à effet translucide plus ou moine accentué. C’est la superposition des fines couches cristallines et de glacis, où la lumière se combine avec la couleur, qui produit les effets chromatiques si diversement nuancés. 6 See n. 1. 7 Of the circa 300 embedded samples, only eleven derive from the exterior grisailles, which were made available for this study by Jana Sanyova and Steven Saverwyns, KIK-IRPA Laboratory. 8 Burroughs 1939, pp. 178-179. The notion of Van Eyck’s technical perfection has its roots in the agendas of Giorgio Vasari and Karl van Mander, which were perpetuated by nineteenth-century historians of the Italian Renaissance. Claims for perfection therefore seem inappropriate for contemporary descriptions of Eyckian methods. 9 Streeton 2011, Table 1. Assessing the significance of differences between the Ghent panels and later works will be taken up briefly thereafter, but is largely beyond the scope of this paper. 10 Cross-sections were examined at the KIK-IRPA in May 2008; see also Coremans 1953, p. 119 and plate LXII. 11 Bosshard 1992, p. 7. 12 For Vermeer, see Sheldon and Costaras 2006, pp. 93-97. 13 A detailed report on the thirteen samples from the Turin painting was written by Leopold Kockaert (dated 12 October 1988). Jana Sanyova (KIK-IRPA) updated the interpretations in a report dated 19 August 2009. See also Spantigati 1997, pp. 23-25. Marigene Butler recorded detailed comparisons between the Turin and Philadelphia paintings after studying both. See Butler 1997, p. 32, pp. 38-39. She considered that their materials and technique were closely comparable. 14 Research at AMOLF (Amsterdam), undertaken by Klaas Jan van den Berg and Margriet van Eikema Hommes, clarified the mechanism of discoloration of copper greens. Their results suggest that discoloration corresponds to the thinness of the layers, the pH of paint and fluctuations in relative humidity, but not necessarily to light penetration. The discoloured material had been presumed to contain reduced copper species, perhaps to copper (I) in a cuprous matrix, but experimental evi-

dence remains insufficient to make firmer assertions. See Van den Berg et al. 2000; Van Eikema Hommes 2004, p. 81. 15 Coremans 1953, p. 72; Van Asperen de Boer, Giltaij 1987, p. 266; Brinkman et al. 1988-1989, pp. 29-34. 16 Francisco Pacheco warned later against applying copper-green paints too thinly his Arte de la Pintura (c. 1638). See Van Eikema Hommes 2004, p. 74. 17 See n. 14. 18 Coremans, Gittens, Thissen 1952, p. 16; Coremans 1953, pp. 71-73; Brinkman et al. 1988-1989, pp. 27-29, p. 32. 19 Coremans, Gittens, Thissen 1952, p. 16; Brinkman et al. 1988/89, p. 30. 20 Various scholars have queried the ‘map of the world of circular shape which [Van Eyck] painted for Duke Philip’ (orbiculari forma quam Philippo Belgarum pricipi pinxit). See, among others, Baxandall 1964, pp. 91-92; Steppe 1983, pp. 89-90; Paviot 1991, p. 58. 21 No underdrawing is visible in infrared reflectography, but using a low-tech alternative – white spirit and direct light – Butler was able to see some underdrawing under thin paint layers (see Butler 1997, p. 32). While the nature of the underdrawing is not currently identifiable, the image was potentially underdrawn with a metal stylus such as silverpoint. 22 Steppe 1983, pp. 92-94; Paviot 1991, pp. 58-59. 23 Examples include the interior of the robe of the Deity Enthroned in the Ghent Altarpiece (Coremans 1953, pp. 73-74), the bed-hanging in the Arnolfini Portrait (Roy 2000, p. 243); and the robe of Margaret van Eyck (see http://www.nationalgallery.org.uk/paintings/ research/the-restoration-of-margaret-the-artists-wife/margarets-reddress, accessed 2 January 2013). 24 Oltrogge 2008, unpublished report, Cologne Institute of Conservation Sciences/Institut für Restaurierungs- und Konservierungswissenschaft Fakultät für Kulturwissenschaften der Fachhochschule Köln. 25 Oltrogge 2008, unpublished report. 26 Brinkmann et al. 1988-1989, pp. 30-31. 27 Spring 2004, pp. 21-24; Spring 2007a, pp. 141-143; Spring 2007b, pp. 78-85. 28 See Vandivere 2011. 29 Weale 1908, pp. xxxviii-xxxix. See also Streeton 2011, pp. 166-171. 30 Gifford 1999, pp. 111-112.

Fig. 27.1 Jan van Eyck, Virgin and Child with Canon Joris van der Paele, 1436, panel, 124.5 x 160 cm, Bruges, Groeningemuseum (inv. no. 0000.GRO0161.I), detail, St Donatian in visible light (left) and IRR (right)

27

Gold-brocaded Velvets in Paintings by Jan van Eyck. Observations on Painting Technique Esther E. van Duijn

ABSTRACT: As part of my PhD study of the painting technique of gold-brocaded fabrics in fifteenth- and early sixteenth-century Netherlandish paintings, I studied goldbrocaded velvets on four paintings by Jan van Eyck: the Ghent Altarpiece, the Virgin and Child with Canon Joris van der Paele, the Virgin and Child with Chancellor Rolin and the Annunciation. Although Van Eyck’s gold-brocaded textiles have been subject of research in the past – most notably by Rembrandt Duits and Lisa Monnas – an overview of the painting technique of these textiles has never been taken into account. In painting gold-brocaded velvets, Van Eyck occupied a special position. Unlike most painters of his time, he had access to real life gold-brocaded textiles in his role as court painter and varlet de chambre to Philip the Good. In this article, I address the question of whether Jan van Eyck painted these velvets from life.

—o— Introduction1 Jan van Eyck is one of the first – if not the first – Northern European artists to imitate gold-brocaded velvets in his paintings using only paint. Before him, gold-brocaded textiles were imitated using a combination of gold leaf and paint.2 These more traditional methods produce decorative and effective results but they do not display the striking realism of painted gold-brocaded velvets. The cope worn by St Donatian in Van Eyck’s Virgin and Child with Canon Joris van der Paele is probably one of the most realistically painted pieces of textile from

that period, setting off the hard and smooth surface of the metal threads against the soft pile of the blue velvet. In early 2011 the Impact of Oil team of researchers was invited to the Groeningemuseum in Bruges to examine the large panel of the Virgin and Child with Canon Joris van der Paele while its protective glass was removed.3 In 2010 and 2011 the team members also studied the panels of the Ghent Altarpiece close-up during the conservation project in the St Bavo’s Cathedral.4 Gold-brocaded velvets can be found in the clothing of the angels in the Singing Angels and Musician Angels panels in the upper register and on the altar cloth in the Adoration of the Lamb panel in the lower register.5 These gold-brocaded velvets appear to be the first dated ones to have been created solely in paint, at least in the Netherlands. Two additional paintings by Jan van Eyck were examined in the autumn of 2012: The Annunciation from the National Gallery of Art in Washington and the Virgin and Child with Chancellor Rolin from the Musée du Louvre in Paris.6 In The Annunciation Gabriel wears a goldbrocaded red velvet cope and similar green velvet dalmatic. In the Virgin and Child with Chancellor Rolin the chancellor wears a gold-brocaded brown velvet houppelande. As far as we know, this is the first example of painted velvet with allucciolato;

404

esther e. van duijn

Fig. 27.2 Close-up of the IRR in Fig. 27.1

little loops of woven gold thread scattered over the brown velvet pattern.7 Together these four paintings contain the most important examples of painted gold-brocaded velvets in Jan van Eyck’s oeuvre.8 Much has already been written about real goldbrocaded velvets, and their depiction in paintings. Jan van Eyck’s brocades are significant in this field of scholarship because of his pioneering role.9 Of great importance is the fact that – as a court painter – Jan van Eyck could have seen for himself gold-brocaded velvets, even though it is highly unlikely that he owned larger pieces of these costly textiles himself.10 With this in mind, the verisimilitude in Van Eyck’s work poses an important question: could the gold-brocaded velvets in Van Eyck’s panels have been painted from life? Although many authors have commented on the superb realism of St Donatian’s cope, only a few publications suggest that it was actually painted after an existing set of church vestments. Viaene has argued that the cope and dalmatic were painted after one of the two copes of gold-brocaded blue cloth in the church inventories of 1417.11 However, the donor of these

copes, Willelmus Vernaechtenzone, died in 1393, so the copes are at least as old as that, more likely even older.12 The velvet depicted by Van Eyck is certainly more recent based on the pattern and the complexity of the velvet. Other authors argue that – even though Van Eyck must have studied gold-brocaded velvets closely – the vestments worn by Donatian were not necessarily painted from life. Pauwels noticed in 1985 that the cope in the painting is seamless. This is unlikely for real copes, which were often made of several pieces of gold-brocaded velvets, sometimes even with different patterns sewn together.13 Duits noted that the description of the two clasps belonging to the gold-brocaded blue copes mentioned in the archives does not match the one painted by Van Eyck.14 Although Monnas does not specifically refer to Donatian’s cope when discussing the question of painting textiles from life in her book, she generally states that fifteenth-century Flemish artists did not paint their gold-brocaded velvets from life.15 She also mentions another fact, which is applicable to Donatian’s cope too: when sewing

gold-brocaded velvets in paintings by jan van eyck

405

Fig. 27.3 Jan van Eyck, Virgin and Child with Chancellor Rolin, c.1430-1437, panel, 66 x 62 cm, Paris, Musée du Louvre (inv. no. 1271), detail, Chancellor Rolin in visible light (left) and an IRR of the sleeve (right)

a cope, the motif of the textile would be seen the right way up on the back, but would appear reversed or oblique on the front.16 In Donatian’s cope the motif appears to be the right way up on the front. When it comes to Rolin’s houppelande, authors seem to be much less certain. Monnas generally asserts that: ‘In a votive painting, where the donor is portrayed before the Virgin, the textile worn by the donor is likely to be accurately portrayed.’17 Only Tietzel claims that Rolin’s houppelande may well have been painted after an existing costume owned by the chancellor, but she does not gives any arguments for this apart from the fact that it was painted very realistically, including the allucciolato details and the fur trimming.18 Studying the painting technique of the different textiles may shed new light on this issue.19 The Underdrawing When looking at the infrared reflectograms (irr) of the four paintings, one clearly stands out: the

irr of the Virgin and Child with Canon Joris van der Paele, which shows a detailed underdrawing for St Donatian’s cope (fig. 27.1).20 This underdrawing not only shows the contours and folds of the drapery but also a very precise drawing of the pattern, including the different heights of velvet pile.21 This depicts a rare type of velvet that it is woven in three heights of pile.22 The lines of the underdrawing are thicker where the height of the velvet pile casts a small band of shadow on the surrounding area. Even the sheen of the velvet is suggested by a thin wash (fig. 27.2, blue arrow).23 A convincing image of velvet was therefore produced before any paint was applied. The visibility of this elaborate underdrawing contrasts with the other irr images. In the Virgin and Child with Chancellor Rolin, the underdrawing is clearly visible in the contours and folds of the donor’s houppelande, but the pattern was not drawn in at this stage (fig. 27.3).24 The underdrawing of the gold-brocaded textiles on the other two paint-

406

esther e. van duijn

Fig. 27.4 Jan van Eyck, The Annunciation, c.1434-1436, canvas (transferred from panel), 92.7 x 36.7 cm, Washington, National Gallery of Art (inv. no. 1937.1.39), detail, Angel Annunciate in visible light (left) and a detail of the cope of the multispectral IRR: composite of three registered sets of false-colour spectral images, in wavelength bands 1100-1400 nm (blue), 1500-1800 nm (green), and 2100-2400 nm (red)

ings, however, is much less visible. For the respective copes of the singing angel in the Ghent Altarpiece and the archangel in The Annunciation, only hints of the underdrawing of the contours and folds of the drapery are visible, while no underdrawing of the pattern was found at all (fig. 27.4).25 The dalmatic worn by Gabriel in The Annunciation and the robe of the organist angel in the Ghent Altarpiece both have paint layers that obscure any possible underdrawing due to their thickness or pigment content. However, it is very unlikely that the pattern on the robe of the organist angel was underdrawn in this phase, because previous research has suggested that it was not planned as a brocaded textile from the beginning but was meant to be a white garment at first.26 Remarkably, in the two textiles mentioned above, the pattern is very similar, which suggests

that Van Eyck reused the pattern of the organist angel on Gabriel’s dalmatic, even though he diminished its size and adapted the motif somewhat to fit the iconology of The Annunciation.27 The pattern in The Annunciation is well defined in normal light, showing green velvet in two heights of pile. This differentiation has been lost on the robe of the organist angel, probably due to pigment degradation and aged varnish layers, but the double height of pile is still visible in the irr image. As mentioned above, it has been argued by some that Donatian’s cope and Rolin’s houppelande were painted from life. Interestingly, one shows a clear underdrawn pattern and the other does not. Of course, the fact that a pattern was or was not underdrawn does not prove whether or not it was drawn from life.28 In the case of the Virgin and Child with Chancellor Rolin, a gold-brocaded velvet owned

gold-brocaded velvets in paintings by jan van eyck

by the chancellor could of course also have been drawn from life on paper or parchment, only to be transferred to the painting at a later stage during the painting process using a medium invisible in irr. It is even possible that the gold-brocaded textile was not planned from the beginning. In irr the underdrawing shows a purse hanging from Rolin’s belt as a symbol of his office; however, it was never executed in paint and it has been argued that this might have been at the instigation of the chancellor himself.29 Could it be that the gold-brocaded velvet was planned for only after the decision was made not to execute the purse in paint? Similarly, the elaborate underdrawing of Donatian’s cope may well have served as a vidimus, shown to – and approved by – the patron before any paint layers were applied. The difference between the two is that here the gold-brocaded velvet was planned from the outset. Without doubt, it would also have helped the artist during the application of the subsequent paint layers, even though the thinner lines and washes soon became invisible under paint. Remarkably, in the underdrawing of Donatian’s cope the pattern was corrected in a few areas during the drawing process, as if the artist was not certain about the exact outline of some details (see fig. 27.2, yellow arrows). The shading of the drapery in the underdrawing has been rendered in both wash and hatching. The hatched lines appear similar in thickness and consistency to the lines of the pattern and were probably applied right after the pattern, because they follow the contour lines of the pattern in most areas. The washes too are applied just after the application of the pattern, because in a few areas, the lines of the pattern are smudged by the application of the wash. This seems to indicate that the modelling of the drapery and the application of the pattern were done in one go and would argue against the use of a model drawing of a flat pattern. It could even be argued from this that the pattern was drawn from life, but even if this were the case there is a counter argument that indicates that a separate model drawing is likely to have existed too. Contrary to the underdrawn

407

pattern of the cope, the pattern on Donatian’s dalmatic – the vestment that he wears underneath the cope and which is partially visible – was not drawn in at the underdrawing stage (see fig. 27.1). Even though the paint in this area is quite degraded today, there is no doubt that the gold-brocaded velvet pattern was indeed painted by Van Eyck. Since it was not underdrawn, this pattern must have been applied at a later stage during the painting process. It can be argued that for this he would have needed a model drawing, probably the same one that he used earlier for the pattern of the cope, since it the same pattern is applied in both vestments. No evidence was found as to how such a transfer took place or in exactly which phase of the painting process.30 Foreshortening the pattern One aspect often pointed out as an indication of whether a pattern was painted from life is the presence of foreshortening in the pattern. If a model drawing of a flat pattern was used instead of a real life model, one ought to be able to see this on the painting, especially if the textile containing the pattern has been folded.31 It has been argued that the patterns in paintings by Van Eyck were indeed foreshortened convincingly.32 However, it is deceptively difficult to see if a pattern really is foreshortened by just looking at the painting. The painted shadows of the textile folds and the application of the highlights imitating the gold thread, all converge to create the optical illusion of a foreshortened pattern, even if the pattern has in fact been applied flat. An additional, very effective trick used by later artists, was to shift a flat pattern over the drapery folds to create an illusion of foreshortening.33 For the study of St Donatian’s cope the underdrawing – without the interference of paint layers34 – helps one to see if the pattern had been foreshortened realistically. For the other examples, apart from the cope of the singing angel, tracings have been made for a similar purpose.35 When studying Donatian’s cope the pattern appears to be truly foreshortened in one place only and that is in the centre left part, where the textile curves to the

408

esther e. van duijn

back (see fig. 27.2, red arrow). In other areas, the illusion of foreshortening seems to have been visually created by a clever positioning of the pattern and highly skilful application of the paint layers, especially of the yellow highlights. A flat pattern on a model drawing may for instance have been transferred to a transparent sheet and this pattern may then have been transferred to the panel, for example by pouncing.36 Using a transparent sheet, the modelling of the drapery on the panel would remain visible in order to shift the pattern over the folds of the drapery.37 For a gifted artist like Jan van Eyck it could not have been too difficult to apply a pattern, for example flat from a model drawing as described above, and then choose one or two areas where part of the flat pattern is adapted to imitate accurate foreshortening, in order to heighten the illusion of a truly foreshortened textile. This idea is confirmed by the other examples where the larger pattern motifs appear to have been applied flat and foreshortening only occurs in the smaller details. This is particularly visible in the pattern of the organist angel in the Ghent Altarpiece (fig. 27.5). In figure 27.6, one of the many small bunches of grapes of the pattern has been foreshortened. However, the grapes first appear to have been left in reserve with rounded, unforeshortened forms (fig. 27.6, white arrow). This suggests that foreshortening took place at a later stage of the painting process, that is, after the initial application of a flat pattern. The fact that this pattern has been reused on Gabriel’s dalmatic in The Annunciation also seems to suggest that a drawing of the pattern must have existed at the workshop. The size of such a drawing remains uncertain, because for The Annunciation the pattern was scaled down to less than half the size of that on the Ghent Altarpiece. Resizing a drawing could be done by squaring, using grids in two sizes, possibly on a separate paper, because on neither painting has any trace of a grid been found. It may also have been done freehand, adapting the aforementioned, smaller details at the same time.

The Paint Layers When the paint layers of the different textiles are examined it becomes clear that Van Eyck did not make use of a fixed, systematically applied layer build-up.38 However, clear parallels can be found between the different paintings, for example in the underpainting. In all cases, this first layer is thin and slightly streaky, and it was employed uniformly without any modelling of the textile. In St Donatian’s cope the layer must have been thin enough for the underdrawn pattern to show through. The colour of this underlayer is warm orange-brown in every case except in Gabriel’s cope, where it is a bright red layer of vermilion.39 This red colour would later serve as the base colour for the red velvet pattern. Similarly, the brown underlayer in the cope of the singing angel was used as a base tone for the red velvet pattern. In both cases, translucent red paint layers were applied to create the velvet.40 In the singing angel’s cope the overall effect is somewhat more subdued than in Gabriel’s cope, owing to the difference in the colour of the underlayer. The next step of the process was not painting the pattern, however, but the application of opaque brown layers to serve as the background for the gold-brocaded areas, i.e. the spaces that would be covered with yellow highlights to imitate gold threads. These layers were applied locally between the areas that were reserved for the velvet pattern. This means that the outlines of the pattern – had they not yet been applied in the underdrawing stage such as on Donatian’s cope – would have been applied on top of the first painted underlayer described above.41 In the robe of the organist angel, this pattern may be visible in some areas as thin dark brown paint. Here and there this paint obtrudes from underneath the more opaque paint layers of the pattern. Through a microscope, it is possible to see that the initial brown outlines of the pattern are overlapped on the other side by the ochre brown paint that serves as the base tone for the gold threads. Such painted lines have not been found in the other painted brocades, even though at this stage the design of the pattern must have

gold-brocaded velvets in paintings by jan van eyck

409

Fig. 27.5 Ghent Altarpiece, Musician Angels, detail, robe of the organist angel, digitally overlaid with a tracing of the textile pattern.

been present in those too.42 With one exception, in all examples the modelling of the gold-brocaded areas has been executed in opaque brown tones ranging from a cool ochre-brown to dark brown in the shadows of the folds. The exception – again – is Gabriel’s cope, where the gold-brocaded areas are

covered with a thin dark brown paint with very little modelling. Occasionally this paint layer is applied so thinly that the bright red underlayer shows through (fig. 27.7, white arrow).43 The next phase was the modelling of the pattern. As noted before, this was achieved in the red

410

esther e. van duijn

Fig. 27.6 Ghent Altarpiece. Musician Angels, close-up detail, robe of the organist angel

velvets by applying red translucent glazes over the warm orange-brown or bright red underlayer. In the case of Donatian’s cope, blue was applied thickly or opaquely to cover the orange-brown layer underneath.44 Similarly, the green paint layers of Gabriel’s dalmatic and the greenish brown paint layers of the organist angel’s robe were applied with opaque paint.45 In the pattern of Rolin’s houppelande the orange-brown underlayer does not show through the thickly applied reddish brown paint of the velvet pattern, but it is still visible sometimes because it was scratched away in thin lines locally in the lower half of the garment, to delineate the two heights of velvet pile in the main motif. The next phases were the applications of the details to heighten the realistic effect of the textile. First lines of dark paint were applied strategically along the pattern to suggest the shadow caused by the pile height of the velvet. The sheen of the velvet was also further enhanced, for example in the

red velvets, by applying a few pinkish scumbles in areas that would catch the light. Especially beautiful is the reddish brown velvet of Rolin, which has been worked out in fine detail to give a measure of how the more degraded velvets such as the blue of Donatian’s cope and dalmatic may have appeared initially. Next was the application of yellow highlights to imitate the gold threads. The brushwork here is often surprisingly free and spontaneous.46 As figure 27.8 shows, none of the painted gold threads are alike. All have been adapted to their specific situation, lighting and location. Van Eyck used several techniques to create the illusion of gold thread, from stippling over thin lines, to scratching in, to dragging fine lines through the wet paint underneath.47 In all cases the overall direction of the yellow strokes is diagonal unlike the direction of actual gold threads in a fabric, which was horizontal. Thus Van Eyck did not paint the threads themselves, but depicted instead the play of light on

gold-brocaded velvets in paintings by jan van eyck

411

Fig. 27.7 The Annunciation, Close-up detail from the Angel Annunciate’s cope

them; he painted what he saw rather than how he knew it to be.48 Of course, many an author has already commended the extraordinary powers of observation displayed by Van Eyck. The high degree of perfection has been taken furthest in the velvets of The Annunciation. Here, the last phases of applying the finer details and highlights have been executed in several stages, going back and forth between the details in both the velvet pattern and the gold-brocaded areas until the desired result was attained. In Gabriel’s cope, one final set of thin red lines was applied strategically over the yellow highlights, imitating the reflections of both the red velvet pattern and the red part of the angel’s wings in the metal thread of the textile (see fig. 27.7, yellow arrow). The other gold-brocaded velvets, especially the ones in the Ghent Altarpiece, look rather simple by comparison. The reason for this probably reflects the differences in size and – related to this – viewing

distance. Although we are uncertain where and how The Annunciation originally hung, its smaller size suggests that it was viewed closer-up than the Ghent Altarpiece. Rolin’s houppelande shows a degree of finish similar to Gabriel’s vestments, although it is executed in fewer phases and with somewhat more confidence. Before concluding, one last observation is in order. It has already been mentioned that the patterns on Gabriel’s dalmatic and the robe of the organist angel are nearly identical and that a model drawing is likely to have existed of this pattern, making it unlikely that Gabriel’s dalmatic was painted after a real gold-brocaded dalmatic.49 This knowledge is strengthened by some observations in the paint layers. First, the red fringe at the lower edge of the dalmatic was not planned from the beginning, but was painted over the finished gold brocade, which initially extended to the edge (fig. 27.9).50 Also, a pentimento found in the pattern of the dalmatic

412

esther e. van duijn

Figs 27.8a-f Close-up details of gold-brocaded velvets (top row, left to right): (a) Ghent Altarpiece, cope of the Singing Angel; (b) Ghent Altarpiece, robe of the organist Musician Angel; (c) Virgin and Child with Canon Joris van der Paele, St Donatian’s cope; (bottom row, left to right): (d) The Annunciation, Angel Annunciate’s cope; (e) The Annunciation, Angel Annunciate’s dalmatic; (f) Virgin and Child with Chancellor Rolin, Chancellor Rolin’s houppelande

seems to point in the direction of a model drawing; one of the almost finished green pattern lines was covered – not very effectively and not completely – by yellow highlights in a later phase (fig. 27.9, white arrows). Had an actual gold-brocaded dalmatic – or even a patterned piece of cloth – been used as a model, the pattern is unlikely to have been adjusted so late in the painting process, even though in theory the artist can of course have decided to change the pattern at such a late moment for aesthetic reasons. Conclusion This article has demonstrated that the question of whether Van Eyck painted his gold-brocaded velvets from life is difficult to answer from a technical point of view. An elaborate underdrawing – such as that found beneath the cope of St Donatian – has several indications that it was drawn from life, but whether

this was executed directly on the panel or using a model drawing as intermediary stage remains uncertain. Given the fact that the dalmatic underneath the cope was not underdrawn seems to indicate that the vestments were not drawn from life as a matching set, as suggested by some in the literature. The fact that the pattern of Rolin’s houppelande was not underdrawn, on the other hand, does not mean that it was not drawn from life. The underdrawing with a purse may have served as a vidimus – just like the underdrawing of Donatian’s cope – and was altered after viewing by the commissioner. It is possible that the gold-brocaded velvet was only planned for at a later stage, but this of course does not mean it was not drawn or painted from life. On the other hand, studying the tracings of the various patterns seems to point in another direction: that of a flat pattern that has been applied to the panel, possibly by using a transparent sheet from

gold-brocaded velvets in paintings by jan van eyck

Fig. 27.9 Jan van Eyck, The Annunciation, close-up detail from the Angel Annunciate’s dalmatic

413

414

esther e. van duijn

which the pattern is pounced, and then adapted locally on the panel to cleverly fit the play of the drapery folds and create a convincing illusion of foreshortening. The fact that one pattern has been used twice on different types of garments – the organist angel’s robe and Gabriel’s dalmatic – also suggests the existence of a model drawing of the flat pattern. As we have seen, the build-up of the paint layers is not similar in all painted brocades, The Annunciation being significantly different from the others. However, parallels between the techniques used on the paintings suggest the beginning of a more fixed workshop method such as found in the work of later artists. Clearly, Van Eyck’s famed powers of observation play an important role in the depiction of his gold-brocaded velvets and he must have studied these textiles intensely. NOTES * This article could not have been written without the help of the following people: J.R.J. van Asperen de Boer (formerly University of Groningen), Till-Holger Borchert (Bruges, Groeningemuseum), Christina Currie (Brussels, KIK-IRPA), Hélène Dubois (Brussels, KIKIRPA), Molly Faries (Bloomington, Indiana University), Melanie Gifford (Washington, National Gallery of Art), Anne van Grevenstein (formerly University of Amsterdam), John Hand (Washington, National Gallery of Art), Friso Lammertse (Rotterdam, Museum Boijmans van Beuningen), Julie Mauro (Brussels, KIK-IRPA), Marc de Mey (Brussels, Flemish Academic Centre for Science and the Arts), Anne van Oosterwijk (Bruges, Groeningemuseum), Cécile Scailliérez (Paris, Musée du Louvre), Ron Spronk (Kingston, Queen’s University / Nijmegen, Radboud University), Margreet Wolters (The Hague, Netherlands Institute for Art History), and the Impact of Oil team, especially Jan Piet Filedt Kok and Abbie Vandivere. 1 This research was carried out within the project The Impact of Oil: A History of Oil Painting in the Low Countries and its Consequences for the Visual Arts, 1350-1550. The aim of this project was to write an integrated history of the introduction, dissemination, and development of the use of oil media in panel painting from 1350 to 1550. The project was funded by the Netherlands Organization for Scientific Research (NWO) and directed by Prof. Dr Jeroen Stumpel and Prof. Dr Jan Piet Filedt Kok. My PhD dissertation All that glitters is not gold. The depiction of gold-brocaded velvets in fifteenth- and early sixteenth-century Netherlandish paintings was successfully defended on 19 June 2013 at the University of Amsterdam. 2 For a discussion on the pre-Eyckian techniques used for depicting gold-brocaded silks, see Van Duijn 2013b (unpublished), pp. 83-84. 3 Jan van Eyck, Virgin and Child with Canon Joris van der Paele, 1436, panel, 124.5 ≈ 160 cm, Groeningemuseum, Bruges, inv. no. 0000.GRO0161.I. 4 Jan and Hubert van Eyck, Ghent Altarpiece, 1432, panel, c. 375 ≈ 520 cm (framed altarpiece, open), St Bavo’s Cathedral, Ghent, inv. nos 413-424. Since this article does not deal with the attribution question, for simplicity’s sake, the name Hubert will not be mentioned after this. 5 Measurements of the unframed panels are: 164.5 ≈ 72.3 cm (Singing Angels), 164.5 ≈ 73.1 cm (Musician Angels) and 138.1 ≈

243.3 cm (Adoration of the Lamb). The painted gold-brocaded red velvet of the altar cloth in the Adoration of the Lamb is simpler in execution than the gold-brocaded velvets worn by the angels and will not be mentioned in the rest of the article. The gold-brocaded dorsers behind the Virgin Enthroned, the Deity Enthroned and John the Baptist Enthroned in the upper register fall outside the scope of this article, because they are gold-brocaded lampas silks, not velvets, and have been created using the applied brocade technique involving gold leaf. Geelen and Steyaert 2011, pp. 382-393; Geelen 2012, pp. 129-142. 6 Jan van Eyck, The Annunciation, c.1434-1436, canvas (transferred from panel), 92.7 ≈ 36.7 cm, National Gallery of Art, Washington, inv. no. 1937.1.39; Jan van Eyck, Virgin and Child with Chancellor Rolin, c.1430-37, panel, 66 ≈ 62 cm, Musée du Louvre, Paris, inv. no. 1271. The manner in which all four paintings were studied is described in n. 35. 7 Monnas 2008, pp. 111, 113. The cushion on the Virgin’s chair in this painting is a gold-brocaded blue velvet. It was studied, but will not be described further in this paper, because it is so small compared to the other examples and its paint layer build-up is very similar although simplified in manner. 8 Other examples of gold-brocaded textiles in paintings by Van Eyck depict lampas silks: gold-brocaded silks with a flat weave, such as the cloth of honour behind the Virgin in the Virgin and Child with Canon Joris van der Paele. 9 Two important recent publications are Monnas 2008 and Duits 2008. 10 He was court painter and varlet de chambre to John III, Duke of Bavaria (1374-1425) between 1422 and 1424 at least and to Philip the Good, Duke of Burgundy (1396-1467) from 1425 until his death shortly before 1441. On the high costs of gold-brocaded silks, see Duits 2008, esp. pp. 70-80, 91 and Monnas 2008, p. 39. 11 Viaene 1965, p. 264. Viaene quotes from the inventory of the church of St Donatian, published by Waele. Weale 1864-1865, p. 105. Other authors refer to the article of Viaene: Janssens de Bisthoven 1981, p. 197; Tietzel 1994, p. 224; Harbison 2012, pp. 58, 264. 12 Weale 1864-1865, pp. 19, 105. 13 Pauwels 1985, p. 233. 14 Duits 2008, pp. 178-179. 15 Monnas 2008, pp. 121-147. Interestingly, in an earlier article on the silks in paintings by Van Eyck she still repeated Viaene’s argument. Monnas 2000, p. 150. 16 Monnas 2008, p. 187. 17 Monnas 2000, pp. 148-149. This statement is repeated somewhat more strongly by Duits, but without further argument. Duits 2008, p. 41. It must be mentioned that in her book from 2008, Monnas no longer refers to this possibility. 18 Tietzel 1994, p. 224. 19 The length of this paper does not allow for a comparison of Van Eyck’s painting techniques with those of contemporaries and later artists, but I added these in my PhD thesis. 20 Infrared reflectography of the Virgin and Child with Canon Joris van der Paele and the Ghent Altarpiece was carried out on-site in the Groeningemuseum by Sophie De Potter of the Royal Institute for Cultural Heritage (KIK-IRPA) in Brussels. The equipment used was an InfraCAM-SWIRÔ (short-wave infrared) video camera, which has a platinum silicide (PtSi) detector and 256 ≈ 256 focal plane array. A 1.5-1.73 micron filter was placed behind the 12 inch focal length lens during image capture and the images (5 cm square) were assembled and processed using Adobe Photoshop. 21 Maryan Ainsworth already noted this aspect of the underdrawing in her article on the IRR of the Virgin and Child with Canon Joris van der Paele and suggested further study. Ainsworth 2003, p. 277. 22 Monnas 2008, p. 116. 23 The use of washes is not uncommon in Van Eyck’s underdrawing. See for example: Van Asperen de Boer, Faries 1990, p. 47 and Gifford 1999, 116 (n. 15). 24 IRR of the Virgin and Child with Chancellor Rolin by J.R.J. van Asperen de Boer and Molly Faries at the Netherlands Institute for Art

gold-brocaded velvets in paintings by jan van eyck

History (RKD). Technical specifications of the equipment: IRR was performed with a Grundig FA 70 television camera equipped with a Hamamatsu N 214 IR vidicon (1975); a Kodak wratten 87A filter cutting-on at 0.9 micron was placed between the vidicon target surface and the Zoomar 1:2 8/4 cm Macro Zoomatar lens. The television camera was mounted on a sturdy Linhof professional tripod with extension pieces, and a 90 cm sledge for moving the camera sideways. The monitor was a Grundig BG 12 with 875 television lines. Documentation was done with a Nikon camera, a 50 mm macro lens, and Ilford film FP 4, ASA 125. The IRR-assembly reproduced here was made with Adobe Photoshop and consists of images which were scanned from photo negatives from the archive of Prof. Dr J.R.J. van Asperen de Boer at the RKD. 25 For the Ghent Altarpiece, Sophie De Potter at the KIK-IRPA carried out the infrared reflectography (see n. 20). The images were studied using the website Closer to Van Eyck: Rediscovering the Ghent Altarpiece (http://closertovaneyck.kikirpa.be). Technical specifications of the False Colour IRR from The Annunciation: Multispectral infrared reflectography (MS-IRR) images were collected by Paola Ricciardi and Doug Lachance in three spectral bands (1100-1400 nm, 1500-1800 nm, 2100-2400 nm) using an indium antimonide (InSb) camera fitted with a custom near-infrared lens. These MS-IRR image sets were processed and registered with the visible-light image by John Delaney to create the false-colour image and overlay it with the visible light image using a custom mosaic algorithm. See also: Conover et al. 2011, pp. 1-9. 26 Brinkman et al. 1988-1989, pp. 34-35. 27 Monnas 2008, pp. 121-123. 28 Interestingly, Ainsworth concluded that St Donatian’s cross and mitre might well have been painted after existing objects, precisely because they appear not to have been underdrawn. Ainsworth 2003, p. 284. A similar conclusion was drawn by Molly Faries regarding portraits where only non-descriptive egg shapes were found in the underdrawing phase. Faries 2003, p. 25. 29 Van Asperen de Boer, Faries 1990, p. 43. On the function of the underdrawing as a vidimus to be seen – and commented upon – by the commissioner, see Faries 2001, pp. 85-86. 30 This is partly due to the poor condition of the blue paint layers, which appear even more degraded in the dalmatic than in the cope. 31 In a dorser on the other hand, the pattern appears flat to the eye, regardless of whether it has been painted from life or after a model drawing. 32 Duits 2008, p. 53 and more generally: Monnas 2008, p. 121. 33 There are numerous examples of flat applications of brocade patterns with an illusion of foreshortening that are more (or less) convincing. See for example: Goddard 1985, p. 403; Périer-D’Ieteren 1985, pp. 47, 58, 89, 108; Van Duijn, Roeders 2012, pp. 5, 6, 8-10 (PDF); Van Duijn 2013a, p. 37. 34 The blue paint layers are very degraded today and no longer show their original surface and modulation. 35 Similar to the underdrawing, a traced pattern can be studied without interference of colour and modelling. Tracings were made on a Melinex sheet from 1:1 photographs of the paintings kept at the Flemish Academic Centre for Science and the Arts (VLAC) in Brussels. These photographs were reproduced from negatives made by the late Father Alfons Dierick of Ghent. The negatives were donated by Dierick’s family to Ghent University, where they are conserved and have been digitized. Even though tracings from photos are less accurate than from the actual paintings, they were still useful. The tracings from the patterns of The Annunciation and the Virgin and Child with Chancellor Rolin were held against their respective paintings; no photographic distortions were detected. It was found that the dimension of The Annunciation tracing was a few millimetres wider than the actual painting. The Rolin tracing seemed an exact match. The tracing from the robe of the organist angel unfortunately could not be checked against the actual painting. 36 Although different techniques to transfer a drawing to a panel – such as tracing or pouncing – are described in contemporary sources, no conclusive evidence has been found – either in sources or on the paintings themselves – of exactly how a model drawing of a flat pattern

415

was transferred to a panel. For a more elaborate discussion on model drawings of textile patterns and possible methods of transfer, see Van Duijn 2013b (unpublished), pp. 98-106. 37 Writing on how to make transparent sheets is found for example in Cennini’s Il Libro dell’ Arte and in the Montpellier Liber diversarum arcium. Cennini (ed. Thompson 1960), pp. 13, 14; Clarke 2011, pp. 98. 244. 38 In all cases the paint layers were studied using a binocular head loupe (magnification 5.5x) and – with the exception of the Virgin and Child with Chancellor Rolin – with a microscope. For the Ghent Altarpiece, the on-site microscope, lent by Ghent University, was used. For the Virgin and Child with Canon Joris van der Paele and The Annunciation, the portable Impact of Oil microscope was used. Apart from the Ghent Altarpiece of which photos were provided through the website Closer to Van Eyck: Rediscovering the Ghent Altarpiece, digital photography was executed using a Fuji Finepix S7000. 39 For the identification of vermilion, see Gifford 1999, p. 108. 40 In Gabriel’s cope, black pigment was added to the red glaze, which may account for the fact that these paint layers are so difficult to penetrate by infrared. Gifford 1999, p. 108. Adding some black pigment to red glazes has also been found on the Ghent Altarpiece. Brinkman et al. 1988-1989, p. 36. 41 The painted pattern seems to follow the underdrawn pattern faithfully, even though the blue paint layers are somewhat difficult to judge, because they are quite degraded and no longer show their original surface and modulation. On the degradation of the blue pigment ultramarine in paintings by Jan van Eyck, see also Gifford et al. and Spring and Morrison in this volume. 42 The outlines of the pattern were applied in a medium difficult to detect in infrared or under the microscope. White chalk, for example, would be effectively erased by subsequent paint layers, because it has almost the same refractive index as oil itself. See also: Van Duijn 2013a, p. 33. On different methods of pattern application found in the brocaded textiles of Early Netherlandish paintings, see also: Devolder 2009, pp. 61-74. 43 This red colour has an effect on the final appearance of the gold-brocaded areas too, giving them a more defined, somewhat harsher look. 44 As has been previously noted, it is difficult to accurately judge the blue paint layers of the cope and dalmatic because they are so degraded, causing the loss of much of the modelling of the velvet pattern. Drying cracks in the blue paint clearly reveal the orangebrown layer underneath. 45 The paint layers of the brown velvet in the robe of the organist angel are difficult to assess due to their degraded nature and obscuring varnish. As has been noted before, the IRR of the robe shows the brushstrokes and a double pile height better than any image in normal light. Some of the lighter areas in the brown velvet show a greenish tinge. It is possible that the brown paint was originally much greener like the pattern on Gabriel’s dalmatic, which is after all almost the same pattern. In literature no reference has been made to this possibility and only future scientific analysis might enlighten us further. 46 Others have noted this too: Jo Kirby, for example, calls Van Eyck’s painting technique ‘impressionistic on a small scale’. Kirby 2012, p. 273. See also Campbell in this volume. 47 These are techniques that Van Eyck also applied in other areas of his paintings. See also Vandivere in this volume. 48 Monnas 2008, p. 116. For a more scientific approach to the phenomenon of specularities on surfaces with tangential grooves, see Lu et al. 1999, pp. 2-7. 49 The outer contours of the larger main motif and the smaller detail in its centre have been adapted to resemble dianthus flowers; Monnas has argued that this was a conscious iconographic choice. Monnas 2000, p. 151. 50 We can compare this to the underdrawing of Donatian’s cope, where the fringe was planned from the beginning, and his dalmatic, which was planned and executed without a fringe.

Figs 28.1a-b Blotting in Jan van Eyck, Virgin and Child with Canon van der Paele, 1436, oil on panel, 122.1 x 157.8 (with frame), Bruges, Groeningemuseum (inv. no. 0000.GRO0161.I): (a) in the Virgin’s red cloak; (b) texture of red dots visible through the Christ Child’s cloth

28

Surface Effects in Paintings by Jan van Eyck Abbie Vandivere

ABSTRACT: Surface effects in paintings by Jan van Eyck enhance or disrupt the ‘perfect’ surface that many people associate with his work. He manipulated his paint in many ways, including blending it wet-in-wet, scraping into the paint, and applying it with more or less impasto. Other surface effects are related to the handling of his materials and the rheology of his oil paint. The presence of finger and palm prints in red glazes suggests that he used his hands to blot them. Reconstructions were useful for visualizing and attempting to reproduce these surface effects. Observations and macrophotographs are presented from several Van Eyck paintings, including the Virgin and Child with Canon Joris van der Paele (1436, Bruges, Groeningemuseum) and the Ghent Altarpiece (1432, Ghent, St Bavo’s Cathedral). The findings from these recent examinations are compared with published observations. This research was conducted as part of the Impact of Oil project.

—o— The surface appearance of Jan van Eyck’s paintings has been described as jewel-like, enamel-like, sparkling and glowing.1 Phrases like elegant… parfaits et mignonnetz (elegant, perfect and delicate), brilliant… émaillé (brilliant and enamelled), and spieghels (mirrored) reflect the value that historical writers placed on the appearance of his paint.2 Van Eyck’s small devotional paintings and portraits seem to have ‘perfect’ surfaces with little or no prominent surface texture, and few visible brushstrokes. Under the microscope, very fine hatching can often be seen, but the illusion from further away is of a flawless surface. In contrast, many of his larger works contain surface effects

that enhance or disrupt this perfect illusion. It seems that the need to apply paint smoothly over a large area required him to adjust his painting technique to include methods like blotting glazes and manipulating the paint with his hands or fingers. He also used other tools to manipulate the surface of the wet paint: a brush to blend one colour into another, and a blunt instrument to scrape into the paint. These surface effects were made possible – and sometimes necessary – by the oil binding medium that Van Eyck used.3 Its slow drying and fluidity allowed him to create a variety of sophisticated effects, including wet-in-wet blending and subtle transitions between colours. Modifications to the medium sometimes caused handling problems, and solutions to these problems created visible surface effects. The following section will introduce the rheology of the types of oil that Van Eyck used. Subsequently, this paper will focus on the various ways that Van Eyck manipulated his paint, and the surface effects that have been observed in his oeuvre. As part of an investigation into the painting technique of Early Netherlandish paintings within the Impact of Oil project, I had the opportunity to examine many works by Jan van Eyck and his contemporaries. Basic non-destructive methods were used, and the project was often granted access to technical documentation and treatment records.4 The observations derived from these examinations were compared with published observations from paintings such as the Arnolfini Portrait (1434,

418

abbie vandivere

London, National Gallery), the Portrait of Margaret van Eyck (1439, Bruges, Groeningemuseum), and The Annunciation (c.1434-1436, Washington, National Gallery of Art). Results from an examination of the Ghent Altarpiece (1432, Ghent, St Bavo’s Cathedral) are discussed as a case study at the end of this paper. In order to give a broader context, Van Eyck’s surface effects will also be compared to those in paintings of other Early Netherlandish artists.5 Van Eyck’s Oil Medium The properties of the oil medium used by Van Eyck and his Early Netherlandish contemporaries allowed for a variety of effects and textures. It can be used to create glazes, because linseed oil has a refractive index close enough to certain pigments – chiefly organic reds and copper greens. It is relatively slow-drying, fluid, and allows colours to be subtly blended together. A certain amount of impasto is also possible. In Van Eyck’s paintings there are some small details that protrude slightly from the surface; for example, opaque dots create highlights on metal and define feathers on angels’ wings.6 In the Ghent Altarpiece, a slightly impasted, stippled texture was used to create the leaves of the trees and headdresses of the female saints in the Adoration of the Lamb. Often, small dots decorate the edges of clothing: tassels and fringes, the ruffles in the woman’s green cloak in the Arnolfini Portrait, and Margaret’s headdress in the Portrait of Margaret van Eyck.7 We cannot know for certain how Van Eyck’s paints would have appeared or behaved when they were still wet, but the results of binding medium analysis tell us something about their properties. Generally, his paints were bound in linseed oil. Sometimes he used oils modified by pre-polymerization (including heat-bodying) or added varnish. Leaving linseed oil to stand in an open dish in the sun for a long period of time would make it more viscous, a process referred to as sun-thickening.8 This simple process could have been carried out by the artist himself. The light and oxygen would

pre-polymerize the oil, and also make it less yellow and slightly increase its refractive index.9 Another way that Van Eyck could slightly increase the refractive index of his red and green glazes was to add a few drops of varnish. This would have given the glazes increased gloss, richness, translucency and colour saturation. Analysis revealed that Van Eyck added a small amount of pine resin to the linseed oil in the woman’s green cloak in the Arnolfini Portrait.10 A minor quantity of diturpenoid resin was also identified in the green dalmatic in the Annunciation, while a sample from the architecture was simply bound in linseed oil.11 All samples from the Portrait of Margaret van Eyck contained heat-bodied linseed oil, a type of pre-polymerized oil that can be detected using modern analytical techniques.12 The oil may have been subjected to gentle heating or boiling, which would have made it more viscous and likely dry faster. Translucent glazes played an important role in Van Eyck’s paintings, especially in broad passages of red and green drapery.13 Several of Van Eyck’s red glazes have undergone medium analysis. A midtone red on the inside of the canopy of the Arnolfini Portrait was bound in linseed oil, but no modifications to the oil are specified. A shadow in the red turban of the Portrait of a Man (Self-Portrait?) (1433, London, National Gallery) was also bound in linseed oil, with a small amount of pine resin added.14 The red glazes in Margaret’s clothing in the Portrait of Margaret van Eyck were based on heat-bodied linseed oil.15 The pre-polymerized oils that Van Eyck used in certain colours were vital to producing a smooth, glassy surface without visible brushstrokes. Glazes made from an untreated (raw) linseed oil tend to show brushstrokes, whereas paints made with sun-thickened oil settle and level out and the strokes disappear soon after the paint is applied. Reconstructions have been useful for visualizing the properties of various pigments when bound in oils modified in different ways.16 Organic red (madder lake) glazes bound in unmodified linseed oil preserved the appearance of brushstrokes. In

surface effects in paintings by jan van eyck

comparison, glazes in sun-thickened oil formed a smooth, well-saturated film without visible brushstrokes.17 The use of pre-polymerized oils also carried certain disadvantages. Reconstructions of red glazes prepared with sun-thickened linseed oil tended to be extremely difficult to handle, and were so thick and gloopy that they were difficult to spread thinly with a brush.18 They also spread beyond the boundaries of the intended brushstroke, and would not create a lasting impasto. When applied in a thick layer, the sun-thickened paints wrinkled, sometimes soon after they were applied. Boiled oil had similar rheological properties.19 Jan van Eyck may have encountered some of these handling and drying problems when painting his red draperies. Perhaps this is why he needed to even out the surface of many of his red glazes with blotting. Blotting and Palm Prints Red textiles feature prominently in Van Eyck’s works, and their surface is almost always a red glaze. Inspection under magnification reveals a fascinating part of his working process. In some large passages of red drapery, Van Eyck used his fingers and palms to dab the surface of his glazes. Dunkerton et al. (2008) identified this phenomenon in the portrait of the artist’s wife, Margaret van Eyck.20 The folds of her red dress are built up of several layers of red glaze on top of an unmodelled vermilion underpaint. In order to thin and even out the glaze, Van Eyck dabbed at the sticky glaze with his fingers or the palm of his hand. I observed a similar phenomenon in the Virgin’s red cloak in the Virgin and Child with Canon Joris van der Paele (1436, Bruges, Groeningemuseum). On top of an opaque red base tone (presumably containing vermilion), he applied organic red glazes in varying thicknesses. Under magnification, the glaze shows evenlyspaced lines broken up into small dots.21 These look like palm prints, which suggests that Van Eyck worked the surface of the red glaze with his hands. This pattern is visible across the whole surface of the cloak, but is especially obvious in the lightest

419

parts of the garment, where Van Eyck wanted the glaze to be the thinnest (fig. 28.1a). The Christ Child’s white cloth was painted on top of the cloak; where the cloth has become abraded, the raised texture of the red dots can be seen (fig. 28.1b).22 The swirling lines of dots in the cloak of the Deity Enthroned in the Ghent Altarpiece also suggest that Van Eyck dabbed the glaze with his hands, as described at the end of the paper. Jan van Eyck’s manipulation of a glaze with his hands or fingers is related to a technique that has been termed ‘blotting’.23 It thinned and evened out the paint, which might otherwise have shown brushstrokes or appeared too thick, gloopy, cracked or wrinkled. The method is described in several historical treatises. Armenini (1598) wrote: ‘veil the sketch evenly with a large brush of miniver; next pat it either with the palm of your hand or with a little wad of cotton wool covered with linen, so that the colour is uniform without any sign of a brushstroke.’24 It is most effective to apply the paint with a brush in a conventional manner, then use a cloth or hand to rub or pat the sticky paint, rather than using the fabric pad to apply the paint.25 The removal method is described by some historical treatises, including Armenini’s.26 He specifically suggested adding a little common varnish to the paint, which would speed drying and increase saturation.27 In Van Eyck’s Arnolfini Portrait, pine resin was identified in glazes which he also may have blotted with his fingers.28 Blotting has been observed in paintings from all over Europe, most frequently in those produced between the fifteenth and seventeenth centuries.29 It has been found in many Early Netherlandish paintings, for example, in works by Quinten Massys,30 and painters who worked in Leiden in the early sixteenth century.31 When a rag, piece of linen, or a fabric-covered pad was used to blot the paint, the weave of the fabric often created a regularly spaced pattern of dots. Sometimes this is visible within the glaze itself (as is the case with the handprints in Van Eyck’s paintings), or where the dots overlap adjacent areas of colour.

420

abbie vandivere

Blotting is not only restricted to red glazes; it has also been observed in green and blue passages. Copper-green glazes can also be viscous and difficult to spread thinly and evenly.32 Reconstructions of verdigris bound in various media showed that many of the same handling properties that were observed in red lake glazes also hold true for verdigris glazes.33 While the glaze made with untreated linseed oil left visible brushstrokes, heat-bodied linseed oil reduced their appearance, and the sunthickened oil left no visible brushstrokes. However, the sun-thickened oil could only be applied in a thick layer, and was difficult to handle. Perhaps this is the reason why Van Eyck may have used his fingers to blot the heat-bodied verdigris glaze of the woman’s green cloak in the Arnolfini Portrait, especially in the thicker parts of her sleeve.34 Van Eyck’s blotting technique – thinning the glaze with his hands – is consistent with his desire to mask his brushstrokes, create an even surface, and achieve a uniform colour. However, in this way, he created a microscopic surface effect. Individual Fingerprints Fingerprints on a painting’s surface not only literally reveal the hand of the artist at work but could also have a technical and visual function. According to Dunkerton and Billinge, dabbing with a finger ‘is almost an instinctive action when using oil paint,’ and Verougstraete and Van Schoute describe the fingers as ‘a sort of second tool, very natural to use, always at hand, cheap and fast.’35 Indeed, the slow-drying, viscous nature of the oil medium encourages a tactile relationship with the paint surface. Several Early Netherlandish paintings contain passages where fingerprints were used intentionally in order to give a particular texture, for instance, Hieronymus Bosch and his follower Jan Mandijn used their fingerprints to depict straw roofs,36 and sixteenth-century Leiden painters used their fingerprints to represent smoke, rocks, foliage and flowers.37 Numerous fingerprints have been found on paintings by Jan van Eyck, sometimes one or two in

isolation on one painting. Although they do not fulfil the same visual functions as the aforementioned examples, they must have been an intentional part of his working process. Most of the individual fingerprints are in brown areas, for instance, in a shadow and fur of the canon’s cloak in the Virgin and Child with Canon Joris van der Paele (fig. 28.2a).38 Fingerprints were also found in the shadow next to the dog’s foot in the Arnolfini Portrait,39 and also in the shadow behind the footstool in the Washington Annunciation (fig. 28.2b).40 It seems that in all these cases Van Eyck dabbed the translucent brown paint with his finger in order to create a more diffuse shadow or tone. By smudging or breaking up the paint on the surface, he allowed the underlying colour to shimmer through. I do not have the forensic skills to identify and compare fingerprints myself, but perhaps in the future they could provide some information about the attribution of various paintings. Scraping into Wet Paint The soft hairs at the tip of a brush are ideal for manipulating oil paint, but the other end of the brush could also be a useful tool for an artist. Using a blunt instrument like the end of a wooden brush handle to draw lines into wet paint is a technique that is usually associated with seventeenth-century painters like Rembrandt or Jan Lievens, but there are a number of earlier examples. Some examples predate Van Eyck, and have been found in so-called pre-Eyckian paintings.41 They also appear in the works of other Early Netherlandish painters, such as Quinten Massys and Geertgen tot Sint Jans.42 Some examples of Van Eyck’s use of a blunt instrument to scrape wet paint have been mentioned in the literature. In the background of the Arnolfini Portrait, he scraped overlapping lines into the dark brown paint to depict the bristles and handle of the dusting brush.43 In a similar manner, he created some of the white ruffles in Margaret van Eyck’s headdress,44 and accentuated the feathers in an angel’s wing in the Virgin by a Fountain (1439, Antwerp, kmska).45 Similar marks were found in

surface effects in paintings by jan van eyck

421

Figs 28.2a-b Individual fingerprints: (a) in a shadow behind St George’s leg in the Virgin and Child with Canon Joris van der Paele; (b) in the shadow behind the footstool in Jan van Eyck, The Annunciation, c.1434-36, oil on canvas transferred from panel, 90.2 x 34.1 cm, Washington, National Gallery of Art, Andrew W. Mellon Collection (inv. no. 1937.1.39)

several areas of the Virgin and Child with Canon Joris van der Paele. He scraped lines into the yellow paint of St Donatian’s brocade to depict the lines of gold stitching in the fabric (fig. 28.3a).46 These lines vary in direction, and in some areas it looks like he returned to manipulate the paint with the brush after the scrapes were made. There are also scraped lines in the translucent veil that drapes around his staff. Some of the lines towards the bottom of the cloth cleanly displace the wet dark brown paint and reveal the lighter brown paint underneath (fig. 28.3b). Further up, the lines are finer and have tapered ends, and the wet paint has been dragged softly beyond its intended boundary (fig. 28.3c). This suggests that the lines in the

lower part – where the veil is behind the staff – were scraped with a blunt instrument. In the middle part – where the veil drapes over the staff – the lines were applied with a fine brush when the paint in the adjacent areas was still wet. A combination of scraped and painted marks also appears in the red tiles in the foreground of the Virgin and Child with Canon Joris van der Paele. To create a speckled pattern, Van Eyck made small staccato marks in the red glaze, which revealed the brown paint underneath (fig. 28.4a). Remarkably, in two of the tiles directly below the Canon’s garment, Van Eyck did not create the stipples by scraping into the glaze; instead, he painted white dots on the surface, presumably after the glaze was

422

abbie vandivere

Fig. 28.2b

dry (fig. 28.4c). Perhaps he forgot to scrape these tiles at an earlier stage, or maybe this was a conscious decision. Regardless, it shows another example of how Van Eyck could use two means to achieve the same end result. We do not know what type of instrument Van Eyck used to scrape into the glaze. It is logical to assume that he used the wooden end of his paintbrush. Alternatively, he could have used an old brush with bristles that had dried clumped together, or another type of tool.47 Some of the marks in the tiles to the left of the carpet have rounded ends, and clean edges with glaze heaped up along the sides. This suggests that he used a blunt instrument like a brush handle to scrape the glaze while it was wet. Other marks have a more triangular shape,

‘feathered’ ends and vary in size (fig. 28.4b). This suggests that he used a brush, perhaps when the glaze was slightly drier. I made reconstructions to figure out what sort of tools made similar marks to the ones found on the Virgin and Child with Canon Joris van der Paele, and other paintings with stippled red glazes.48 Three different instruments – a brush, a pointed stick and a blunt stick – were used to scrape small lines in red glazes at different time intervals. The glazes were bound in several types of binding media, including untreated linseed oil, sun-thickened oil and boiled oil.49 The untreated oil retained the texture well, but also created visible brushstrokes. The glazes made with boiled and sunthickened oils tended to settle, and some of the

surface effects in paintings by jan van eyck

423

Figs 28.3a-c Scraping in the Virgin and Child with Canon Joris van der Paele: (a) lines scraped into the yellow paint in St Donatian’s brocade; (b) lines scraped into the bottom of the veil around St Donatian’s staff: (c) lines painted with a brush, where the veil drapes around St Donatian’s staff

marks (especially those made with the pointed stick) partially filled up with glaze soon after they were applied. The two types of marks observed in the Virgin and Child with Canon Joris van der Paele could best be replicated with a blunt wooden stick and a brush. As the glazes started to dry, it was more difficult to make marks with the brush, and the marks were not as ‘clean’; this seems similar to what was observed in some of the tiles in that painting (fig. 28.4b). The slow drying and viscosity of the oil medium made it possible to manipulate the paint in this manner. The tool displaces the glaze, which can be seen heaped up along the edges of the stroke. This differs somewhat from the sgraffito technique used in early Italian paintings, where the egg tempera paint was scratched off from the underlying gold, rather than displaced by the tool. Surface Effects in the Ghent Altarpiece Many of the surface effects described above – blotting, fingerprints and scraping wet paint – are present in the Ghent Altarpiece.50 Remarkably, they frequently occur alongside each other in certain panels. This may have an implication for authorship, or the work of different hands. Here, the artist will be referred to as ‘Van Eyck,’ although we are not certain whether these areas were painted by Jan van Eyck himself.

In one of the lunettes on the outside of the altarpiece there appears to be a combination of surface effects. Parallel lines in a dark brown shadow in the background of the lunette above the Virgin Annunciate might be fingerprints. Below, Van Eyck created the pages of the book by scraping into the thick yellow paint and revealing the brown layer underneath. The blunt ends of the lines and the ‘piling up’ of paint on either side suggests that it might have been executed with the end of his brush (fig. 28.5a). In the female donor panel below the Virgin Annunciate, a fine wavy pattern in the glaze of Elisabeth Borluut’s red cloak suggests blotting with a cloth or, more likely, a hand (fig. 28.6a).51 It is especially visible in the highlights, where presumably the artist tried to thin the glaze to reveal the lighter-coloured underlayer beneath. The open altarpiece also shows a multitude of different techniques. In the red glaze of the Enthroned Deity’s cloak there are some faint patterns of parallel lines that look like palm prints (fig. 28.6b). The prints are most visible in the orangey highlights in the upper half of his body, presumably where a red glaze was applied on top of a vermilion-containing underlayer.52 In the trim of John the Baptist’s cloak, Van Eyck appears to have scraped into the yellow paint to reveal a brown layer beneath. It is unclear whether the artist used the

abbie vandivere

a.

b.

surface effects in paintings by jan van eyck

425

c.

Figs 28.4a-c Tiles in the Virgin and Child with Canon van der Paele: (a) tile to the left of the carpet, probably scraped with blunt instrument. Inset: reconstruction of red glaze, scraped with a blunt instrument (wooden stick); (b) tile further to the left of the carpet, possibly scraped with brush. Inset: reconstruction of red glaze scraped with a brush; (c) tile to the right of St George’s foot stippled with white paint

426

abbie vandivere

a.

b.

c.

Figs 28.5a-f Scraping in the Ghent Altarpiece: (a) Virgin Annunciate, book in roundel above Virgin; (b) Singing Angels, parting in hair of angel in green on left; (c) Singing Angels, brocade clothing of singing angel; (d) Knights of Christ, fringe on red flag; (e) Adoration of the Lamb, wings of an angel; (f) Adam, hairs between navel and groin

surface effects in paintings by jan van eyck

427

d.

e.

f.

428

abbie vandivere

Figs 28.6a Blotting in the Ghent Altarpiece: (a) Elisabeth Borluut, proper right sleeve

bristles or the wooden end of his brush to manipulate the wet paint because the paint has settled somewhat and softened the edges of the strokes. In most of the other panels, the trims were applied in a more conventional manner with a brush and yellow paint, and are meant to imitate gold. Two surface effects were used alongside each other in the Singing Angels panel. The angel in the upper left has hair that extends over the sky, and within the hair are fingerprints and scraped curls. This shows that the paint must have been wet when both were applied (fig. 28.7).53 The parting of the hair of the angel in green suggests a

wet-in-wet application, possibly displacing the wet paint with a brush (fig. 28.5b). Parts of the yellow foliate pattern of the brocade of the angel in the foreground were scraped (fig. 28.5c). The yellow borders around the angels’ necks also show different application methods: the angel in green on the left has scraping in both a hatched and crosshatched pattern and the angel in red brocade next to him shows scraping in a lower layer with fine lines hatched in yellow paint on top (figs 28.8a-b). The border around the neck of the blue angel in the middle looks like it was stippled on with a brush (fig. 28.8c). This shows the variety of approaches

surface effects in paintings by jan van eyck

429

Figs 28.6b Deity Enthroned, cloak, below proper right shoulder

that Van Eyck used to differentiate the borders of each garment. A similar variety can be seen the flags in the Knights of Christ, where most of the fringes are painted in a conventional manner with hatched lines, but one has a scraped border. In addition to the lines mentioned above, there are small isolated cases of scraping into the paint throughout the inside of the altarpiece. These include: the curl of hair on top of a Pilgrim’s forehead, the blue fringe of a red flag in the Knights of Christ panel (fig. 28.5d), the wings of an angel in the Adoration of the Lamb panel (fig. 28.5e), and the hairs between Adam’s navel and groin (fig.

28.5f). In the last example it is unclear whether the hairs of the brush or the wooden end were used; the brown paint is so fluid and thin that a dry brush might easily displace it. Discussion In some cases, surface effects reveal steps in Van Eyck’s working process. The variety of ways he applied or removed paint from the gold trims of the angels’ garments suggests he might have been experimenting with how many ways he could create subtle differentiations of texture for each individual. This may also be true for the different

430

abbie vandivere

Fig. 28.7 Scraping and fingerprints in the Ghent Altarpiece: Singing Angels, hair of the angel at top left

ways he textured the tiles in the Virgin and Child with Canon Joris van der Paele. Alternatively, it could reveal something about his materials. Perhaps in the tiles he began by scraping the wet glaze with a blunt instrument (fig. 28.4a), but as the red glaze started to dry he found it more difficult to apply the texture – for instance, in the tiles where he appears to have used a brush (fig. 28.4b). It could have been so dry by the time he got to the area below the Canon’s garment that he was forced to depict the texture by applying white dots to the surface (fig. 28.4c). The evidence of blotting with hands and fingers also reveals something about the handling properties of his paint. Perhaps the medium he chose for his red glazes was a sort of compromise. Untreated linseed oil would have shown brushstrokes if applied over a large surface. Pre-polymerized oils would preserve fewer brushstrokes, but sometimes had poor handling properties. Blotting could help to solve both of these problems. It could eliminate the appearance of brushstrokes and thin the glaze sufficiently. It did, however, leave microscopic visual evidence of the working method as dots or lines in the glaze. These dots may not have been of great concern to Van Eyck; they are not easily visible to the naked eye, and seem mostly to appear on large works that were less frequently examined close-up. Surface effects are much more common in Van Eyck’s larger paintings than his smaller ones. The need to apply paint over a large surface may have brought certain challenges: applying glazes to large areas without visible brushstrokes, having areas dry before they were completed, and creating effects that were not necessarily possible with brushes alone.

surface effects in paintings by jan van eyck

431

a.

b.

c.

Figs 28.8a-c Gold trims in the Ghent Altarpiece, Singing Angels: (a) ‘scarf ’ around the neck of the angel on the left, showing scraping in a hatched and cross-hatched pattern. The trim beneath his neck was painted with small brushstrokes; (b) border of angel wearing red brocade, showing scraping in a lower layer with fine lines hatched in yellow paint on top; (c) border around the neck of the blue angel in the middle, probably stippled on with a brush

432

abbie vandivere

NOTES * This research was done as part of my PhD research within the project: The Impact of Oil: A History of Oil Painting in the Low Countries and its Consequences for the Visual Arts, 1350-1550. I would like to thank the museums that made their paintings available for study, the members of the Ghent Altarpiece project; the organizers of the Back to the Roots course, Doerner Institute, Munich; and Marjolijn Bol, TillHolger Borchert, Esther van Duijn, Kathrin Dyballa, Stephan Kemperdick, Indra Kneepkens, Jan Piet Filedt Kok, Anne van Oosterwijk, Jana Sanyova, Christiaan Vogelaar, Arie Wallert and Lidwien Wösten. I would also like to draw attention to the publications on the painting technique of Jan van Eyck written by Jill Dunkerton at the National Gallery, London. 1 Variations on these terms have been used in modern writings by: Panofsky 1971; Gombrich 1950; Bol 2012, p.18; De Mey 2008; Dunkerton 2000, p. 290; and Effman 2006, p. 18. 2 Quotations from Lemaire de Belges 1505, Paillot de Montabert 1829-1830 (see Effman 2006, p. 18), and De Heere 1559 (see Cuypers, Brinkman 1996, p. 18). 3 Myths and theories about the Eyckian binding medium are summarized in Roy 2000 and Effman 2006. 4 Unless otherwise mentioned, the observations are based on my examination with the naked eye and under magnification with a head loupe or stereomicroscope. The paintings were examined and photographed in the museum environment. During research on location, a portable microscope (Zeiss OpMi-1 86372) was used in most cases to examine the paint surface. Digital photographs were taken using a Canon EOS 40D camera and a Canon EFS 60mm macro lens. Because of time constraints and the fact that the surface of the paintings could not be touched, no scale bar could be included in the photographs. 5 For more comparative examples, see Vandivere 2013, pp. 113133. 6 For stipples on the wings of the Antwerp Virgin by a Fountain, see Depuydt-Elbaum 2003, p. 170. 7 Dunkerton 2008a: Van Eyck’s handling of oil paint. 8 Sun-thickening is a modern term. Historically, oil was allowed to polymerize and thicken using a variation of this method (Kneepkens 2012, pp. 18-20). A modern ‘stand oil’ is not a suitable equivalent, because it is usually made by boiling the oil in the absence of oxygen: see Saitzyk 1987. 9 Kneepkens 2012, p. 12; Bol 2012 pp. 135-136; Billinge et al. 1997, p. 41. 10 Billinge et al. 1997, p. 41 and White 2000, p. 104. The processing of the oil is not specified, so it could either have been used raw, or heat pre-polymerized: see Effman 2006, p. 24. 11 Gifford 1995, p. 8. 12 Dunkerton 2008a: Van Eyck’s handling of oil paint. A prepolymerized oil that was not exposed to high temperatures (for instance, a sun-thickened oil) cannot be easily identified with the analytical techniques that are currently available. Oils that have been heated to a higher temperature (heat-bodied oils) can, however be identified using GC-MS: see Billinge et al. 1997, p. 41. 13 For more information about glazes, see Bol 2012, pp. 123-169. 14 Both analyses are mentioned in White 2000, p. 104. 15 Dunkerton 2008a: Margaret’s red dress. 16 Reconstructions of red glazes in untreated and heat-bodied linseed oils were carried out by Kneepkens 2012, Dunkerton 2000, p. 292, and Vandivere 2013, Appendix 4d. 17 Kneepkens 2012, pp. 65-69, 135. 18 Billinge et al. 1997, p. 41 describes the consistency of paints made with pre-polymerized oils as ‘rather viscous and jelly-like’. 19 Kneepkens 2012, pp. 69-71, 85, 116-120. The boiled oil levelled and settled slightly less quickly, and a slight impasto could be

created; however, the brushstrokes were more visible than with the sun-thickened oil. 20 Dunkerton 2008a: Margaret’s red dress. Although here Dunkerton specifies that ‘Van Eyck blotted the sticky paint with his fingers or the palm of his hand,’ she admitted that it was difficult to know with certainty whether it was blotted using this method or with a cloth. See also Wolfthal 2003, p. 92. 21 The lines are mostly perpendicular to the brushstrokes in the underlying opaque paint layer, which might have made the lines spotty. 22 The surface has certainly suffered some abrasion and the lead white has presumably become more transparent over time, but perhaps this cloth was always intended to be semi-translucent. 23 ‘Blotting’ is not a historical term, but has been adopted by various authors, including: Wösten 2010; De Melo, Sanyova, Cruz 2011, Vandivere 2011, Kneepkens 2012, and Sheldon et al. 2012. 24 Olszewski 1977. Other historical sources about blotting are listed in De Melo, Sanyova, Cruz 2011, p. 2. 25 Different types of blotting reconstructions were carried out by: Wösten 2010; De Melo, Sanyova, Cruz 2011, Kneepkens 2012, Sheldon et al. 2012, and Vandivere 2013. Both methods of blotting give a similar visual effect, but the removal method reduces wastage of the red paint. 26 According to De Melo, Sanyova, Cruz 2011, pp. 1-2, some treatises are not as clear on whether the fabric pad was used to apply or remove the paint. 27 Olszwesky 1977, p. 275; De Melo, Sanyova, Cruz 2011, p. 4 and reconstructions pp. 7-8. 28 Campbell 1998, p. 184; White 2000, p. 104; Dunkerton 2000, pp. 290-291, fig. 13. 29 De Melo, Sanyova, Cruz 2011 describes the technique used in Portuguese paintings, but also lists other paintings that feature blotting (Table 2, p. 3). Sheldon et al. 2012 observed palm prints, fingerprints and blotting with a textile in several red lake glazes in paintings from Tudor Britain. 30 Dunkerton 2008b. 31 Wösten 2010 and Vandivere 2011, pp. 8-9. 32 Vandivere 2011, p. 9. For more information about green glazes, see: Woudhuysen-Keller 1995, pp. 65-69, and Van Eikema Hommes 2004, pp. 51-89. 33 Bol 2012, pp. 140-145. 34 Dunkerton 2000, p. 290, and White 2000, p. 104. 35 Dunkerton, Billinge 2005, p. 48, and Wolfthal 2006, p. 91. 36 See, for example Bosch’s Haywain Triptych, after 1510, Museo del Prado, Madrid, and Jan Mandijn’s Temptation of St Anthony, c.1550, Frans Hals Museum, Haarlem: illustrated in Bagley-Young 2007, p. 4. 37 Vandivere 2011, n. 61. For more examples, see Wolfthal 2006. 38 In the shadow of the Canon’s cloak, a fingerprint is partially underneath an adjacent area of red glaze, which proves that it is original. 39 Campbell 1998, p. 184, and Dunkerton, Billinge 2005, pp. 48, 59. 40 Gifford 1995, p. 87. Fingerprints in the Ghent Altarpiece will be discussed later in this paper. 41 Stroo 2009, pp. 226 and 230, and Vandivere 2013, pp. 115119. 42 For examples of scraping in paintings by Quinten Massys, see: Dunkerton 2000, fig. 5, pp. 289-290. For Geertgen tot Sint Jans, see Vandivere 2013, pp. 123-129. Reconstructions of Geertgen’s textured fabrics are described later in this paper, and in Vandivere 2013, Appendix 4d. 43 Dunkerton 2000, p. 290; Dunkerton, Billinge 2005, pp. 48, 57; illustrated in Campbell 1998, fig. 16, p. 184. 44 Dunkerton 2008a: Van Eyck’s handling of oil paint. 45 Depuydt-Elbaum 2003, fig. 13, p. 171. 46 For more information about Van Eyck’s gold-brocaded fabrics, see Van Duijn in this volume.

surface effects in paintings by jan van eyck

47 The blunt character of the lines mean that it is unlikely he used a sharp instrument, although there are a few examples in Van Eyck’s oeuvre where he used a pointed tool – like a stylus – to make incisions: see Van Asperen de Boer 1979, p. 175, and Billinge, Veroegstraete, Van Schoute 2000. 48 The inspiration for these reconstructions was the hairy texture on two paintings by Geertgen tot Sint Jans from the Gemäldegalerie, Berlin: St John the Baptist in the Wilderness, and the Virgin and Child. The results are included as Appendix 4d in Vandivere 2013. 49 The binding medium of the red glazes in the Virgin and Child with Canon Joris van der Paele have not been analysed. The reconstructed paints discussed here were made from madder lake bound in cold-pressed linseed oil (Kremer Pigmente 73054), sun-thickened oil (Kremer Pigmente 73011) and boiled oil (prepared by Kneepkens 2012 pp. 43-45: boiled oil minimal). The tools used to scrape the glaze were: a brush, the sharp end of a barbecue skewer, and the blunt end of a

433

barbecue skewer. Marks were made immediately and after 2 hours, 1 day and 4 days. 50 The images captured as part of the project Closer to Van Eyck: Rediscovering the Ghent Altarpiece (http://closertovaneyck.kikirpa.be) have been invaluable for visualizing the surface effects. 51 This description was based on observations made before the recent conservation treatment. The restoration revealed that the clothing of Elisabeth Borluut had been overpainted, and the paint in this area may not be original. 52 For the build-up of the Enthroned Deity’s red clothing, see Coremans 1953, pp. 73-74, and Brinkman et al. 1988-1989, pp. 36-37 and 46. 53 There is some doubt as to whether the paint in this area is original. The colour and texture of the brown paint that overlaps the sky differs from the rest of the hair. Hopefully the conservation treatment will provide more information.

PART III VAN EYCK IN CONTEXT: OTHER MEDIA, RECEPTION AND LEGACY

Fig. 29.1 St Christopher, wall painting, Mechelen, tower of St John’s Church

29

Mural Paintings before Jan van Eyck. A Remarkable Discovery from around 1400 in St John’s Church in Mechelen Marjan Buyle and Anna Bergmans

ABSTRACT: During restoration work on the gothic Church of St John in Mechelen, traces of wall paintings were found behind the eighteenth-century organ. Two monumental portrayals were discovered, together with architectural polychromy and painted stone consoles. The paintings, which represent St Christopher carrying the Christ Child and St George with the dragon, date back to 1400 and show great artistic quality. They are unique testimony to the painting tradition before Jan van Eyck. During conservation work, analyses of the painting technique and the materials were made. The historical and art-historical examination of the paintings aims to place them in the historical context of ducal Burgundy and the artistic milieu of around 1400 in the Low Countries.

—o— When the mural paintings were discovered in the tower of St John’s Church in Mechelen in the autumn of 2008, the importance of the finding soon became apparent (figs 29.1, 29.3).1 The paintings were almost entirely covered in thick layers of whitewash2 (fig. 29.2). The trend for stripping the walls in church interiors from the mid-nineteenth century up to the 1970s has meant that discoveries of wall paintings are limited to parts of buildings where whitewash was, fortuitously, not removed, such as the space behind the organ. An ensemble of mural paintings dating from around 1400 is obviously extremely rare and valuable. Apart from the fact that it literally enriches Flanders’ heritage, the finding is also of great importance in terms of art history.3

Two Monumental Portrayals The north wall of the tower features a very large image of St Christopher carrying the Christ Child across the river. On the left, his companion the hermit waits with a burning lantern. The scene is set in a landscape with a river, rocky banks, and trees (fig. 29.1). On the south wall St George appears on horseback, resplendent in full armour, piercing the neck of a dragon with his lance (fig. 29.3). On the left the princess watches, kneeling devoutly with her lamb. A fortified city is visible in the background. On the right is a cluster of elm, oak and lime trees. The mural paintings are situated in what now appears to be the first floor of the tower; however, the present wooden floor was only put in place after the eighteenth-century organ loft was installed (fig. 29.4b). In the Middle Ages the mural paintings would have been visible to everyone who entered the church through the west door (fig. 29.4a). Each figurative image is set in an illusory alcove with a pointed arch, outlined in black paint. The alcoves are now 5.74 metres high and 3.96 metres wide; the very bottom of the image is hidden by the floor. In addition to the two monumental portrayals of St Christopher and St George, this space also contains interesting architectural polychromy and beautifully carved consoles with preserved colour-

438

marjan buyle and anna bergmans

Fig. 29.2 St George and the Dragon, wall painting, detail, the princess, shown during the uncovering of the painting

ing (fig. 29.5). At the moment, questions of how long the paintings were visible and why, and when they were covered over, cannot be fully answered. After the Counter-Reformation many medieval paintings were no longer deemed suitable and had to disappear. The apotropaic properties associated with St Christopher and the related excesses of devotion also led to the disappearance of his image. The major wear displayed by the mural paintings suggests that they were visible for a relatively long period. The text on the vault may provide some clue as to when the paintings were plastered over. It refers to damage in 1580 and restoration in 1602. One hypothesis is that at that time the vault and the wall surfaces were completely whitewashed. If so, these mural paintings would have been visible from around 1400 to 1602.

Layer Structure Immediately after the tower at the west end of the church was built the paintings were applied to the inner brick wall on a layer of plaster consisting of the usual lime and sand. First the painter filled the ‘alcove’ with a uniform base of red ochre. Onto this he painted his underdrawing. He then applied the colours, with the all-important nuances and shading. He used fairly thick, black outlines to define the figures. In the case of St Christopher it appears that the red background, which follows the outlines rather untidily, was only applied after the figures were completed. In any case the stencilled flowers, which turn the background into a kind of cloth of honour, were painted last.

mural paintings before jan van eyck

Fig. 29.3 St George and the Dragon, wall painting, Mechelen, tower of St John’s Church

439

440

marjan buyle and anna bergmans

Figs 29.4a-b (a) Simulation of the original situation in the tower; (b) simulation of the present situation in the tower

Underdrawing Hitherto, there has been little interest in the subject of underdrawings in mural paintings. There are various reasons for this: issues of visibility (the technology available to research into panel paintings is of little or no use at all in this case); the fact that only a small number of mural paintings survive from the late fourteenth to the fifteenth century, making a comparison of works by the same master impossible; the substantially less interest in underdrawings in mural painting north of the Alps compared to the spectacular sinopia in fresco painting south of the Alps; the lack of specific publications on the subject (the presence of underdrawing is at best mentioned in passing in restoration reports); and the lack of a methodology for the exact record-

ing of underdrawing in restoration documentation for mural paintings.4 In this case, wear of the paint layers has left the painted underdrawing very visible. Painted with a brush in jet black, it is purely linear, without any hatching or shading. An attempt was made to document the underdrawing in full size using tracing paper, but this proved not to be feasible, not only because it is still partly covered by the paint layers but also because the difference between the outlines and underpainting is quite subtle. In particular, the painting’s large dimensions made it difficult to keep such a large sheet of film in place: while the drawing is being made, the film must be lifted repeatedly to perform certain checks. Thus we were forced to limit the tracing of the underdrawing to

mural paintings before jan van eyck

the most interesting parts of the image, particularly those places where pentimenti were visible. The underdrawing in Mechelen is strikingly visible even to the naked eye. It is so extensive that in current views it plays a major role in the readability of this ancient and worn painting: now the underdrawing stands in for the original composition in certain places, which was probably not the original intention. The horse’s mane in the underdrawing is so detailed that one might wonder whether the intention was for these black lines to filter through the mane’s nervous and pasty white painting. The underdrawing is highly skilled. Precise, and painted without any hesitation, it betrays the hand of a master experienced in monumental compositions. Only the outlines, folds and facial characteristics are drawn in; there is no suggestion of volume

441

or shading. The underdrawing is painted in fairly thick, black lines. In some places it is difficult not to confuse it with the outlines that were painted at a later stage. In executing the underdrawing the artist eschewed technical aids and drew freehand, straight on to the wall. Only the figures and the large animals have underdrawings. None was found on the landscape, trees or rocks. Strikingly, the princess’s lamb also displays no trace of underdrawing. There are many pentimenti; however, they do not relate to the content and only involve the positioning of and alterations to the heads and hands in particular and sometimes the folds. Obvious pentimenti are visible in the position and arched form of the shield on St George’s back. Alterations to the Christ Child’s head and hands can also be seen.

442

marjan buyle and anna bergmans

Fig. 29.5 Corbel

Different Hands Our working hypothesis is that at least two painters were involved. The first was responsible for the architectural polychromy on and around the vault ribs, and the polychromy of the four consoles. He also painted the alcoves with their stencilled decorations. The second painter filled the alcove with a red background (unless the first painter was also responsible for this) and painted his black underdrawing on top. He then continued to add different layers of colour with sparing use of technical resources such as the aforementioned stencils and a strip of applied brocade. The notion of there being two painters is confirmed by the fact that they used different colour palettes and pigments. Strangely enough, the painter responsible for the architecture and landscape used more expensive pigments, such as pure vermilion for the red areas. The painter of the

figures, however, chose pigment mixes to paint the beautiful red areas such as St George’s stunning cloak. He made only sparing use of more expensive pigments, such as lead-tin yellow type I and indigo, which he reserved for important figures and smaller but important parts: the Christ Child’s yellow halo and his blue garment, the princess’s blue dress. The painters used completely different styles to paint faces (Christopher and George compared with the consoles). This division of labour should not surprise us and fits the occupational structure of medieval craftsmen perfectly, in which a large degree of specialization and an extensive division of labour was common in craft guilds. Colours and Pigments The importance of colour in a medieval mural painting must not be underestimated. The association of colour with lustre and light presupposes a

mural paintings before jan van eyck

direct link with divine revelation. Colours and light in medieval church interiors are never accidental but help create the illusion of a Heavenly Jerusalem. Analyses of the pigments and stratigraphies produced interesting results.5 Red earth (iron oxide red) was used for the red base coat. The underdrawing was painted in charcoal black. St George’s silver armour, now visible as a grey layer, contains tin leaf and minute traces of lead. In order to obtain specific shades of colour, the artist mixed several pigments together: in St George’s stunning cloak vermilion, red lead, red ochre and calcium carbonate were identified. The work becomes even more refined as he built up the colours in two layers: the light green on St George’s saddle consists of a first layer of lead-tin yellow type I and a second layer mixed from lead tin yellow type I, lead white and copper green. In order to obtain dark green he used the same mixture, though without the yellow base, as is clear from the sample’s stratigraphic cross section. In the image of St Christopher the light blue colour used to paint the Christ Child’s garment was obtained by mixing indigo with lead white (fig. 29.6). The same mixture was probably also employed in the St George painting, for the princess’s dress, which was then enhanced with white stencil designs and with relief decoration on the waistband. The light yellow used on the Child’s halo is lead-tin yellow type  I. St  Christopher’s purplish-red garment is vermilion mixed with red ochre. Lastly, the pigment used for the background with the stencilled flowers is red earth (iron oxide red), the same as was used for the base coat. Condition The general state of preservation of the mural paintings in St John’s Church in Mechelen is better than average, because the configuration of the figures has been almost entirely preserved and they have not previously been restored. Thus no old restoration products or overpainting are present. Fortunately, the few large missing sections,

443

caused by previous damage and holes in the walls, are not in any important areas. The four faces, for example, of Christopher, the Christ Child, George and the princess, are remarkably well preserved, a result of the apparent absence of losses and the fact that they are created from successive paint layers and therefore exhibit less wear. The wear and thinning of the paint layer is present over the whole of the painting, though this in no way impedes one’s ‘reading’ of the images. Wear to the top layer means that the underdrawing has been revealed once more in most parts of the image. Here and there the poor condition of the paint layer is regrettable. Artistic Context The wall paintings of SS George and Christopher are of exceptional importance to the history of art from around 1400. We refer more specifically to art from between circa 1380 and circa 1420, executed in what has come to be known collectively as the ‘International Style’ and which manifested itself in all areas of artistic production. In recent decades, art from around 1400 has attracted much scholarly attention, occasioned by several major exhibitions.6 However, such major events have largely ignored the art of wall painting. Nonetheless, there is a demonstrable relationship between wall painting, miniature painting, panel painting and other forms of visual art. There are no known archival sources for the wall paintings in Mechelen. But comparative research does make it possible to place the works in art-historical perspective. Around 1400, artistic output in Flanders and Brabant was closely associated with production at the royal court in Paris and the ducal court of Burgundy. Since the signing of the Peace of Ath in 1356, Mechelen and Antwerp had been ruled by Louis of Male, Count of Flanders. In 1369, his daughter Margaret married Philip the Bold, youngest son of the French king and first of the Valois dukes of Burgundy, who on Louis’s death in 1384 became Count of Flanders, Franche-Comté, Rethel, Antwerp, Mechelen,

444

marjan buyle and anna bergmans

Fig. 29.6 St Christopher, detail, St Christopher carrying the Christ Child on his shoulder

mural paintings before jan van eyck

Artois and Nevers.7 On 21 March of that year, the ducal couple made their Joyous Entry into Mechelen.8 Extensive research has been carried out into the representative function and the significance of the artistic production at these magnificent courts and the travelling artists who worked at them. Movable and immovable works of art adorned the regal residences as well as religious and public buildings. Wall painting also played a prominent role, as is apparent from archival sources.9 Related Wall Paintings Comparable to some extent to the wall paintings in Mechelen is the slightly earlier St Christopher from the collegiate church of Semur-en-Auxois, transferred to the municipal museum. The attention to detail, the naturalism, the vibrant colours and the expressive quality of the figures make this a very important example of Burgundian painting. Based primarily on resemblances to the cartoons for the Apocalypse Tapestry series in Angers, Fabienne Joubert dates this wall painting to 13701375 and attributes it to Jean de Bruges, who worked for Philip the Bold and is mentioned in the accounts for 1371 and 1372 of the ducal court in Dijon.10 In Bruges, a remarkable late-fourteenth-century depiction of St George fighting the dragon adorns the east wall of a room in the house at Spinolarei 2, as part of a complex iconographic programme.11 This is a more static portrayal than the one in Mechelen and is rendered in full profile. In a series of niches below the painting of St George are ten male figures representing the virtues. On the south wall is an unusual triumph of fifteen, rather than the customary nine, heroes.12 This imagery has been associated with the crossbowmen’s guild, which used to meet in the neighbourhood.13 Around 1400, SS Christopher and George adorned a wall of the Mary chapel in the Church of Our Lady in Halle. The now lost murals are known through various copies.14 Below an architecturally articulated baldachin, Christopher wades through

445

the water with a blessing Child on his shoulder; in a rocky landscape, George poses victoriously atop the dragon he has just pierced with his lance. His silhouette is not unlike that in the wall painting in Mechelen: broad-chested, with a narrow waist. He is clad in armour and a surcote with a segmented belt worn low. The ensemble in St Catherine’s Church in Duisburg (Tervuren) is an excellent example of a high-quality wall painting from the Low Countries from around 1400.15 As has recently been pointed out, there are some striking similarities between these monumental paintings and the panel from Kortessem now in the Royal Museums of Fine Arts in Brussels (inv. no 4883).16 The ensemble in Duisburg testifies to the close artistic connections with the circles of Parisian and Burgundian courtly art and to the intense exchanges of artists and models that must have taken place. In St Rumbold’s Church in Mechelen, the niches behind the Baroque altar of St Anne, in the east wall of the south transept, feature some early fifteenth century paintings of saints.17 St Alexis and St Dorothea are represented against a red background. A third saint in the same series, namely John the Baptist with a lamb, is presently visible in the church, to the right of the altar.18 The Flemish miniature art of that same period was first subjected to extensive study in the context of the exhibition and colloquium in 1995.19 More recently, the Royal Institute for Cultural Heritage (kik-irpa, Brussels) published the first part of a work on ongoing research in the Low Countries into contemporary panel painting.20 The wall paintings in Mechelen may be placed in this artistic context. Admittedly, there is a considerable difference in technique with the other contemporary arts, particularly in terms of the luxurious application of gold leaf21 and the plentiful use of various relief decorations22 and engraving techniques on sculptures and panels. Only the waistband of the princess incorporates applied brocade and gold foil has not been applied.

446

marjan buyle and anna bergmans

Costume Insofar as dating the Mechelen wall paintings is concerned, fashion provides important clues. Around 1400 clothing from was characterized by an unprecedented variety of types and designs of attire, subtle combinations, bright colours and striking fabrics; in sum, it was marked by a richness of invention.23 Dagged sleeves, collars, shawls and headgear made for truly bizarre effects. Men’s fashion was even more diversified than women’s. The costumes in Mechelen are entirely in keeping with this fashion. The silhouette of St George is quite pronounced and typical of the late fourteenth century: a broad chest and very narrow waist.24 Around his hips he wears a heavy segmented belt and his surcote has wide sleeves with edges cut into dagges (fig. 29.7). He is portrayed as a prominent knight. The rarity and costliness of the materials and pigments used in the production of certain garments meant that such items were generally reserved for the wealthy elite. This was the case, for example, with the belt with little bells that were made of gold.25 The helmet with mail collar and open visor, finally, was part of the standard armour of a knight and appears on numerous representations of the saint from this period, such as the statuette of St George in the altarpiece by Jacques de Baerze (Dijon, Musée des Beaux Arts). The princess wears a blue gown with opulent brocade patterns and an ermine-lined collar. Its wide sleeves hang all the way to the ground. The fashionable high waist is accentuated by a belt richly embellished with relief decoration (fig. 29.8). St Christopher’s attire is of an entirely different order. He is clad in a plain red tunic under a wide white coat with green lining in a flowing fabric that is draped across his shoulders and flutters about him. It lends him a sense of dynamism that contrasts with the static landscape.26 Style The underdrawing has no depth and the perspective of the composition is worked out entirely in the paint layers, particularly in the folds of the

garments and the expressive detail of the facial features. The flesh tones and faces are meticulously constructed in successive layers of colour. Certain vestimentary details have been highlighted with stencil patterns, as in the splendid red brocade surcote worn by St George. The broad folds and shadows are already worked out in the red paint layer, while both stencilled brocade patterns are flatly applied. They are painted yellow, but are clearly supposed to evoke gold brocade. The princess’s blue gown was also decorated with a stencilled white fleur de lis pattern originally. The voluminous effect of the folds in the fabrics is expertly executed and produces quite a sculptural effect. This is particularly noticeable in the well preserved sections of the painting, such as in the greens and whites of the flowing coat to the right of St Christopher. The rather abstract volume of the oak, lime and elm leaves has been worked from dark to light: from black to dark green, with pale green highlights. In addition to these technical, stylistic details, the rendering of the landscape is also characteristic of the art of around 1400. The combination of a realistic landscape with abstract background features, as in the painting of St Christopher, is also commonly encountered in contemporary miniatures. These imaginary settings are characterized by a sense of perspective that is still naive. Analysis of miniatures has shown them to possess quite specific properties: saturated colours, modelling through the application of colour, a preference for realistic details and for the representation of textures, fabrics with brocade patterns that also existed in reality, colour fields accentuated by clear contour lines.27 All these elements also appear in the wall paintings in Mechelen. The term ‘realism’ does not apply in a strict sense to art from this period: these paintings are not truthful representations of reality as it presents itself to us. The proportions of the trees, rockery and buildings in relation to the human figures are entirely fanciful, for example. On the other hand, all the details – including the leaves on the trees,

mural paintings before jan van eyck

447

Fig. 29.7 St George and the Dragon, detail, St George’s long dagged sleeves and the cluster of trees

the feathers on St George’s helmet, the foliated decoration of his shield, and his horse’s mane – are rendered quite realistically. The realistic details in the expressive and energetic visages of the two saints are equally striking. Other salient features include the fluidity of the draped attire and the wonderful suggestion of airiness. Christopher’s cloak wafts in the wind; George’s garments are agitated by the movements of himself and his horse as he spears the dragon. Such details add to the dramatic tension. In terms of composition and iconography, the paintings are akin to miniatures produced at the courts of Burgundy and France around 1400, particularly those by the Boucicaut group.28 The Boucicaut Master probably came from the north but he worked in Paris. His oeuvre and the wall paintings in Mechelen are closely related in terms of the representation of space, the plastic detailing

and the iconographic modelling. The representation of St George on horseback, an image that would come to full prominence in the course of the fifteenth century, as in the work of Rogier van der Weyden, would also appear to derive from the miniatures of the Boucicaut Master.29 The shape and structure of the trees are identical. On this basis, the paintings may be dated to the first decade of the fifteenth century. The placement of the figures in the image plane and their position inside the tower, the modelling of the individualized faces and the realistic details are all indicative of a highly talented artist. As such, they also substantiate the view that artistic talent is not incompatible with the use of mechanical procedures such as stencilling. We conclude that the paintings discovered in St John’s in Mechelen belong to the so-called International Style of around 1400, with strong

448

marjan buyle and anna bergmans

Fig. 29.8 St George and the Dragon, detail, the princess in an opulent blue gown with white stencilled brocade patterns, ermine collar and fashionably high waist accentuated by a splendid belt

mural paintings before jan van eyck

influences from Parisian art. The paintings produced by artists from the Low Countries during this era are quite distinct. How exactly these artists influenced one another is a topic for further research, but artists are known to have travelled around, taking their models with them.30 In this context, we refer to the striking similarity between the head of St George and that of a wild man in a drawing from around 1400 attributed to Jacquemart of Hesdin, a miniaturist from Artois who worked at the court of John, Duke of Berry, in Bourges. The drawing is part of a sketchbook consisting of six panels of prepared boxwood.31 The inclination of the two heads, their profile with a square forehead, the long, heavy nose and the eye with broad brows are all very similar. Moreover, the hairline of the wild man and the rim of George’s helmet correspond almost precisely. The wall paintings from Mechelen add to our art-historical insight into painting from the period in question. The fact that these paintings were not movable objects but an integral part of the interior of a building makes their discovery all the more interesting. The question remains: were they produced around 1400 by a local artist from Mechelen? An Artist from Mechelen? It goes without saying that the presence of the Burgundian ducal court acted as a catalyst for artistic production and exchange in a prosperous town such as Mechelen. Inside the city gates, no fewer than five sizeable Gothic churches were built: St Rumbold’s, St John’s, the Church of Our Lady Across the Dyle, St Catherine’s and the former Church of St Peter and St Paul. The new Schepenhuis or Aldermen’s Hall reflected the city’s municipal pride and autonomy.32 The special political status of the city, first under the counts of Flanders and subsequently under the dukes of Burgundy, undoubtedly provided a stimulus for Mechelen to flaunt its wealth, prestige and power. The construction of St Rumbold’s, for example, may be seen as a statement confirming the independence of this enclave in Brabant.33 The concept

449

of the new choir was modelled after that in the cathedral at Amiens, with an ambulatory and seven radiating chapels.34 Above the entrance to the north transept, there used to be a stained-glass window with portraits of Louis of Male and his wife and daughter.35 Artistic production is also known to have been stimulated by events such as visits by rulers, joyous entries and other festive occasions. The accounts of the court of Burgundy for 1413 record a payment A maistre Vranque, paintre, demourant à Malines, pour paindre et faire la figure de mademoiselle Katherine de Bourgogne, fille de MdS, paié comptant… VI fr. XV s.36 In other words, a painter identified as Master Vrancke, who used to work in Mechelen, was paid for producing a portrait of Catherine of Burgundy, daughter of John the Fearless. This must have been quite a significant assignment, and the obvious implication is that there was a highly qualified painter working in Mechelen. The hypothesis put forward by Bella Martens that this otherwise unknown painter Master Vrancke may have been none other than the famous painter Meister Francke, who is first mentioned in Hamburg in 1407, has never been substantiated.37 It was Emmanuel Neeffs, who also studied the Burgundian accounts published by De Laborde, who suggested that the portrait painter may be identified with a Master Vrancke from Mechelen. But which one?38 Few late medieval paintings produced in Mechelen have been preserved. Archival sources do, however, confirm that from the early fourteenth century the city had painters on its payroll. They used to paint banners and emblems, adorn public buildings, and create festive decorations.39 In the period around 1400, one name sticks out in the urban accounts: Vrancke van Lint, who is mentioned from 13 October 1386 onwards in public works entries as a city painter.40 Neeffs published some valuable information about the work created at Vrancke van Lint’s workshop.41 The payments are recorded very clearly in the city accounts.42 Master Vrancke decorated wind vanes, banners and coats of arms; he carried out a paint job at the

450

marjan buyle and anna bergmans

premises known as the Huis van Bornem (13941395), polychromed a sculpture in the facade of the Vleeshuis (1399) and created decorations for the city gates.43 He also produced an altarpiece for the Schepenhuis (1404-1405).44 On the occasion of the funeral mass for Margaret III, Countess of Flanders, held in St Rumbold’s on 11 April 1405, he decorated the catafalque and executed the polychrome decoration of seventy-four shields with the arms of Burgundy and Flanders.45 Shortly thereafter, on the occasion of the Joyous Entry of the Duke of Burgundy, John the Fearless, on 23 April 1405, he decorated eighty-four coats of arms. In 1405, he received payment for his work, for fabrics and for painting materials when Mechelen joined the duke in defending Flanders against an invasion by the English.46 He further decorated the fireplace of the council chamber in the Schepenhuis with gold, silver and azurite (1408-1409). Eight years later, finally, Vrancke van Lint produced twentyfive medallions for the antependium of the oratory of the magistrate. This lengthy series of painting assignments coincides with the art-historical dating of the wall paintings in the tower of St John’s. Mechelen’s municipal accounts provide evidence of a considerable artistic output of ephemeral as well as enduring works of art, as was customary.47 In fact, in the fourteenth-fifteenth centuries, no such distinction was made in the context of artistic production. The available documentary evidence would suggest that Vrancke van Lint was the only known local artist to execute such commissions towards the end of the fourteenth and the beginning of the fifteenth centuries. Moreover, the municipal accounts indicate that his production was quite prestigious: clearly his services were sought to enhance the splendour of the city and the court of Burgundy. On this basis, we hypothesize that he was also commissioned to create the wall paintings in St John’s. Be that as it may, the arrangement of the more than life-size figures in the compositions and within the tower space would certainly suggest that the artist was accustomed to executing this kind of monumental work.

Patron and Purpose Neither concrete data nor archival sources are available to tell us who commissioned the wall paintings. Any hypothesis on this matter must therefore be based on contextual clues. The location of the paintings is crucially important in this respect: the intention was clearly for the saints to be visible to the faithful. The placement of St Christopher suggests that the west portal was the main entrance to St John’s. Hence St George on the opposite wall occupied an equally prominent position. This has a special significance. In the absence of source materials, the specific functions of St John’s tower and of the portal could not be determined.48 Nonetheless, there is clearly a general underlying meaning, as described by Durand de Mende.49 The purpose of the tower is defensive. The church entrance is an important sacred space that provides access to the Heavenly Jerusalem and thus is associated with Christ. The representation in this location of two saints who are deemed to ward off and combat evil ties in very well with the medieval visual idiom. Given that St George was the patron of the nobility, his painting may have been commissioned by a member of that class. The House of Burgundy, too, had adopted George as its patron. And Mechelen’s schuttersgilden or ‘armed guilds’, whose duties included responsibility for their city’s safety, are also possible candidates. The Oude Voetboog, oldest of the town’s crossbowmen’s guilds, had been in existence since before 1315.50 As membership was for life and numbers were restricted, a Jonge Voetboog or ‘young’ crossbowmen’s guild had been established as a kind of recruitment pool. It is not entirely clear when the latter association was set up, but it most probably already existed in the fourteenth century. Both guilds took St George as their patron.51 The Mechelen schutters fought for John the Fearless, as well as his son and successor Philip the Good, on one occasion against an English army during the siege of Calais.52 St Christopher is a popular saint who provides protection against a sudden death. His cult was

mural paintings before jan van eyck

widespread and his image appeared in many late medieval church interiors. On the other hand, Christopher was also the patron saint of Mechelen’s kolveniersgilde or arquebusiers’ guild, though it is uncertain when this militia was established. The earliest documents are the letters from Philip the Good dating from 1453 in which he requests the magistrate of Mechelen to send eight kolveniers to reinforce the ducal army in Lille and help crush a rebellion in Ghent.53 Though their weaponry had been in use since the early fifteenth century,54 the kolveniers were first included as a guild in the city accounts only in 1504.55 On 23 April 1405 John the Fearless was officially welcomed in Mechelen as the city’s new ruler.56 Such festivals were known as Joyous Entries, an event in which the militias traditionally played a prominent role. John the Fearless had taken part in the crusades and owed his nickname to his bravery during the Battle of Nicopolis in Bulgaria (1396). To raise the necessary funds, his father Philip the Bold borrowed considerable amounts, with Mechelen also contributing its share.57 John was spared from decapitation by Sultan Bajazet in return for substantial ransoms. Already venerated as a hero for his military exploits in the service of Christendom, he further acquired the aura of a martyr on account of his incarceration in Turkey.58 His release was made conditional upon the payment of a large sum of money, to which the people of Mechelen again contributed substantially.59 Upon his return, and notwithstanding the defeat suffered at the hands of the Turks, he was given a hero’s welcome. He and his father visited several towns and cities, including Mechelen, where they were showered with gifts.60 For all of the aforementioned reasons, the wall paintings in St John’s could be interpreted as a salute from the town militias to their sovereign. However, further clues in this respect are lacking. The fact that the paintings are located in St John’s, the patron saint of John the Fearless, is in itself not a decisive argument. Nonetheless, St George could be argued to have been portrayed as a nobleman

451

and as a true crusader, covered in St George’s crosses. Flags and banners bearing his image played an important role, not only in wars, but also in grand official ceremonies in centres of political, spiritual or economic power.61 St  George is the embodiment of chivalrous virtues and hence numerous sovereigns since the fifteenth century had themselves portrayed in his guise. Jean II le Meingre, better known as Boucicaut, Marshal of France and crusader, is believed to have had himself portrayed as St George.62 It is therefore conceivable that John the Fearless is represented as a miles Christi in the wall paintings in Mechelen. In that case, the painting would tie in with the many other fifteenth-century city ornamentations where the sovereign is represented as an antique or a biblical hero.63 On the other hand, one should not lose sight of the general apotropaic properties of the two saints, both of whom were deemed to ward off evil. It is particularly unfortunate that the small coat of arms above St Christopher is poorly preserved and hence illegible, because a heraldic interpretation may well have facilitated our reading of the painting. Conclusion It is quite evident that the newly discovered wall paintings are unique. There are very few wellpreserved wall paintings from around 1400, and the fact that these examples have never been restored or overpainted makes them all the more exceptional. They are, moreover, relatively complete and their artistic quality is outstanding. Stylistically and costume-wise, the images tie in with the International Style of around 1400. The wall paintings add an exceptional page to the history of art from the Low Countries around 1400 and complement our understanding of the better known and more intensively studied panel painting and miniature art of the same period. The fact that these paintings are inextricably connected with the building in which they were discovered implies that they were produced in situ, and, moreover, by a competent artist or workshop.

452

marjan buyle and anna bergmans

NOTES * We are most grateful to Wim Blockmans, Thomas Coomans, Ilona Hans-Collas, Henri Installé, Mireille Madou, Cyriel Stroo and Dominique Vanwijnsberghe for their stimulating feedback. 1 A first preliminary report of the discovery was published in: Buyle 2008. 2 The uncovering and conservation-restoration work was carried out by Marjan Buyle, Els Jacobs and Philippe Schurmans of the Flanders Heritage Agency. 3 The full report of the conservation-restoration work and the technical, material and art-historical research was published in Relicta. It is also available for consultation online via OAR: Buyle, Bergmans 2013. A French article is published online: Buyle, Bergmans 2012. The art-historical study has been published in the Antwerp Royal Museum Annual 2010: Bergmans, Buyle 2010. 4 This methodology is currently being developed as a handbook for restoration documentation for mural painting restoration, by the Flanders Heritage Agency. 5 Analyses of the pigments were performed by Marina van Bos and Ingrid Nijs (KIK-IRPA). The full analysis report is published as a supplement to Buyle, Bergmans 2013. 6 For example, Cologne 1978, Paris 2004, Dijon 2004, New York 2005, Nijmegen 2005, Luxembourg/Budapest 2006, Strasbourg 2008 and Rotterdam 2012. 7 On the historical and cultural developments during the Burgundian era, see in particular Prevenier, Blockmans 1983 and subsequent spin-offs. See also the recent synthesis: Boone 2011. 8 Van Doren, 1, 1859, p. 75. 9 Nys 2004. 10 Joubert 1992; Dijon 1992, pp. 156-157; Dijon 2004, pp. 294295. 11 Buyle 1995. 12 See Van Anrooij 1997. 13 Noad 2002 (with bibliography 1995-2002). 14 Bergmans 1998, pp. 138-142, 313-314. 15 Bergmans 1988; Bergmans 1994; Bergmans 1998, p. 333. 16 Stroo 2009, p. 249. 17 Bergmans 1998, pp. 233-234, 325. 18 Bergmans 1998, p. 325. 19 Smeyers, Cardon, 1995. See also the recent synthesis by Deneffe 2011. 20 Stroo 2009. 21 Baert 2009. 22 Geelen, Steyaert 2009. 23 Blanc 1997, p. 9. 24 Martin 1967, pp. 73-94. 25 Fingerlin 1971, n. 538 and p. 160. 26 Cf. the analysis of the Boucicaut Master’s Christopher by Blanc 1997, pp. 104-106.

27 Smeyers et al. 1993, p. 175. 28 Meiss1968; Châtelet 2000. 29 Strasbourg 2008, pp. 97-99. 30 Scheller 1995. For a discussion of ‘Flemish’ miniature art, see Vanwijnsberghe 2011. 31 New York, Pierpont Morgan Library, Ms M. 346, fol. 3v; Scheller 1995, pp. 218-225. 32 Van Uytven 1991, pp. 75-80. 33 Blockmans 2010, p. 424. 34 Coomans 2009, p. 15. 35 Neeffs 1876, p. 21; Laenen, 19342 , 339. The window was removed in 1767. 36 After De Laborde, 2, 1, 1849, p. 97. 37 On Meister Francke, see among others Martens 1929; Hirschfeld 1970. 38 Neeffs 1876, p. 501. 39 Neeffs 1876, pp. 89-100. 40 On city painters, see Wisse, 1998. 41 Neeffs 1876, pp. 97-99. 42 SAM, registers of the city accounts for 1386/1387-1416/1417. They were also registered in the chronicles by J.B. Rymenans and in Chronologische aenwijser by Gyseleers-Thys (SAM Oud Archief, Fonds CC Extrait des archives). 43 SAM, R. 1064 (1397-1398), f. 107, 123v; SAM, R. 1072 (1404-1405), f. 156v. 44 Neeffs 1876, p. 98. 45 SAM, R 1072 (1404-1405), f.164v. 46 SAM, R 1072 (1404-1405), f. 169. 47 Neeffs 1876, pp. 89-142. 48 See also Kuys 2006. 49 Durand de Mende ed. Barthélemy, 1854, p. 22. The Rationale divinorum officiorum of William Durand of Mende: A New Translation of the Prologue and Book One, New York Columbia University Press, 2007. 50 Van Melckebeke 1913, p. 10. 51 Van Uytven 1991, pp. 71-72. 52 Van Melckebeke 1913, p. 28. 53 Van Melckebeke 1874, p. 1. 54 Van Autenboer 1972, p. 4. 55 Van Melckebeke 1874, p. 9. 56 Van Doren, 1, 1859, p. 80. 57 Van Doren, 1, 1859, p. 79 58 Blockmans, Prevenier 1997, pp. 51-52; Paviot 2003, pp. 17-57. 59 Van Doren, 1, 1859, 79. 60 De Barante 1835, 2, pp. 25-26; 424-428. 61 Braunfels-Esche 1976, p. 81. 62 Braunfels 1974, pp. 385-387; Braunfels-Esche 1976, pp. 117119. 63 Vandenbroeck 1981, pp. 15-16, 48-50.

Fig. 30.1 Jan van Eyck (?), The Fishing Party (?), Paris, Museée du Louvre, Département des Arts graphiques (inv. no 20.674)

30

The Fishing Party by Jan van Eyck (?). A Technical Analysis Claudine A. Chavannes-Mazel with Matthias Alfeld, Koen Janssens, Geert Van der Snickt (MA-XRF), Peter Klein (Dendrochronology), Micha Leeflang, Margreet Wolters (Infrared Reflectography), Carel Van Tuyll van Serooskerken, André le Prat (Musée du Louvre)

ABSTRACT: In this paper we will explain that the Eyckian drawing known as The Fishing Party, long thought to be either an unfinished drawing, or a sketch for a mural or tapestry, is in fact composed of two fragments that once belonged to a single full-coloured drawing. MA-XRF scans prove that, with the exception of vermilion red and a copper-green, the pigments have completely faded. Furthermore, as the middle segment of the original drawing has been lost, it becomes clear that the two groups of noblemen and women are not standing on the left and right banks of a small brook, but rather on the shoreline of a lake or pond. At the time when the two fragments were placed immediately adjacent to each other, in the early seventeenth century, the ends of the hunting sticks were extended as fishing rods, a row of stones was added in the foreground and the background was repainted.

—o— The drawing known as The Fishing Party in the Louvre’s Department of Prints and Drawings is a composite image composed of two fragmentary leaves of paper in a slightly dilapidated condition (fig. 30.1).1 At an unknown date, the two original fragments were carefully pasted onto an underlying rectangular piece of paper with the dimensions 244 ≈ 388 millimetres and mounted on a wooden board.2 The seam where the two fragments were joined is still clearly visible (fig. 30.4). Further

examination of the original fragments reveals that the top corners of both leaves were once rounded off (fig. 30.3), and that the bottom right corner of the right-hand leaf was either missing or removed (fig. 30.10). In the composite drawing as we see it today, painted details were added in the areas bordering the two fragments (the two top corners, the triangular space in the middle at the top, and the bottom right-hand corner), evidently in order to create a unified image (figs 30.2, 30.3). Framing this rectangle – and in places overlapping the edges of the original fragments (fig. 30.10) – is a separate band of paper that has also been pasted onto the wooden support panel. This paper frame is brick red in colour, with a strip of gold and a thick black line painted along its inner edge, thus visually demarcating the boundary of the aforementioned rectangle. The red band of the paper frame bears an inscription in gold lettering: vetervm bvrgvndiae dvcvm conivgvmque filiorvm filiarvmqve habitvs ac vestitvs (‘The customs and attire of the old dukes of Burgundy, their spouses, theirs sons and daughters’). As it appears to the viewer, the drawing presents an idealized vision of a noble hunting party. Two groups of sumptuously dressed nobles stand across

456

claudine a. chavannes-mazel

from each other on either side of a brook: eight women on the left, with nine men observing them from the right-hand shore.3 One woman and two men appear to be fishing with rod and line in hand, though no fish can be seen. Simultaneously, three of the lords and two of the ladies hold birds of prey aloft on their gauntlets; some are holding up small bits of food with their free hand. In the foreground, two dogs run towards each other while a third dog looks on. In the background on the right stands a large edifice, roofed and crenellated, surrounded by turrets and towers. On the left, behind the group of women, we see what appears to be a row of lowlying dunes, covered with trees and various sorts of undergrowth. To complete the picture, the trees and the sky in the background are painted in their natural colours: green and blue. Eye-catching details are the vivid red vermilion colour of the men’s robes and the luxurious copper-green sleeves worn underneath. This is in contrast to the somewhat colourless appearance of the women’s apparel, in which traces of red can be seen only on their veils, hats and sleeves. On both sides, details in black and white can also be observed in the clothing. All of the figures are depicted wearing gold jewellery, varying from tiny gold buttons and buckles to heavy gold chains worn around the shoulders. In the eyes of a contemporary observer, individual members of the group were perhaps clearly identifiable via several minutiae, for instance as members of the Order of Saint-Antoie-en-Barbefosse and the Order of the Garter. Today, however, interpretations regarding the precise identity of the dignified group represented here diverge significantly.4 The lack of consensus concerning the dramatis personae also characterizes the discussion surrounding the drawing’s subject. Previous scholars have proposed themes ranging from a simple fishing or hunting party to the representation of an important diplomatic or political event. The scene has also been interpreted as an allegory of love. It was Otto Kurz who gave the drawing its most frequently used title: a fishing party.5 Châtelet prefers ‘Allegorical Depiction of the Conference at Biervliet’,6

and Magdi Toth-Ubbens is convinced we are looking at Queen Isabeau of France and the English king, Henry V, negotiating his marriage to Isabeau’s daughter, Catherine de Valois in 1420.7 Indeed, allegory, love and betrothal are the most common themes associated with the drawing.8 No less enigmatic is the drawing’s function, which has likewise given rise to ample speculation: are we looking at an unfinished drawing, a model for a tapestry, or – given the rounded corners of the two original fragments – a sketch for a diptych or murals in a room with a vaulted ceiling?9 Numerous extant late-medieval tapestries and murals feature the subject of noble men and women in combination with water and fish, such as the fourteenthcentury frescoes in the Castello Roncolo in Bolzano or the murals in the Castello del Buonconsiglio in Trento of around 1400.10 As elusive as any of these considerations may be, the most challenging aspect concerns the date of the original drawing, as well as the later additions and/or alterations made to it. When Otto Kurz first published his observations on this problematic sheet, he recognized its quality. At the same time, he was uncertain regarding its dating, based on what he described as ‘a very marked difference in style between the figures in the foreground which show the style of the early fifteenth century, and the landscape behind them which, with its depth and feeling for atmosphere, is unthinkable before the late sixteenth century’.11 Since then, most scholars have agreed that the background was added later, possibly by Joris Hoefnagel (1542-1600/01), as was suggested by J.G. van Gelder in a postscript to Kurz’s article.12 Moreover, it is generally believed that the ‘missing’ parts were added at the same time as Hoefnagel’s purported intervention, with the drawing as a whole then pasted onto the oak wood panel. In his catalogue of Early Netherlandish drawings in the Louvre, Frits Lugt was more critical, describing the work as a composition entièrement dans l’esprit de Jan van Eyck, copiée soigneusement par un artiste de la fin du XVIe siècle.13 In terms of its function and date,

the fishing party by jan van eyck

more recent authors remain as cautious, with the drawing inevitably viewed as an original work or a copy and dating either from the early fifteenth century or the late sixteenth century. Pierre Colman is the only one who has made a concise determination: L’auteur, c’est Jan van Eyck.14 Barring the discovery of any new findings the mystery concerning the group’s identity, as well as that of the drawing’s exact context and meaning, will never be unravelled. For this reason, new hypotheses regarding these aspects of the drawing’s history will not be introduced here. Nor will the focus be on the purported attribution to Jan van Eyck. The argument for dating the original drawing to Jan van Eyck’s period in Holland (before 1425) – as opposed to the sixteenth century – is based upon the meticulous rendering of the figures’ attire. This topic is beyond our present scope, however, and will be addressed in another article to follow.15 Instead, the emphasis will be placed on resolving a number of key issues that concern the drawing itself. This will be done by applying new techniques that were hitherto inconceivable and presenting new insights regarding the discoloration of old pigments. One issue to be considered is whether the two groups were originally presented in the same manner as they are currently found, i.e. facing each other directly in the open air on the two banks of a small brook. No previous author has ever approached this matter. In 1958, Lugt explicitly stated that the drawing consists of two pieces pasted together on a wooden panel. Yet the implications of his findings, such as the possibility of an alternative layout, have never been further pursued. An additional issue to be addressed focuses on the drawing’s function as a sketch for a mural or tapestry and its classification as an unfinished drawing. Finally, a key question concerns the matter of the drawing’s purported subject: are we indeed really looking at a ‘fishing party’? Technical Analyses In August 2011, The Fishing Party was subjected to both a detailed visual investigation as well as vari-

457

ous non-invasive, scientific analyses conducted at the Louvre conservation studio. First, a close examination with the naked eye was carried out with the aid of high-resolution photography. What we expected to find was an unfinished sketch with red vermilion pigment, some copper-green and hints of gold, applied directly to the paper. We instead discovered that the entire surface of the original paper is covered with a beige-coloured layer composed of a substance, the nature of which cannot be determined.16 Only in those few places where the beige layer is damaged – for instance, at the lower right of the arm of the female figure with the fishing rod (fig. 30.2) – can one directly see the original paper underneath. Hence, the colour is likely to have been added to the drawing by applying pigments onto the paper, with this beige layer as the binding medium. Virtually all of the pigments have since faded, with the exception of vermilion red and copper-green. It is the disappearance of these pigments that elucidates why, for instance, the emblem of the Order of the Garter on the left knee of the man in the foreground is colourless and not blue as it should be. Once one becomes aware that the composite image comprises different layers of paint, other areas emerge that clearly betray evidence of overpainting. Such is the case with light green and light blue in the background used to depict the sky and foliage, which occasionally spills over onto the black line of the red paper frame (figs 30.2, 30.3). Such colours were applied at a later date and could not have been part of the original drawing. Stones or trees have been added to the composition in three of the drawing’s four corners, as well as the missing triangle in the middle at the top. Furthermore, in the middle section of the painting, the trees on the left do not correspond with those on the right. In short, one gets the impression that the relationship between the two pieces in the original version of the drawing was markedly different than how they are presented to us now. The strange termination of the brook both in the foreground and background (fig. 30.4) – it

458

claudine a. chavannes-mazel

ends in a void between the two groups of trees – only strengthens this observation and confirms the suspicion that the row of stones in the foreground is a later addition as well, designed to function as a repoussoir device. In the same week, dendrochronology on the wooden panel was carried out by Peter Klein. When considering that the two original drawings are certain to have been in a frail condition at the time the other areas were added to complete the rectangular scene, it seems likely that the wooden base was required for additional support. The dendrochronological dating of the wooden support is therefore significant. Applying the master chronologies of oak originating from the German/Netherlandish regions, Klein dated the youngest heartwood ring to 1603, with a felling date of 1612

onwards most likely.17 This date corresponds perfectly with the earliest known description of the drawing, i.e. entry number 131 in the inventory of Emperor Matthias, drawn up at the time of his death in Vienna in 1619. Here it is described as ein taffel von miniatur, wie die alte burgundsche herzog und herzogin pflegten gekleidt zu sein, mit einer vischerei (‘a tableau with miniature, such as the old Burgundian duke and duchess used to be dressed, with a fishing scene’.)18 It is therefore likely that the two pieces were refitted into a rectangular format and pasted onto the wooden panel at some point between 1612 and 1619. In such a case, the connection to Joris Hoefnagel would have to be dismissed, as he died in 1600 or 1601. In addition to dendrochronology, further examination of the drawing was made with infrared

Fig. 30.2 The Fishing Party, detail of Fig. 30.1, top centre

the fishing party by jan van eyck

459

Fig. 30.3 The Fishing Party, detail of Fig. 30.1, top right corner

reflectography (irr) (fig. 30.5). Unfortunately, no underdrawing was visible. A vague shadow behind the trees at the top to the right of the castle could possibly be interpreted as an additional tower, but this is unverifiable at present. Far more spectacular were discoveries obtained through elemental analyses, which ultimately led to the unveiling of some of the drawing’s secrets. Using macro-X-ray fluorescence (ma-xrf), Geert Van der Snickt and Matthias Alfeld scanned the drawing for two consecutive days. Elemental distribution images of mercury (Hg, from vermilion), gold (Au), lead (Pb, mainly from lead white), copper (Cu, from copper green and/or blue), tin (Sn, from lead tin yellow), calcium (Ca, chalk), iron (Fe, earth pigments), as well as correlation plots of zinc (Zn) and copper (Cu), reveal a very different picture of the drawing from how we see it today. Perhaps the greatest insight can be obtained from the gold distribution image (not shown, as the details are too small to be reproduced here), which reveals that small amounts of gold were once applied to the drawing. The gold found in the jewellery, the tiny buttons and clasps can still be seen with the naked eye, especially when one is aware of its presence. Furthermore, the scan established that gold was also used to embellish miniscule details that could not be seen otherwise, such as the spires of the castle. Even underneath the red

Fig. 30.4 The Fishing Party, detail of Fig. 30.1, bottom centre

frame, hints of gold can be discerned where it overlaps two of the spires. Minute detailing of this nature is unlikely to have been added later. This then raises the question of why a costly gilding technique would be applied to what was conceived merely as an unfinished sketch? Two important conclusions may be drawn from the gold distribution image. First, evidence of this sort demonstrates that the castle belongs to the original drawing and is in no way related to the later additions. As we shall see, the castle in the background figures prominently in virtually every scan made, except for the Hg scan. Contrary to assumptions that have been made in the past, one may therefore conclude that it belonged to the original design. Second, one may rule out the notion that the drawing merely had some kind of menial function. The presence of gold implies that the drawing – or pair of drawings – was initially conceived as a finished work of art and not as a sketch.

460

claudine a. chavannes-mazel

Fig. 30.5 The Fishing Party, IRR

In addition to the gold distribution image, the distribution images of the four most informative elements acquired through ma-xrf will be addressed here in successive order: mercury (fig. 30.6), calcium (fig. 30.7), lead (fig. 30.8), and iron. Figure 30.6 shows the presence of mercury, typical for vermilion (mercury sulfide). Vermilion was used to paint the men’s bright red garments as well as the accents on the women’s clothes, for example the round hat of the lady in the foreground and the veil worn by the young woman with the fishing rod. Traces of vermilion are likewise distinguishable in the faces of both the men and women. This latter finding is particularly relevant. Were the drawing an unfinished piece of art – with the left half never completed – vermilion accents would not have been added to the women’s fragment. Instead, the evidence suggests a finished drawing in which the colours (other than the vermilion of the women’s dresses) have all faded.

In the distribution image for calcium (fig. 30.7), it can be seen that calcium is primarily found in the chalk or gypsum present in the paper and/or in the preparation layer. The calcium signal is low or absent in areas where the paper was covered with a paint containing heavy elements such as mercury or lead (e.g. the men’s garments). In these areas, the superimposed layer attenuates the X-ray fluorescence of calcium. By contrast, the signal is stronger in parts where the overlying layer comprises lighter elements (e.g. the castle). Among the women, we see that the dresses of two of the figures display a higher calcium signal. The colour of the clothing worn by the women must therefore have varied. An additional point to be made concerns the foreground, where one can observe in this image that the water plants on the right fragment do not continue onto the left fragment, again providing evidence that the right and left do not match.

the fishing party by jan van eyck

While lead is a component of various pigments and additives, in this specific case, its presence in the lead distribution image is primarily linked to lead white (fig. 30.8). In spite of its toxicity, lead white (lead carbonate hydroxide) was the most commonly used white pigment, even well into the twentieth century. Its popularity stemmed primarily from its high covering power and excellent drying properties. In the drawing at hand, lead white was used in its pure form to add white accents, such as the highlights of the women’s dresses. Its primary application, however, was in combination with other pigments, where it was introduced to soften a pigment’s tone and to enhance the paint’s opacity. Its overall presence, including the areas of beige, confirms that the entire surface of the drawing was covered with paint. The bluish/blackish shade found in several areas of white – several of the men’s faces, for example – is a result of the aging of lead white, a chemical process that is not yet completely understood (fig. 30.9). The lead distribution image also confirms the presence of overpainting on the frame, as it reveals a detail that is normally hidden beneath the red border: at the lower right, the number ‘63[…]’ can clearly be seen (figs 30.8, 30.10). Frédéric Reiset, when compiling his inventory in 1866, is known to have catalogued the drawing as number 636. The red frame was therefore overpainted at a later time, in this case, after the last re-edition of Reiset’s inventory in 1887.19 Such findings are also confirmed when examining the back of the wooden panel. Apart from the Louvre inventory labels, the panel’s reverse side shows that previous paper frames have been removed at an earlier date and replaced by the current frame. Accordingly, the gold inscription does not date from the seventeenth century, but is rather a careful, much later reproduction. Another indication that the frame was overpainted on at least one occasion is the way in which the layer of brick-red paint carelessly spills over the gold line of the frame. This is especially noticeable along the drawing’s top edge on the right and along the bottom (figs 30.3, 30.4).

461

Finally, the Fe scan (not shown) reveals that iron is present in two different forms. It is chiefly encountered in the form of iron gall ink, the medium that was used to draw the brownish/blackish lines in the drawing. Iron also occurs in the iron oxide earth pigments that were used to paint the frame. The presence of earth pigments in the paint layer itself could neither be proven nor disproven. It is probable that they were used to add further accents. What Do the Scans Tell Us about the Original Design? Apart from the fading of the pigments, the scans provide convincing grounds for distinguishing between those aspects of the drawing that are original and those which are not. What we now know is that the entire surface of the original drawing – or pair of drawings – was completely covered with paint, because the material binding the pigments, i.e. the beige-coloured layer, is present everywhere. With the exception of copper-green and vermilion, all other pigments have faded to the point where there is virtually no longer any trace of colour. One area where a residual hint of blue can still be seen is on the right sleeve of the woman with the round headdress (fig. 30.4). All things considered, however, the degree of fading that occurs on the two fragments is virtually the same: any perceived difference in this respect is misleading, as it is merely related to the greater prevalence of vermilion and green on the right fragment when compared to the left. As such, we may reasonably conclude that the two leaves shared the same fate over time. Unquestionably, three of the original drawing’s four corners are missing and have been repainted. The triangular piece between the two fragments has also been repainted to compensate for the rounded corners. Moreover, the green of the leaves and the blue of the sky were added after the picture frame, as these colours overlap the black line marking its edge (figs 30.2, 30.3). Once we are accustomed to the drawing’s complexity as a whole, then the incongruities between the two fragments

claudine a. chavannes-mazel

Fig. 30.6 The Fishing Party, Hg scan (mercury)

Fig. 30.7 The Fishing Party, Ca scan (calcium)

Fig. 30.8 The Fishing Party, Pb scan (lead)

the fishing party by jan van eyck

463

Fig. 30.9 The Fishing Party, detail of Fig. 30.1, the group of men on the right bank of the brook

become more obvious. The trees in the background, for example, do not meet up in the middle, as one would expect. In the foreground, the vegetation is found on the right-hand fragment only. All in all, one finds virtually no match along the vertical seam where the two fragments meet. The only plausible conclusion that can be drawn is that an essential portion of the drawing is missing – one that holds the key to understanding what was once originally depicted. The lack of correspondence between the two fragments as well draws one’s attention to the distinctly different angles from which the two groups of aristocrats have been portrayed. The women are not only seen from a slightly higher angle, they are also depicted somewhat larger than the men. This discrepancy in perspective suggests that in the ini-

tial composition, the two groups were not shown facing each other on opposite sides of a babbling brook, but were depicted standing along the shoreline of a more extensive body of water, such as a pond or lake, at two different locations. In other words, there is much more to the missing area between the two surviving fragments than the current presentation suggests. Any pronouncement regarding what was originally depicted in the missing middle section of the drawing is simply circumspect. What we do know is that the group’s active gestures, the dogs running towards each other, the birds of prey and the rods are certainly elements of the drawing as it was originally conceived. Moreover, if we take the poses and gestures of the noble men and women at face value, there is little doubt that a number of

464

claudine a. chavannes-mazel

these figures were holding some sort of rod or staff in their hands from the outset. Both the woman holding the rod and the kneeling man pull back their extravagant sleeves to allow freedom of movement. As the drawing appears in its present state, the key to understanding the activity in which this noble group is involved is conveyed by the presence of fishing tackle, i.e. fishing rods, angles, lines and sinkers. But were this actually the case, why then would angling ever entail hunting with birds of prey or require the presence of dogs? And correspondingly: if the two groups are not fishing, what activity would then have been depicted in the original drawing? The solution is to be found in contemporary hunting scenes, for instance in the calendar miniature for the month of August in the Très Riches Heures du Duc de Berry, which features a hunting party led by a falconer. The falconer carries a long stick for scaring wild prey out of trees and bushes.20 The so-called DijonVersailles paintings – post-medieval copies of fifteenth-century originals – reveal the function of this hunting implement even more clearly.21 In the backgrounds of these two paintings, several youths are seen along the edge of a lake, waving their hunting sticks in and around the reeds in order to scare birds out of the water. Indeed, long canes were used by nobles when hunting with birds of prey and dogs, as frontispieces to later dictionaries clearly illustrate.22 Consequently, one may conclude that the activity depicted in the original drawing was a typical hunting scene that at some point was transformed into a scene of fishing by converting the sticks into fishing rods. The manner in which the ends of the rods now extend from one fragment onto the other – i.e. as fishing rods – could not possibly have been an element of the drawing’s original concept. Hence, the young woman’s fishing rod in the right fragment is a later addition made to the original drawing, as are the fishing lines, which curl oddly between rod and sinker. Alterations of this nature are likewise found in other areas of the drawing in its present state.

Although part of the original design, the castle in the background was highlighted in white paint during restoration. However, the manner in which this was done clearly illustrates that there was no strict adherence to the architecture’s original layout: dark spots indicate the presence of window apertures that were not highlighted. Accordingly, the exact layout of the castle can no longer be determined. Additional changes have also been made to the countryside on the right. In the infrared reflectogram, a dark area can be seen directly behind the heads of the men (fig. 30.5). To the right of the castle, a vague silhouette (a tower?) has been removed, as can be observed in the lead distributions. The landscape behind the noblewomen poses a different problem. The beige layer covers this part of the drawing, just as it does the rest. This tells us that something was painted in this part of the drawing. In all probability, some form of landscape was once depicted here. On the other hand, we have concluded that the green and light blue colours of the trees and sky were added later: the leafy green that has been added is used only to fill in the space left open by the artist that painted the trees. Finally, the lead-white image shows a darker section that extends downward from the height of the shoulders of the lady fishing. In terms of the original composition, this brings the authenticity of the dunes and trees above that horizontal line somewhat in doubt. For the time being, we can only suggest that additional research is necessary. General Conclusions To summarize, it is clear that the two pieces in their original state could not possibly have been placed opposite each other in the manner we see them today. That said, the diptych format is not as strange as it might first seem. Each of the two pieces of paper, then of a size that can no longer be determined, must have been glued – with their corners still intact – onto another piece of paper, for instance, the page of an album with a collection of drawings. At the time they were removed from this

the fishing party by jan van eyck

page, the uppers corners were cut and rounded off. Such ‘intrusions’ were virtually common practice: a substantial number of drawings and etchings from anonymous collections reveal outlines that were devised in this manner.23 For the time being, it suffices to say that the drawing was damaged and worn in various places, requiring a serious restoration that was undertaken in the early years of the seventeenth century. This was carried out with great care and knowledge in order to preserve the coherency of the two fragments’ appearance. Stones were placed in the foreground as repoussoir devices to enhance the total effect; blue paint was added to suggest a flowing brook; the ‘hunting’ sticks in the hands of the men and woman were transformed into fishing rods. A brick-red frame with an inscription in Latin was added at that time, overlapping the original drawing (fig. 30.10). Tints of blue and green were introduced to cover the trees and skies, giving the drawing a unified appearance – so typical for the early seventeenth century. These tints in the background overlap the black line of the passepartout, proving that they do not belong to the original concept. In the late nineteenth century, the red of the frame was again almost completely repainted, with the gold letters rewritten in the same style. With this alteration, the number of Reiset’s inventory was painted over, as revealed by the lead distribution image. One last question remains: why did someone take the trouble to repair two fragments of a drawing made on paper that was in an incomplete state, devoid of their original colour (with the exception of the copper-green and vermilion)? The detailed representation of sumptuous clothing could not have been the only feature of value to the observer. It is true that in the seventeenth century, historical

465

fashion enjoyed significant popularity. In its own right, the Recueil d’Arras may be viewed as a costume book of its era. 24 Certainly, artists like Rubens and Rembrandt made sketches from model books of this nature. In our case, however, it would have been much easier to copy an image, rather than reworking the figures’ costumes with such painstaking care. The panels in Dijon and Versailles are fine examples of such a practice, comprising careful copies of representations that depict medieval festivities, for which no original has been preserved.25 Perhaps the drawings held some kind of emotional value, be it the recognition of illustrious ancestors or a recorded event in that individual’s ancestral history. Its value may perhaps have been based on the drawing’s place in a prestigious collection, and/or the implicit quality of the original fragments. Prior to its restoration, the drawing could not have had much of an aesthetic value due to discolouration; and even after its restoration, it still remains a patchwork at best. The male faces remain unhappy caricatures. The drawing’s first mention in the catalogue of 1619 does not provide any clues, other than a stated emphasis on its Burgundian provenance and the costumes that were represented.26 Of one thing, however, there can be no doubt: the drawing’s careful restoration has ensured its survival. Few drawings of this kind have enjoyed so fortunate a fate, with many lost for all posterity. This fact alone lends the drawing a hidden and respected significance – one that remains a secret to this day. What is clear, however, is that it is more suitable to call the drawing: ‘Fragments of a Burgundian hunting scene’, than ‘A Fishing Party’, since it was recreated as a Fishing Party only centuries after its genesis.

466

claudine a. chavannes-mazel

Fig. 30.10 The Fishing Party, detail, bottom right corner (Pb and Hg scans) with the frame overlapping the figure furthest right

NOTES * I thank Carel van Tuyll and André le Prat from the Louvre’s Department of Prints and Drawings for their generous support. I thank Thomas Belyea, Amsterdam, for the English translation. 1 Paris, Musée du Louvre, Département des Arts graphiques, no. 20.674. 2 The condition of the two paper fragments at this time was apparently such that they could not be flattened entirely, resulting in a number of uneven horizontal creases and ripples that can still be seen with the naked eye. 3 Lugt 1968, pp. 5-6 and pl. 6. 4 From Kurz 1956 onwards, interpretations vary. Colman 2006 gives a detailed historiography, pp. 8-11; Châtelet 2010, p. 316.

5 Kurz 1956. 6 Châtelet 2010, p. 316, Châtelet 1980, p. 197, no. 21. 7 Tóth-Ubbens 2012, passim. 8 Dhanens 1980, pp. 160-162; Colman 2006. 9 Lugt 1968, p. 5. 10 IJpelaar 2012. For tapestries with comparable subjects, see Kurz 1956, fig. 5, and Pächt 1999, p. 117. For murals, see Vaivre 1985, figs 16, 17. For miniatures, see IJpelaar 2012. Christian Klamt, Utrecht, introduced us to the interesting phenomenon of ‘Traditionsbücher’. In the oldest and most famous, the Codex Falkensteinensis (München, Bayerisches Hauptstaatsarchiv MS Kloster Weyarn 1, c.1164-1170), a man catches a fish with a fishing rod in the water around the castle of Hartmannsberg on fol. 11r. See Rösener 2000 and http://www. bayerische-landesbibliothek-online.de/codexfalkensteinensis (accessed January 2013).

the fishing party by jan van eyck

11 Kurz 1956, p. 124. 12 Van Gelder in Kurz 1956, p 131. Joris Hoefnagel was a Flemish artist residing at the imperial court in Vienna from 1591 until his death in 1601 (Hendrix 2007-2012). 13 Lugt 1968, p. 5. 14 Colman 2006, p. 18. 15 Margaret Scott will co-author part of this research. In that article, we will include the copy of the drawing, painted in 1841 by Charles Alexandre Debacq (coll. châteaux de Versailles et Trianon MV 4020, inv. no. 3716; LP 4660). 16 Because our approach was non-invasive, we were unable to take a sample of the beige layering. This would be one way to determine the exact nature of this layer. 17 Report on the dendrochronological analysis of the Fishing Party panel (Jan van Eyck, Paris, Musée du Louvre inv. no. 20674), 2 September 2011, by Peter Klein: ‘The support of the painting (paper on wood 25.2 ≈ 39.7 cm) is made from one oak board. The wood concerned is originating from the Netherlandish/German regions. Using these master chronologies the rings could be dated as follows: Board I 139 growth rings 1603 – 1465 The youngest heartwood ring was formed out in the year 1603. Regarding the sapwood statistic of Western Europe an earliest felling date can be derived for the year 1610, more plausible is a felling date between 1616….1620……1626 + x. With a minimum of 2 years for seasoning an earliest creation of the support is possible from 1612 upwards. Under the assumption of a median of 15 sapwood rings and 2 years for seasoning a creation of the support is plausible from 1622 upwards.’ 18 Köhler 1906-1907, inv. H (Nach 1619 Juni 18), p. X. The word taffel does not necessarily refer to wood. Compare with other entries in Köhler’s inventory, pp. VIII-XIII.

467

19 Reiset 1866, p. 340. His inventory was reprinted without changes in 1878 and 1887, with a description of no. 636 on p. 340. On the back of the panel, the inventory number ‘I2025’ has been branded into the wood and was certainly made before the ’63[…]’ number on the front. At a later time, an inventory number was added in pencil, ‘N 20674’. Finally, paper labels with the recent Louvre inventory numbers are ‘case 11 20.674 N III 14540’ and ‘P 10’. The oldest number mentioned in the catalogue of 1619, i.e. no. ‘131’, is found nowhere on the panel. 20 Chantilly, Musée Condé MS 65, fol. 8v. Cazelles, Rathofer 1988, p. 42. 21 Dijon, Musée des Beaux Arts, inv. no. 3981: ‘Fête champêtre à la cour de Philippe le Bon’; Versailles, Musée national des châteaux de Versailles et de Trianon, MV 5423; RF 1141: ‘Marriage de Philippe le Bon, janvier 1430’. Constans 1980, no. 4995; Buren 1985, passim; Châtelet 2007, p. 96. For the most recent bibliography, see Rotterdam 2012/2013 no. 78. Kemperdick writes on p. 278 that the men in the background are fishing – which they are not. 22 The title page to the most famous of these, A Jewell for Gentrie, was published in 1614, but goes back to earlier versions, probably by Juliana Berners. See also Hoffmann 1985. 23 For instance, Dresden 2005, cat. nos 1, 2, 4, 5, 18, 25, 43, 44, 46, 47. 24 Châtelet 2007. 25 See above, n. 22. 26 See above, n. 19.

Fig. 31.1 Joan Reixach, St John, 1446-1448, fragment of the Predella of the Evangelists, panel, 62 x 47 cm, Valencia, Museo de Bellas Artes (deposit from the Diputación de Valencia)

31

Van Eyck in Valencia Bart Fransen

ABSTRACT: In this article I argue that Jan van Eyck visited Valencia as part of the ducal embassy to Aragon in the autumn/winter of 1426, which was the first of two embassies sent by Philip the Good to negotiate his marriage to Eleanor of Aragon, sister of Alfonso V of Aragon. This visit probably enabled a number of Valencian painters to contact the Burgundian court painter and gain direct access to his technical and pictorial innovations. Within a small network of Valencian artists, some even acquired a painting and a drawing attributed to him. Exceptionally, tracings after Eyckian works were imported and literally taken over, as is evidenced by the study of Dalmau’s Virgin of the Consellers. Dalmau made his tracings after the Ghent Altarpiece in Van Eyck’s workshop, before the work was installed in the Vijd Chapel. The duration of Dalmau’s stay in Van Eyck’s workshop should therefore be taken into account when studying the genesis of the Ghent Altarpiece.

—o— More than any other Spanish region, Jan van Eyck’s oeuvre received a very early response in Aragon, and particularly in Valencia, where his models were used among a small network of Eyckian artists. Moreover, it is very likely that the painter to the Burgundian court stayed in Valencia while on one of the diplomatic missions for the Duke of Burgundy, Philip the Good (1396-1467), even if no archival document explicitly mentions his name as part of one of the embassies. Jan van Eyck in Aragon From 1 July 1427 to 15 February 1428 a Burgundian embassy of around forty members,1 headed by Lourdin de Saligny and Renaud de Fontaine, Bishop of

Soissons, was sent to Aragon to negotiate – inconclusively, as it turned out – the marriage of Philip the Good and Eleanor of Aragon (1402-1445), the sister of Alfonso  V of Aragon (1396-1458).2 En route to Valencia the ambassadors passed through Paris, Troyes, Chambéry and Nice; on their way back they made extra stops in Ibiza and Menorca. Jacques Paviot has argued that Jan van Eyck could not have been part of this embassy, firstly because payments made in July and August 1427 suggest that he was in the Low Countries, and secondly – and primarily – because a Johannes pointre was received by the painters’ guild in Tournai on 18 October 1427, the feast day of St Luke, a date that is incompatible with the duration of the embassy.3 From October 1428 to December 1429 Van Eyck did take part in the grand Burgundian matrimonial embassy to Portugal, which numbered around seventy five members and was led by Jean de Roubaix, councillor and first chamberlain to the duke and much more experienced than the aforementioned Lourdin de Saligny.4 The adventurous journey to Lisbon and back, via the ports of southern England and Galicia, is documented in a well known travelogue probably written by a Portuguese individual, which survives in two copies, one French and one Spanish, the latter only partially preserved.5 Van Eyck joined this mission not as an ambassador per se6 but as a court painter, charged with painting a portrait of the Infanta, Isabella of Portugal (1397-1471), Philip’s intended bride.7 The portrait was painted in the Portuguese city of

470

bart fransen

Avis in January/February 1429 and its iconography is known through a seventeenth-century copy drawing conserved in Lisbon (Arquivo Nacional da Torre do Tombo (see fig. 21.1).8 Once the negotiated marriage treaty and Van Eyck’s painting had been sent off to Philip the Good, the ambassadors spent four months visiting several places on the peninsula. In contrast to the detailed description of the period spent in Portugal, the travelogue mentions those excursions only briefly, probably because its author was not one of the party. He does tell us, however, that some members of the embassy – Van Eyck is not specified – visited Santiago de Compostella, the duke of Arjona (in Zamora ?), the king of Castile (in Valladolid ?), the city of Granada and ‘other places’. Some authors have thought that the group went to Aragon,9 but this is due to a misreading of the French word ‘Arjonne’ which refers not to Aragon but to Arjona. According to recent findings in Aragonese archives, published by Blasco Vallés and Reche Ontillera, a Burgundian embassy had already been sent to Aragon in 1426, before the mission of 14271428 mentioned above.10 The safe-conduct for that first embassy was written by Alfonso V and dated 28 September 1426. Although Jan van Eyck is not mentioned in the document it is likely that he was part of the Burgundian delegation, which included between twenty and forty members and was led by Lourdin de Saligny.11 The four ambassadors listed by name and surname in the safe-conduct are Lourdin de Saligny, André de Toulongeon, Jean de Terrant and Jean Hibert.12 On 31 July 1426 those four men had already been paid by Philip the Good to undertake the secret mission ‘to certain places faraway and for certain matters that affect him greatly and are close to his heart’.13 Paviot pointed out that the same Burgundian accounts include a payment, dated 26 August 1426, to Jan van Eyck, also for a ‘secret journey to certain places far away also ordered by the duke on which nothing needs to be declared.’14 In my view, this payment to Jan refers to the same secret mission to Aragon in 1426 and confirms that Jan van Eyck was part of that first embassy.15

The exact dates of the journey are not recorded.16 In principle the safe-conduct dated 28 September 1426 was valid until Easter 1428 (20 April), but it is likely that the embassy returned much earlier. Based on the available documentation, we can assume that the ambassadors left the Low Countries in October/November 1426, stayed in Valencia in December (in which month a letter from Alfonso to Philip the Good is dated)17 and came back to the Low Countries in January/February 1427, in any case well before 1 April, when Alfonso V wrote another letter to the duke offering a safe-conduct for the second mission that left on 1 July 1427.18 Most scholars take it for granted that the first mission to Aragon was sent with the aim of negotiating a marriage between Philip the Good and the Catalan noblewoman Isabella of Urgell (14091459). This hypothesis, first put forward by James Weale in 1912, has infiltrated the literature and is widely treated as fact.19 In her study on the embassies of Philip the Good, Anne-Brigitte Spitzbarth indicated that the 1426 and 1427 embassies to Aragon were both related to the same marriage, the first one being a preparation for the second and larger one that was intended to finalize the alliance between Aragon and Burgundy.20 Although Spitzbarth is confused about the identity of Philip the Good’s prospective bride,21 it is clear from the Aragonese documents of 1427 that the ambassadors’ goal was a marriage with Eleanor of Aragon, sister of Alfonso V, a plan that was abruptly cancelled by Alfonso in a letter to Philip dated 14 August 1427.22 If Jan van Eyck did go along on the first journey to Aragon in 1426, his role might have been the same as it was in the 1428 mission to Portugal, that being to paint a portrait of the possible marriage candidate, Leonor of Aragon. That Van Eyck was not sent on the second mission of 1427 was probably because he had already completed his task on the first. A Valencian Network The presence of Eyckian works in Aragon is well studied as far as the royal collection of Alfonso  V, who resided most of the time in Naples, is concerned.

van eyck in valencia

In his residence in Naples he kept a triptych of the Annunciation (originally made for Giovanbattista Lomellino), a mappamundi (possibly by Van Eyck), a St George and the Dragon and an Adoration of the Magi.23 In this article, however, our focus is the city of Valencia and the immediate impact of Jan van Eyck’s oeuvre in a network of local painters. A most interesting testimony to this Eyckian network is the relatively early presence of a panel by Jan van Eyck (or workshop) in the collection of Joan Reixach (c.1411-1485), an important Valencian painter with a productive workshop, active between 1437 and 1482. Listed in his will of 1448 is una taula de pintura de la historia com Sent Francesch reb les plagues, acabada ab oli de la ma de Johannes (‘a painting on panel with the history of St Francis receiving the stigmata painted with oil by the hand of Johannes’).24 This was probably a third version of Van Eyck’s St Francis Receiving the Stigmata, known through the works in Turin (Galleria Sabauda) and Philadelphia (Museum of Art).25 In his will, Reixach left the painting to his good friend Andreu García, a well-connected Valencian priest and an outstanding promoter of the arts, who had also amassed a private collection of design drawings and models which according to García’s will would in turn be inherited by Reixach.26 The Van Eyck painting in Reixach’s collection was not, of course, an object for devotion – at least not in the first place – but served principally as a model for his own work and that of his pupils. Several derivative works by Reixach himself are known, such as the Stigmatization of St Francis in the predella of his St Ursula Altarpiece of 1468 (Barcelona, Museo Nacional d’Art de Catalunya), or the panel sold by Christie’s in London in 1998 (formerly in the Balanzó collection).27 The Eyckian model was also followed by other Valencian contemporaries, notably the Master of the Porciúncula, whose version, located in Castellón (Convento de Capuchinas), seems to be literally inspired by the Eyckian iconography (fig. 31.2).28 In the 1440s Reixach was already producing paintings in which his knowledge of Van Eyck’s

471

innovations can be observed, as for example in the three Evangelists from the banco of a lost Marian altarpiece, made between 1446 and 1448 for the chapel of the Hospital of Our Lady of the Innocents (Valencia, Museo de Bellas Artes).29 The St John the Evangelist is the first surviving painting in which, instead of a traditional gold background, Reixach used a modern setting, placing the saint in a fifteenth-century interior, with a view of a modern city in which the inhabitants wear contemporary Burgundian dress (fig. 31.1). It is clear that the name and the work of Jan van Eyck were well known in a network of painters around Joan Reixach. Other significant painters in this Eyckian network were Luis Alimbrot or Lodewijk Hallincbrood (active 1432-1463), Jaume

Fig. 31.2 Master of the Porciúncula, St Francis Receiving the Stigmata, panel, 149 x 98 cm, Castellón, Convento de Capuchinas

472

bart fransen

Baço (c.1410-1461) and Lluís Dalmau (active 1428-1460). Luis Alimbrot must have been familiar with Van Eyck’s oeuvre before leaving Bruges around 1437-1438 to settle in Valencia, where he bought a house right next to that of Joan Reixach, in the street of St Vincent (in the parish of St Martin), where Jaume Baço also had a house.30 Jaume Baço, called Jacomart, was Reixach’s associate in many projects.31 As court painter to Alfonso V between 1442 and 1448, he must have known the Van Eyck paintings in the royal collection in Naples. The monumental Annunciation attributed to Jacomart (Valencia, Museo de Bellas Artes) may predate his departure for Naples in 1442 and thus would indicate a rather early knowledge of Van Eyck’s Annunciation in the Ghent Altarpiece.32 Last but not least, there is the Valencian painter Lluís Dalmau, who travelled to Bruges in 1431 and may have seen Jan van Eyck working on the Ghent Altarpiece, as will be discussed further below. Reixach, Jacomart and Dalmau had all worked for the king of Aragon and had the right to place the royal arms above the entrance of their houses.33 To this network still must be added the name of Berenguér Mateu, a Valencian painter documented between 1423 and 1446 and brother of Jaume Mateu (1382-1452), a well-known exponent of the International Style in Valencia.34 Although none of his documented works are conserved, we know that Berenguer Mateu shared with this network a particular interest in Van Eyck and was an acquaintance of Andreu García, mentioned earlier as the inheritor of the St Francis Receiving the Stigmata from Joan Reixach’s collection. The inventory of Andreu García’s goods, dated 1452, has recently been studied by Encarna Montero who brought to light the fact that Andreu García had received from Berenguer Mateu the following objects as a guarantee for a loan of six florins: a book on painting, four panels with painted heads and una myga ymatge en paper, de ploma, de mà de Johannes (‘half an image, on paper, in pen and ink, by the hand of Johannes’).35 This last entry is likely to refer to a drawing by Jan van Eyck: firstly because other Valencian

documents refer to him by the Latinized version of his first name, ‘Johannes’,36 and secondly because Valencian painters named John were called ‘Joan’ and used a Latinized version of their name only when it appeared as part of a Latin text.37 Thus we may assume that the use of Latin solely for the name of the painter indicates that he was from outside of Valencia. Judging by the high esteem in which Jan van Eyck was held within this network, and the direct connections between Valencia and Bruges, it is probable that this drawing was by the master (or his workshop). This is the only evidence that a drawing by Van Eyck left his workshop at such an early date. It is hard to imagine what was represented on the drawing. ‘Half an image’ may refer to a halflength portrait but this is speculative. The main conclusion about this entry is that it demonstrates that it was possible for a Valencian painter to acquire a drawing by Van Eyck well before 1452. This is all the more fascinating when we take into account the fact that at that time, drawings were valuable possessions, belonging to the workshop or the family business in which they were produced, carefully preserved and often reused in the design of other art objects. Like the Stigmatization of St Francis in Reixach’s collection, the Van Eyck drawing in Berenguer Mateu’s collection was first and foremost a source of inspiration for his own paintings, but the fact that he left it with other objects as a guarantee for a loan of six florins indicates that the Van Eyck drawings had an important economic value long before art collectors started to buy them. Jan van Eyck and Lluís Dalmau The most literal echo of Jan van Eyck’s ars nova in Spain is without doubt the large Virgin of the Consellers by Lluís Dalmau (fig. 31.3) (Barcelona, mnac). The evidence regarding the commission and execution of the painting is extremely valuable: firstly the detailed contract dated 29 October 1443; secondly the mostra or preparatory drawing of the composition; thirdly the artist’s signature and the date, 1445, at the bottom of the panel. These

van eyck in valencia

473

Fig. 31.3 Lluís Dalmau, Virgin of the Consellers, 1445, panel, 2.72 x 2.76 (painted surface), Barcelona, Museu Nacional d’Art de Catalunya

facts, along with iconographical and historical data, have been much discussed in the literature.38 The influence of Jan van Eyck is also constantly mentioned, with a focus on iconographic similarities: the angels’ faces, close to those in the Ghent Altarpiece; the Valencian tiles, initially planned to be square but finally using a variety of forms and thus complicating the perspective; the brocade patterns, not simply painted as if on a flat surface, as

Dalmau’s predecessors did, but following the folds in the robe; the suggestion of physical contact by delicate touching hands; and the figure of the Virgin enthroned, literally inspired by the Virgin and Child with Canon Joris van der Paele in Bruges (Groeningemuseum), finished in 1436. But Dalmau not only copied some iconographic elements, he also inserted the date of execution and his own signature on the panel: sub anno

474

bart fransen

mccccxlv per ludovicum dalmau fui depictum. Exceptional for Aragonese painting, it is the painting itself that speaks:39 fui depictum, just like Jan van Eyck when he signed his works: johannes de eyck me fecit. Dalmau also followed Van Eyck’s concept of light and space. The ceiling in the Virgin of the Consellers is in fact an illusionistic continuation of the real architecture of the chapel, where the panel occupied an entire wall, as mentioned in the contract. The two pendants even seem to penetrate the space of the viewer. The introduction of the landscape and singing angels seems to have been Dalmau’s own initiative, because the contract with the councillors stipulated that the background should be gilded with Florentine gold.40 The same Eyckian formula is used for the representation of pearls and precious stones: instead of using pastiglia and gold leaf, as agreed upon in the contract, Dalmau opted to depict them in paint, just as Van Eyck did. In 2008 a technical microanalysis was carried out on the materials and techniques of the Virgin of

the Consellers by scientists at the Museo Nacional d’Art de Catalunya, which showed that the medium was a drying oil, and that the structure of the preparation and paint layers is comparable to that of Jan van Eyck.41 This strongly supports the hypothesis that Dalmau must have spent significant time in Van Eyck’s workshop, in order to obtain knowledge of techniques that were unknown in Aragon. It is well known that Dalmau was sent from Valencia to Flanders by Alfonso V of Aragon in 1431, and although the purpose of the trip is not specified in the document it can be assumed that the painter was sent to be trained in the ars nova of Jan van Eyck. This was not an isolated example of such internship. The same thing happened to Zanetto Bugatto, for instance, who in 1460 was sent by the Duchess of Milan to be an apprentice in Rogier van der Weyden’s studio. The Italian painter stayed no less than three years in the north. The exact date of Dalmau’s return to Valencia is not known, but his presence there is documented again from July 1436 onwards.42

Fig. 31.4 Overlay of the tracing of the head of St John the Baptist by Jan van Eyck (Ghent Altarpiece, Ghent, St Bavo’s Cathedral) onto the head of St Andrew by Lluís Dalmau (Virgin of the Consellers, Barcelona, Museu Nacional d’Art de Catalunya)

van eyck in valencia

Fig. 31.5 Overlay of the tracing of one of Jan van Eyck’s Singing Angels (Ghent Altarpiece, Ghent, St Bavo’s Cathedral) onto the head of the left singing angel by Lluís Dalmau (Virgin of the Consellers, Barcelona, Museu Nacional d’Art de Catalunya)

If we align the dates of Dalmau’s stay in Flanders in relation to the production of the Ghent Altarpiece we have the following chronology: on 11 September 1431 Dalmau was paid to undertake the journey to Flanders; the date 6 May 1432 is inscribed on the frame of the lower register of the Ghent Altarpiece; the foundation of the altar in the Vijd chapel, however, is dated 13 May 1435; Dalmau was back in Valencia by July 1436. It is hard to judge how long it would have taken Dalmau to travel from Valencia to Bruges in the

475

Fig. 31.6 Overlay of the tracing of the furthest right Singing Angel by Jan van Eyck (Ghent Altarpiece, Ghent, St Bavo’s Cathedral) onto the head of one of Lluís Dalmau’s singing angels (Virgin of the Consellers, Barcelona, Museu Nacional d’Art de Catalunya)

winter of 1431-1432. If the Ghent Altarpiece was finished and installed in St John’s Church on 6 May 1432, then Dalmau would have had little time to get in touch with Van Eyck and receive permission to copy his new creation before it left his workshop. It is arguable that Dalmau made his copies not in the Vijd Chapel, where the height of the upper register is too great to allow the making of a detailed copy, but in Van Eyck’s workshop itself, thus before the panels were installed in the church.

476

bart fransen

a

b

Figs 31.7a-b (a) Overlay of the tracing of the furthest right Singing Angel by Jan van Eyck (Ghent Altarpiece, Ghent, St Bavo’s Cathedral) onto the partly-visible head of one of Lluís Dalmau’s singing angels (Virgin of the Consellers, Barcelona, Museu Nacional d’Art de Catalunya); (b) Same detail of the partly-visible head of one of Lluís Dalmau’s singing angels in IRR

If, on the other hand, the complete Ghent Altarpiece, more precisely the upper register, was only finished by 1435, as argued by Hugo van der Velden,43 then Dalmau had time enough to make the journey to Flanders, to install himself in Bruges, to get involved in Van Eyck’s workshop and to make the exact copies of the work in progress in Van Eyck’s studio at that time. Two simple observations seem to support this possibility. First of all, the Eyckian elements introduced by Dalmau in his Virgin of the Consellers originate mainly from the upper register of the Ghent Altarpiece: the angels, St John and the tiled floor. The lower register did not really influence the work of Lluís Dalmau.44 This raises the possibility that during Dalmau’s stay,

only the upper register of the Ghent Altarpiece was being executed in Van Eyck’s workshop. Secondly, comparative research based on tracings confirms that the size and the contours of several copied heads in Dalmau’s work are so close to the ones in the Ghent Altarpiece, that Dalmau must have made exact tracings of the Eyckian heads and that he used them in 1445 (ten years later!) in his Virgin of the Consellers.45 A first example is the figure of John the Baptist, in which the position of the mouth, the nose and the eyes is identical (fig. 31.4) to those in the Ghent Altarpiece. The use of green for John’s mantle indicates that Dalmau had seen Van Eyck’s painting in a finished stage. A second example is the angel on the

van eyck in valencia

extreme left in Dalmau’s panel, which closely matches the one in Van Eyck’s panel of the Singing Angels, specifically in the contours and the position of the diadem, the eyes, the eyebrows, the nose, and the (slightly shifted) open mouth (fig. 31.5). A third and last example, in the left window of Dalmau’s panel, are the two angels seen in frontal view, the second of whom is only partially visible. Here Dalmau seems to have used the same Eyckian model twice. The first face shows the eyes, eyebrows, nose and chin in the same position – only the mouth is slightly different (fig. 31.6). The second face, only partially visible, is based on the same model (fig. 31.7a), but remarkably, Dalmau decided to change the position of the angel’s right eye: in a first stage he underdrew and painted the eye in the same position that it appears in the Eyckian model; in a final stage he positioned the eye slightly higher (fig. 31.7b). These proportional similarities between the masterpieces of Van Eyck and Dalmau confirm the idea that during his stay in Bruges Dalmau had the privilege of direct access to the panels of the upper register of the Ghent Altarpiece in Van Eyck’s workshop, and that he was allowed to make faithful copies of these in the period between his arrival and the altarpiece’s completion. In summary, we can conclude that in fifteenthcentury Spain it was first in Valencia that certain painters had a particular interest in imitating the technical and pictorial innovations of Jan van Eyck, not only because Alfonso V had Eyckian works in his collection, as is often said, but principally because in a relatively early stage a small network of Valencian artists were able to gain direct access to the work and the working method of the Burgundian court painter. They might have been in touch with Van Eyck when he was in Valencia in the winter of 1426-1427; some of them even acquired a painting and a drawing attributed to him; and exceptionally, as in the case of Dalmau, tracings made directly after Eyckian works in the North, were imported and literally taken over.

477

N OTES 1 On the estimated number of participants, see Spitzbarth 2013, pp. 475-476. In any case the embassy could not have exceeded sixty members, which was the maximum indicated in the safe-conduct Alfonso issued in 1427. See Blasco Vallés, Reche Ontillera 2007, pp. 92, 412. 2 The main document on this embassy, its participants, duration and itinerary, is the mission report dated 20 August 1429, conserved in Lille (ADN, N 19·fol. 257r-259r) and published by Spitzbarth 2013, pp. 616-621. 3 Paviot 1990, p. 86, nn. 22, 24. 4 Spitzbarth 2013, pp. 296, 473, 476. 5 The complete version, in French, which Paviot erroneously thought lost (Paviot 2013), is well conserved in Brussels (Archives Générales du Royaume, Chambres des Comptes, registres, 132, fols 157-166) and was first published by Weale 1908, pp. lv-lxxii. The incomplete version, in Spanish and conserved in Paris (Bibliothèque Nationale de France, Archives et Manuscrits, Ms. Portugais 20, Ancien fonds no. 10245, fol. 105-129 v) has been published by de Vasconcelos 1897, pp. 10-45. 6 Spitzbarth 2013, p. 297. 7 Isabella of Portugal became Philip the Good’s third wife in January 1430. A marriage between Philip and Isabella had already been negotiated following the death of the first duchess, Michelle de Valois (1395-1422), but nothing came of it. Instead, Philip married Bonne of Artois (1396-1425) in 1424 but she also died shortly afterwards. A third marriage for Philip was first projected with Leonor of Aragon, a project that seems to have been almost completed, only to be suddenly cancelled due to the interference of the son of John I of Portugal, Peter, Duke of Coimbra (1392-1449), who negotiated a double alliance with the house of Aragon: he himself married Isabella of Urgell and his brother Edward married Leonor of Aragon. So Philip the Good finally married the Portuguese princess with whom a marriage project was already initiated by January 1424, as stated in a letter to the Republic of Venice dated 31 January 1424. See Spitzbarth 2013, p. 343, n. 448. 8 Fransen 2009, p. 107; Fransen 2012, p. 76. See also Kemperdick in this volume. 9 Van Puyvelde 1940, p. 23. 10 Blasco Vallés, Reche Ontillera 2007, pp. 91-95. 11 These numbers are based on an estimate of twenty-one members, following the calculation by Spitzbarth and the maximum of forty members indicated in Alfonso’s safe-conduct of 1427. See Spitzbarth 2013, pp. 473, 475-476 and Blasco Vallés, Reche Ontillera 2007, pp. 92, 412. 12 It is not clear to whom the certain Iohannis militem, cambellanum refers. It might be the squire Jean de Baissey or one of the two couriers, both called Jean; only one of them is known by his surname, Coq. See Spitzbarth 2013, p. 295, n. 282. 13 This payment is dated 31 July 1426 and the amounts mentioned are: for Lourdin de Saligny 1440 francs (of 40 groats) and a gift of 500 francs; for André de Toulongeon 720 francs and a gift of 250 francs; for Jean de Terrant 480 francs and a gift of 150 francs; and for Jean Hibert 360 francs and a gift of 100 francs. Paviot 1990, p. 86. 14 The complete reference reads as follows: tant pour faire certain pelerinage que mondit seigneur pour lui et en son nom lui a ordonné faire, don’t autre declaracion il n’en vault ester faicte, comme sur ce que par icilui seigneur lui povoit ester deu a cause de certain loingtain voyage secret que semblablement il lui a ordonné faire en certain lieux que aussi ne vault aultrement declarer. The first lines deal with a ‘certain pilgrimage’ that probably has nothing to do with the diplomatic character of the missions discussed here. The word pelerinage cannot allude to any diplomatic mission as these are always referred to as voyage or ambassade. See Spitzbarth 2013, p. 45. 15 The fact that the payment to Jan van Eyck is not listed together with the payments to the ambassadors is no reason to think

478

bart fransen

that he was not part of the same mission. A comparable example is the payment to a certain Remon Monnessou who took part in the second embassy to Aragon, for which he received 55 pounds (Courtrai, B. Ville, Cod. 322, XVI, fol. 88v), though he is never listed together with the other ambassadors and does not appear in the final report of the embassy. See Spitzbarth 2013, online database http://ambassadeurs. plb.free.fr, embassy record no. 277 (accessed 4 November 2014). 16 Paviot, in this volume and Paviot 2013 thinks that the journey took place during the summer and autumn of 1426 but this contradicts his earlier statement in which he argued that two payments to Jan van Eyck (26 August and 27 October) imply that Jan was resident in the Low Countries during the summer and autumn of 1426. See Paviot 1990, p. 86. 17 In a letter dated 20 December 1426, the Aragonese king informs the Burgundian duke that his ‘instructions have been followed’. See Blasco Vallés, Reche Ontillera 2007, pp. 412. 18 Blasco Vallés, Reche Ontillera 2007, pp. 412-413. 19 See, for example, Paviot in this volume; Blasco Vallés, Reche Ontillera 2007, pp. 92-95, 411-414; Mira 2007, p. 265; Lacarra Ducay 2007, p. 246; Bermejo 1990, p. 559; Dhanens 1980, p. 47; Pemán y Pemartín 1969, p. 33. 20 In 1427 Philip the Good sent the same four ambassadors to Aragon as in 1426, but added two important men: the Bishop of Soissin, Renaud de Fontaine, for the religious aspects of the marriage, and a Genoese merchant Oliviero Maruffo, who probably dealt with the financial aspects of the agreement but was also useful because he knew Aragon very well, having lived there for many years. See Spitzbarth 2013, p. 382, n. 592. 21 Erroneously Spitzbarth mentions ‘Éléonore, the mother of Alphons V’ as Philip’s prospective bride and elsewhere she is identified as ‘the daughter of John V‘. Spitzbarth 2013, p. 295, 343. 22 The reason for the cancellation of the Burgundian-Aragonese marriage is that Alfonso V had meanwhile accepted a deal with Peter, Duke of Coimbre and son of John I of Portugal, concerning a double alliance between the house of Aragon and the house of Avis: Isabella of Urgell married Peter, Duke of Coimbre, and Leonor of Aragon married Peter’s brother Edward. See Blasco Vallés, Reche Ontillera 2007, p. 414 23 On these works, and on an Adoration of the Magi that was also attributed to Van Eyck, see Toscano 2007, pp. 352-357; Challéat 2007, pp. 371-374; Cornudella 2009-2010, pp. 47-54; Jones 2014, pp. 30-43. 24 Benito Doménech 2001, p. 24; Gómez Frechina 2001, pp. 67-75; Valencia 2001, pp. 118-120; Gómez Frechina 2007, pp. 390392; Jones 2014, pp. 35-37. 25 Valencia 2001, pp. 106-117. 26 On the relationship between Joan Reixach and Andreu Garcia, see Ferre i Puerto 1999. On Andreu Garcia, see Montero Tortajada 2013.

27 London, Christie’s, 10 July 1998, lot 193 (as Valencian School); Gómez Frechina 2001, p. 69; Jones 2015, p. 35-37. 28 Valencia 2001, pp. 118-123. 29 Valencia 2001, pp. 208-215; Almudín 2007, vol. 2, pp. 364367. 30 Bermejo 1990, p. 561; Benito Doménech 2001, p. 33. 31 Benito Doménech 2001, pp. 31-45; Gómez Frechina 2007, pp. 392-396. 32 Valencia 2001, pp. 192-196; Benito Doménech 2001, pp. 32-36. 33 García Marsilla 1996-1997, p. 46. 34 Madrid /Valencia 2001, pp. 185-189. On the documented works of the brothers Mateu, see also Mocholí Roselló 2009, pp. 975978. 35 Montero Tortajada 2015, pp. 196-197; Montero Tortajada 2016. I am very grateful to Encarna Montero for sharing this recent finding with me. 36 See the archival references to the abovementioned Stigmatization of St Francis, de la ma de Johannes in the collection of Joan Reixach and of the well-known St George, de ma de mestre Johannes lo gren pintor del illustre duch de Burgunya, sent from Valencia to Naples in 1444. Valencia 2001, pp. 118-120. 37 For example, this is the case for Joan Reixach in the inscription IOHANNES REXACH FECIT CIVIS VALENCIE IN ANNO M°CCCC LX° OCTAVO on the St Ursula Altarpiece of 1468 (Barcelona, MNAC). See Barcelona 1992, pp. 294-298. 38 Simonson Fuchs 1982; Berg Sobré 1989, pp. 288-297; Barcelona 1992, pp. 298-301; Barcelona/Bilbao 2003, pp. 296-301; Ruiz Quesada 2007, pp. 248-256. 39 I thank Didier Martens for drawing my attention to this. 40 Simonson Fuchs 1982, p. 50; Berg Sobré 1989, p. 295; Ruiz Quesada 2007, p. 244. 41 Salvadó et al. 2008, pp. 48-50, 56. 42 Simonson Fuchs 1982, p. 48. 43 Van der Velden 2011a, pp. 38-39; Van der Velden 2011b, pp. 140-141. 44 Dalmau’s view of a landscape is undoubtedly related to Early Netherlandish painting but the link with the Ghent Altarpiece is less convincing than the other Eyckian elements mentioned above. More precisely the seascape, the little winding river and the steep mountains on the horizon indicate that another source of inspiration must be taken into account. 45 I would like to thank Bernard Petit (Brussels, KIK-IRPA), Mireia Mestre Campà, Mireia Campuzano, Núria Pedragosa y Carme Ramells (Barcelona, Museu Nacional d’Art de Catalunya) and Rafael Cornudella Carré (Universitat Autònoma de Barcelona) for their help, which was crucial in this part of the research.

Fig. 32.1 Jan van Eyck, Virgin and Child with Saints Catherine and Michael and a Donor (the Dresden or Giustiniani Triptych), 1437, oil on oak panel, 33 x 27.5 cm (central panel), 33 x 13.5 cm (wings)

32

Jan van Eyck’s Genoese Commissions. The Lost Triptych of Battista Lomellini Maria Clelia Galassi

ABSTRACT: The majority of Van Eyck’s Italian patrons were from Genoa, reflecting the prominence of the Genoese merchants working in Bruges in the first half of the fifteenth century: Anselm Adornes bequeathed to his daughters two paintings of St Francis receiving the stigmata (most likely the paintings now in Turin and Philadelphia), which he probably inherited from his father and uncle; the coat of arms in the Dresden Triptych suggests that the altarpiece was commissioned by a member of the Giustiniani family; and Bartolomeo Facio wrote about the triptych (now lost) made for Battista Lomellini which was in the Neapolitan collection of Alfonso of Aragon in 1456. After a brief discussion of the relationship between Jan van Eyck and the Genoese community in Bruges, this paper focuses on that lost Lomellini Triptych. I will propose a reassessment of the painting’s history and its commission in light of new documentary data and reconsidered visual evidence.

—o— Introduction The excellent relationship between the Genoese community in Bruges and the dukes of Burgundy was established in the 1390s when Philip the Bold conferred on the Genoese group the status of ‘Nation’, accompanied by a series of favourable economic privileges including reduced taxation.1 This position was further strengthened by privileges granted in 1414 during the dukedom of John the Fearless and again in 1421 during the dukedom of Philip the Good.2 The volume of Genoese trade with Bruges increased significantly during Van Eyck’s lifetime, as did the presence of the Genoese in the city. According to economists, no fewer than

thirty-three Genoese ships were docked at the port of Bruges in 1427, more than twice the number recorded in 1414.3 Trading was accompanied by naval brokering and banking,4 and we know the names of around 150 Genoese bankers and traders in Bruges in the fifteenth century whose activity and occupancy is recorded in archival documentation.5 After 1422, Genoese moneylenders became supporters of the court and it is known that in 1425 members of the Genoese Nation financed Philip the Good in the war against the Kingdom of Aragon,6 among them a Giustiniani and a Lomellini, both scions of families connected with Van Eyck commissions. It is therefore possible that Genoese patrons enjoyed privileged treatment in the ordering Van Eyck’s paintings as a result of the excellent relationship they had established with the court of Philip the Good.7 The number of Genoese businessmen living in Bruges between 1431 and 1440 was the largest of the century – an aspect of relevance to this paper, given that the period includes the dates of the Giustiniani and Lomellini commissions. It is not surprising, therefore, that in 1441 the Genoese Nation decided to enlarge the loggia of the consulate, the Saaihalle in Vlamingstraat, adding a private wing to the house of the consul.8 The Donor of the Giustiniani [Dresden] Triptych Apart from the two paintings depicting St Francis receiving the stigmata that Anselm Adornes

482

maria clelia galassi

Fig. 32.2 Jan van Eyck, Virgin and Child with Saints Catherine and Michael and a Donor (the Dresden or Giustiniani Triptych), left wing

bequeathed to his two daughters and which he most likely inherited from his father (Pieter) and uncle (Jacob) respectively,9 we know of two commissions to Van Eyck that arose within the milieu of the wealthy and cultivated Genoese Nation.

The first is the small Virgin and Child with Saints Catherine and Michael and a Donor that is now housed in the Gemäldegalerie Alte Meister in Dresden, often referred to simply as the Dresden Triptych (fig. 32.1). The second is a triptych depicting the Annunciation, St John the Baptist and St Jerome that was made for Battista Lomellini; the work itself is lost but it is described in detail in Bartolomeo Facio’s Viris illustribus (1456).10 The coat of arms painted onto the upper left corner of the left wing frame of the Dresden Triptych, now definitively recognized as original, suggests that this small travelling altarpiece (33 ≈ 27.5 cm when closed), signed and dated on the frame of the centre panel by Van Eyck in 1437, was commissioned by a member of the prominent Giustiniani (‘Justiniani’) family.11 Many members of the family conducted business in Flanders during the fifteenth century – not only in Bruges but also in Middelburg and Antwerp – particularly in trading alum from the Greek island of Chios.12 The presence of the Archangel Michael behind the modish kneeling donor (fig. 32.2) suggests that the latter’s name could have been Michele, although many doubts arise when seeking to identify him. As usual, when we deal with noble Genoese families, specific people are difficult to identify due to the large number of individuals sharing the same name and doing the same things at the same time. For example, during the early decades of the fifteenth century there were five different people named Francesco Spinola active as merchants and bankers in Bruges. They included Francesco of Pietro, who lived in the city from 1420 to 1426, when he was in his early to mid twenties, and whose ledger provides reliable documentation of the business networks of this branch of the Spinola family.13 The elegant portrait of the donor of the Giustiniani [Dresden] Triptych suggests that he, like Francesco Spinola, must have been a young and successful businessman. And like the Spinolas, the Giustinianis were a large clan comprising many branches with several individuals sharing the same name, as in this case of Michele. In the past, he has

jan van eyck’s genoese commissions

been identified by Weiss as Michele son of Marco, who in 1430 lived outside Genoa and who signed a petition to the Genoese government requesting permission to live within the city.14 This document is dated seven years before the completion of the triptych, however, so we must assume that either Michele stayed longer in Bruges, or that he occasionally returned to Bruges, or that one of his relatives acted as an intermediary in commissioning the painting for him. The verisimilitude of the portrait is proof of some sort of personal contact, at some point, between the painter and the donor. On the other hand, in his recent monograph on Van Eyck, Albert Châtelet suggested that the young man could be ‘Michele of Antonio’, another Michele who is found in the Giustiniani genealogical tree in the second quarter of the fifteenth century.15 The hypothesis is acceptable, although not very compelling, since as yet we have no evidence of his connection to Flanders. In a 2011 article, Noëlle Streeton discussed the possibility that rather than a reference to donor’s name the presence of the Archangel Michael standing behind him might be a more general allusion to the devotion of the Greek Orthodox Church, connected to the Giustiniani trading society in Chios.16 Following this hypothesis, Streeton’s intention was to identify the donor – or at least the intermediary between the actual donor and the painter – as the merchant Raffaele Giustiniani, whose name is recorded several times in the 1438 ledger of the Milanese bank Filippo Borromei and Co. in Bruges.17 This fascinating document has recently been transcribed by James Bolton and Francesco Guidi Bruscoli at Queen Mary University of London and offers a rich insight into the financial milieu of the Flemish town in this period.18 Although Raffaele is not the only Giustiniani recorded in the bank’s transactions – the 1438 ledger lists the other eight members of the family – he was the only one to have held an account with this bank, and he seems to have been the only one who certainly resided in Flanders in a stable way, as confirmed by a reference to him as ‘Raffaello Giustiniani di Bruges’.

483

In reality, from other documentary sources, we know of other Giustinianis who sporadically lived in Bruges during the 1430s, such as the Domenico who is recorded as being in Bruges in 1434-1435.19 However, Raffaele, who is still recorded in Bruges in 1439 and in 1441 as merchant jenevois [sic] resident en la ville de Bruges,20 appears to have been the most prominent Giustiniani in Flanders during Van Eyck’s lifetime. Therefore, it seems likely that he had an important role in the commissioning of the triptych, as the donor or, even better, as the local intermediary with Van Eyck. As we shall see, most intriguingly, Raffaele had close business relations in Bruges with the Lomellini family, particularly with Gerolamo Lomellini, son of Battista, the latter being the donor of the lost Lomellini Triptych. The Lomellini Triptych: Circumstances of the Commission Thanks to Bartolomeo Facio’s very precise description of the painting, we know that the donor of the triptych was Battista Lomellini, whose portrait was depicted on the reverse of one wing, together with his wife, depicted symmetrically on the reverse of the other wing.21 The Lomellini family was comfortably settled in Flanders, active in trading alum, banking, brokering and insurance.22 About twenty members of the family resided for long periods in Bruges,23 such as Barnaba, consul to the Genoese Nation in 1431,24 and Egidio (‘Giles’), his eight sons and three daughters, whose attainment of burgess status was listed in the Bruges Poortenboek in 1463.25 Battista is a name that commonly occurs in the Lomellini genealogy and from documentation we know that two different men named Battista conducted business in Bruges during Van Eyck’s lifetime, Battista son of Battista, and Battista son of Giorgio.26 In his landmark study on Van Eyck and Italy, Weiss convincingly indicated Battista of Giorgio as the possible donor, considering the close and friendly relations between him and Bartolomeo Facio before the latter moved to Naples in 1444.27 As early as 1428, Battista of Giorgio began intense activity as a naval insurer of merchant ships on the routes from Genoa,

484

maria clelia galassi

Cadiz and Chios to Bruges-Middelburg-Southampton.28 This kind of business does not necessarily require a residence in Bruges but it is likely that he lived there from time to time. At any rate, we must assume that any contact with Van Eyck to settle the commission and, above all, to have a preliminary portrait produced, probably on paper, must have occurred during an unrecorded period while Battista was in Bruges. On the other hand, the portrait of his wife was probably produced based on oral descriptions, since her presence in Bruges seems unlikely. In his recent study, Albert Châtelet suggested that the triptych might have been jointly commissioned by Battista of Giorgio and his brother Gerolamo as a family piece, since St Jerome was depicted in one internal wing, symmetrically juxtaposed with St John the Baptist on the other.29 Châtelet’s hypothesis also states that Facio’s description was incomplete, and that the triptych had portraits of two couples: Battista with his second wife Caterina Lomellini on one external wing and Gerolamo with his wife Orietta Lomellini on the other. Caterina and Orietta were sisters, and according to Châtelet this double family relationship makes the hypothesis of a family commission more plausible. Personally, I doubt that Facio would have omitted such an important detail when describing a painting that he could have seen with his own eyes in Genoa, and later in Naples. It is also probable that at the date of the commission, Battista was still married to his first wife, Giacoba Gentile.30 Nevertheless, Gerolamo probably played an important role in his brother’s commission, probably as an intermediary with the painter while the triptych was being executed, given that he is recorded as being based in Bruges from March 1437.31 That year is also the date of the completion of the Giustiniani [Dresden] Triptych, and it seems likely that the two commissions arose more or less at the same time – perhaps the Lomellini commission immediately after that of Giustiniani – as a form of mutual emulation. It seems likely that this type of portable ‘travelling altarpiece’ became a fashionable status symbol

in the milieu of young and wealthy international businessmen. In 1437, Battista Lomellini was a brilliant young merchant, exactly like the donor of the Giustiniani [Dresden] Triptych. He may well have seen the Giustiniani altarpiece in Genoa or when it was still in Bruges and decided to order a similar work from Van Eyck. His brother Gerolamo, who was based more regularly in Bruges at that time, was charged with monitoring the painter’s progress, and it is perhaps for this reason that he suggested his patron saint be depicted on one of the wings. It is worth noting that Gerolamo Lomellini was an account holder at the Borromei Bank, as was Raffaele Giustiniani, the supposed donor of the Giustiniani [Dresden] Triptych. Moreover, the bank ledger records show that in July 1438, Raffaele and Gerolamo opened an account together and maintained this joint account for the entire year, with a turnover of an additional one hundred Flemish pounds.32 Although these comments do not fall under the topic of this paper, it is also interesting to note that one of Girolamo’s sons, Gregorio, documented in Bruges in 1456,33 married a Vivaldi, namely Maddalena.34 The possibility that Gregorio and Maddalena were the donors depicted by Petrus Christus with the respective coat of arms – Lomellini for the man, Vivaldi for the woman – in the triptych that is now dismembered between Washington (National Gallery of Art, wings) and Frankfurt (Städel Museum, Virgin and Child Enthroned with Saints Francis and Jerome, supposed to be the central panel, dated 1457)35 seems to be very plausible, despite requiring some further clarification. If this hypothesis could be corroborated, this second commission would be confirmation of the apparent predilection for Flemish painting by this branch of the Lomellini family. Returning to Van Eyck’s Lomellini Triptych, if my assumption is correct that a commission followed immediately after the completion of the Giustiniani [Dresden] Triptych, the painting would have arrived in Genoa between 1437 and 1438. That makes Nicole Reynaud’s idea that Barthélemy d’Eyck could have seen the Lomellini Triptych in Genoa during his

jan van eyck’s genoese commissions

stopover while travelling to Naples with René d’Anjou between April and May 1438 extremely plausible. The later 1443-1445 Triptych of the Annunciation in Aix-en-Provence (central panel still in Aix-en-Provence, Church of the Madeleine; right wing in Brussels, kmskb-mrbab and left wing in Rotterdam, Boymans-Van Beuningen Museum and Amsterdam, Rijksmuseum) would therefore represent Barthélemy’s answer to Van Eyck’s model, and could be regarded as possible visual evidence for the lost Lomellini Triptych. The Lomellini Triptych: Circumstances of the Transfer to Naples At present, how long the Lomellini Triptych remained in Genoa and when it entered Alfonso’s collection in Naples has not been established. When Bartolomeo Facio wrote his De Viris Illustribus in 1456 the painting was already in Naples and this date represents our terminus ante quem. In his 1956 article, Weiss speculated that the painting was taken to Naples and gifted to Alfonso in 1444 during a diplomatic mission from Genoa in which both Facio and Battista Lomellini participated.36 Weiss’s hypothesis is still considered to be the most plausible,37 and is often quoted as a confirmed event. According to this reconstruction, therefore, Van Eyck’s painting remained in Genoa for a very short period, less than a decade. I intend to demonstrate that a different scenario is more plausible and in order to do so I will briefly outline the political context that characterized the tense years when Alfonso of Aragon tried to become king of Naples. The 1444 mission to Naples represented a crucial moment in the perpetually stormy relationship between the Republic of Genoa and the dynasty of Aragon. Some years before, in 1435, it was a Genoese fleet, captained by Biagio Assereto, that stopped Alfonso’s first attempt to conquer the Kingdom of Naples. During the naval battle near the island of Ponza, the royal galley was captured and Alfonso and his family were taken captive. The king was transferred to Savona and later to Milan – where Duke Filippo Maria Visconti negotiated his release

485

unbeknown to the Genoese – while his father Juan was detained in Genoa. The intervention of the duke of Milan in favour of Alfonso initiated a strong Genoese rebellion against the duke himself, after which Genoa always supported King René of Anjou in the war against Alfonso.38 It was only after Alfonso was firmly established as king of Naples in 1442 that the Genoese were forced to recognize his governance. In fact, the goal of the 1444 mission was to establish a diplomatic relationship between the Republic of Genoa and the Aragon Kingdom of Naples, after many years of conflict and following the failure of a previous diplomatic mission of the Neapolitan Caraffello Caraffa and Giovanni Tedesco to Genoa in 1443. The tense atmosphere of the negotiations is described by Bartolomeo Facio, secretary of the Genoese delegation, in his Rerum gestarum Alfonsi regis libri. The Genoese delegates – Battista Lomellini and the jurist Battista Guano – had to face ‘many immoderate requests’ and refused to sign the treaty without the prior approval of the Genoese Senate. 39 The mission came close to failing again, and only after lengthy debates did the Genoese Senate decide to accept Alfonso’s conditions.40 According to the diplomatic agreements set out in the peace treaty, Genoa was required to present an annual tribute to the king in the form of a huge gold basin. This basin was the official gift, and it seems very unlikely that in this tense situation Battista Lomellini would have added, as a personal gift, a painting that he had ordered only a few years before and that was meant for his own private devotion. Visual evidence also indicates that the triptych remained in Genoa for a longer period. In fact, there are two paintings (one produced in Genoa, the other connected to Genoa), both dated 1451, which seem to be a clear reflection of an Eyckian model. Whilst I am not the first to observe this,41 I would nevertheless like to emphasize the point again, since it is sometimes neglected in critical debate. The two paintings in question are an Annunciation by Justus of Ravensburg, which is a wall painting in the Genoese cloister of Santa Maria di Castello (fig. 32.3), and St Jerome in his

486

maria clelia galassi

Fig. 32.3 Justus of Ravensburg, Annunciation, wall painting, 1451, Genoa, Santa Maria di Castello

Study by Antonio da Fabriano in the Walters Art Gallery in Baltimore (fig. 32.4). Jos Amman of Ravensburg, known in Genoa as ‘Justus de la Magna’ is recorded to have been in Genoa from March to September 1451. Three documents state that he had a workshop in the city, where two other foreign artists – Jaen de Tournai and the gilder Leon de Bruges – also worked. His recorded altarpiece for the silk merchant Antonio Caffarotto has been lost and the impressive cycle in Santa Maria di Castello – where, in addition to the Annunciation at the centre of the wall, he also painted the entire vault with figures of prophets and sibyls – is the only work of his that survives. He is recorded again in Ravensburg in 1452, but his German activity is completely obscure.42 In a recent essay, Serena Romano lengthily debated the prob-

lem of Justus’s artistic training and his cultural debts in relation to Van Eyck and Petrus Christus. Romano tends to minimize the hypothesis of any direct contact with the Flemish triptych, and her conclusions suggest that Justus could have been aware of it thanks to the diffusion of Flemish models through Provence and Rhine painters.43 Nevertheless, it is interesting to note that the Annunciation seems to reflect, albeit in a very naive way, solutions and details carefully selected from Van Eyck’s repertory, in particular in the Virgin’s pose (similar to that in the Ghent Altarpiece), the three-light window in the background (similar to that in the Virgin and Child with Chancellor Rolin) and the household objects in the niche (similar to those in the Lucca Madonna). Scholars have already highlighted the presence of these iconographic

jan van eyck’s genoese commissions

Fig. 32.4 Antonio da Fabriano, St Jerome in his Studio, 1451, tempera and oil (?) on panel, 88.4 x 52.8 cm, Baltimore, The Walters Art Gallery (inv. no. 37.439)

488

maria clelia galassi

Fig. 32.5 Justus of Ravensburg, Annunciation, detail

jan van eyck’s genoese commissions

Fig. 32.6 Justus of Ravensburg, Annunciation, detail

489

490

maria clelia galassi

Fig. 32.7 Justus of Ravensburg, Annunciation, detail

quotations, but, in my opinion, the way in which Justus tried to imitate Van Eyck in his method of rendering, through using a dry mural technique, the effects of natural light and the visual quality of different materials, are more relevant.44 It seems

that the painter tried to emulate a Van Eyck painting that was directly under his observation, and that he was able to study it very carefully in that moment. In other words, Justus’s knowledge of Van Eyck’s painting seems to have been direct and

jan van eyck’s genoese commissions

491

Fig. 32.8 Antonio da Fabriano, St Jerome in his Studio, detail of the original frame

recent, through first-hand experience of a Van Eyck painting while he was working in Santa Maria di Castello. We cannot find any other examples at this time in Genoa where the still life is so realistic (fig. 32.5) and where the transparency and light reflections and refractions (figs 32.6, 32.7), as well as cast shadows, are depicted so convincingly. It is interesting to note that in his Annunciation Justus emphasized the aspects that Bartolomeo Facio celebrated in the Lomellini Triptych: a figure of the Virgin Mary ‘notable for its grace and modesty’; an Angel Gabriel ‘of exceptional beauty and with hair surpassing reality’; ‘a library done with rare art’; the light seeming to be ‘real sun-light’.45 If we accept the hypothesis that Van Eyck’s triptych was still in Genoa in 1451, we can establish that Justus was given access to study it at Battista Lomellini’s house. Justus’s donors, the brothers Emanuele and Lionello Oliva, were in fact spice and wool merchants plying their trade from Genoa and the Middle East to Northern Europe, and belonged to the same social milieu as Battista Lomellini. The Oliva family business of providing insurance services to merchant ships travelling the Atlantic route is

also well documented and in some cases they were partners with Battista Lomellini himself.46 Antonio da Fabriano’s St Jerome in his Study, now in the Walters Art Gallery in Baltimore, (fig. 32.4), could also not have been conceived without an Eyckian model, since it exhibits a very early reflection of Flemish painting which at this date – 1451 – Antonio would not been able to find in other paintings in Genoa or central Italy. The painter produced the work when he had already returned to his hometown, Fabriano. But we know from documents that he was active in Genoa from at least 1447 to 1448, and probably longer.47 During that time it is likely that he had the opportunity to see the Lomellini Triptych.48 With its figure of St Jerome it must have been an attractive model, but the trompe-l’oeil bookshelf, so carefully described and highly appreciated by Bartolomeo Facio, must have been its most alluring feature. Antonio’s capacity to deeply understand Van Eyck’s innovation is less convincing if we compare it to Justus’s results, since Antonio’s painting always maintains a rough realism that is typical of his central Italian education and a thick consistency to the paint that

492

maria clelia galassi

is particular to his egg-tempera technique. Nevertheless, what is remarkable is Antonio’s attention to the different sources of natural light that come into the studio frontally and through side windows, and his confidence in rendering cast shadows. On the original frame, the presence of an elegant inscription with the painter’s signature clearly refers to Eyckian models, such inscriptions being very unusual in Italy at this time (fig. 32.8). Thus, we have evidence that the Lomellini Triptych was still in Genoa in 1451. In 1977 Ferdinando Bologna suggested that the painting could have been purchased by Alfonso of Aragon between 1452 and 1455,49 when the king bought other Flemish pieces of art through his agent Andreu Pol, namely a series of tapestries and in linteis picturae by Rogier van der Weyden.50 This chronology coincides perfectly with the various factors that have already been considered and it can be slightly delayed if we consider the political relationships between Genoa and the Kingdom of Naples during the 1450s. In fact, in the mid-1450s, the pacific role played by the ambassador Jacopo Bracelli, and the consistent business connection between King Alfonso and the Genoese Eliano Spinola – the holder, inter alia, of the monopoly of the coral trade – also encouraged artistic connections. We know that in 1456, the Genoese goldsmith and silversmith Simone Caldera went to Naples to offer the king an exceptionally large diamond on behalf of Eliano Spinola, and that in the same year the sculptor Domenico Gagini, after a long tenure in Genoa’s cathedral, also moved to Naples to work at the Castel Nuovo.51 Within this framework we can assume that the Lomellini Triptych also found its way to Naples, perhaps in 1456, exactly at the time that is described by Facio in Alfonso’s collection. At present, I have no evidence to explain why Battista Lomellini decided to sell – or to gift – his precious triptych to Alfonso. However, if we accept that the triptych exhibited the portrait of his first wife, it is possible that the decision was a consequence of his second marriage, which rendered the painting obsolete. Further research may reveal the date of this second marriage and perhaps provide clues for answering this question.

Conclusion My research on the lost Lomellini Triptych allows us to state certain points and to propose some hypotheses. The commission of the Lomellini Triptych, probably a fashionable ‘travelling altarpiece’, arose within the milieu of the young Genoese men conducting business in Bruges, and who were looking for voguish status symbols. The donor, Battista, probably used his brother Gerolamo as an intermediary with the painter. Gerolamo, who resided in Bruges at least in 1437 and 1438, was the business partner of Raffaele Giustiniani, who was recently proposed by Streeton to have been the donor, or at least the intermediary, of the Giustiniani [Dresden] Triptych. It seems likely that the commission for the Lomellini Triptych occurred immediately after the Giustiniani [Dresden] Triptych, between 1437 and 1438. Thus, my reconstruction is consistent with Streeton’s, and our two proposals corroborate and complement each other. Visual evidence suggests that the triptych was still in Genoa in 1451, when Justus of Ravensbourg saw and studied it whilst he was painting the Annunciation in the cloister of Santa Maria di Castello. Political events suggests it was sold – or gifted – to Alfonso of Aragon around 1455-1456, so immediately before Bartolomeo Facio had seen it in the king’s collection. Finally, if we accept the hypothesis that Justus of Ravensburg’s Annunciation was inspired by the central panel of the Lomellini Triptych, it can provide a useful visual tool for reconstructing the lost Van Eyck painting. We can deduce that its composition was similar to the Ghent Altarpiece, taking place in a house interior rather than a church, as in The Annunciation now at the National Gallery, Washington, with a three-light window giving a view the landscape in the background. Perhaps the interior was seen through the columns of a porch, with figures of prophets carved on the capitals. Since Barthélemy d’Eyck also depicted this type of carved column in his Annunciation in quite an identical manner, both artists could have taken this detail from the Lomellini Triptych.

jan van eyck’s genoese commissions

NOTES 1 Finot 1900, pp. 47-57. 2 De Simoni, Belgrano 1867, pp. 399-406. 3 Petti Balbi 1996, p. 38. 4 Heers 1971, pp. 308-321. 5 Petti Balbi 1996, pp. 81-83. 6 Finot 1900, pp. 94-95. 7 On Genoese commissions to Van Eyck, see: Algeri 1991; Parma 1999; Parma 2002; Galassi 2006 with previous bibliography. 8 Petti Balbi 2005, pp. 185-190. 9 Geirnaert 1999. To be probably identified with the small, devotional paintings in the Galleria Sabauda of Turin and in the Museum of Art of Philadelphia 10 Baxandall 1964, pp. 102-103. We don’t know when the triptych was lost. Weiss (1956, p.12) hypothesized it was brought to France in 1501 by Frederick of Aragon and was destroyed in Tours in 1504 during the fire in Frederick’s palace. 11 Neidhardt, Schölzel 2000; Neidhardt, Schölzel 2005. 12 Doehaerd, Kerremans 1952, ad indicem, under Justinianus. 13 Petti Balbi 1997. 14 Weiss 1956, pp. 1-2. 15 Châtelet 2011, pp. 147, 271. 16 Streeton 2011. 17 The ledger is held in the Borromeo-Arese family archive at the Palazzo Borromeo on Isola Bella, Libri mastri 7 and 8. 18 http://www.queenmaryhistoricalresearch.org/roundhouse/ default.aspx. 19 De Simoni, Belgrano 1867, pp. 407-408, 410; Remo 1923, p. 233. 20 De Simoni, Belgrano 1867, p. 415; Gilliodts-Van Severen 1904, pp. 639-640; Remo 1923, p. 283. 21 Baxandall 1964, pp. 102-103. 22 See the many documents collected by Doehaerd, Kerremans 1952, ad indicem. 23 Petti Balbi 1996, pp. 84, 91-94. 24 De Simoni, Belgrano 1867, p. 395. 25 Bruges, Stadsarchief, Poorterboeken, p. 317.

493

26 See the Lomellini genealogical tree in Battilana 1833, vol. III, in particular tables 6-7, 12-13, 31, 34. 27 Weiss 1956, p. 3. 28 On his activity, see the many documents collected by Doehaerd, Kerremans 1952, ad indicem. 29 Châtelet 2011, pp. 147, 271-272. 30 Battilana 1833, vol. III, table 6. 31 In a letter dated 23 March, 1437, the Signoria of Genoa recommends Girolamo Lomellini to the Duke of Burgundy. In the letter Girolamo is said to be about to leave for Flanders to do business (che va in Fiandra per interessi): De Simoni, Belgrano 1867, p. 413; Remo 1923, p. 282. 32 Streeton 2011, p. 5. 33 Petti Balbi 1996, p. 93. 34 Battilana 1833, vol. III, table 6. 35 For the triptych’s hypothetical reconstruction: Lane 1970; New York 1994, entries 12-13, pp. 131-14. 36 Weiss 1956, p. 9. 37 See for example Cornudella’s assumption: Es muy probable que esta ocasión propiciase la venta o el regalo del Tríptico Lomellini al rey (20092010, p. 48). I am grateful to Bart Fransen for bringing Cornudella’s article to my attention. 38 Balbi 1962. 39 Pietrangella 2004, p. 368. 40 Pietrangella 2004, p. 369. 41 See Parma 2002; Borchert forthcoming. 42 Algeri 1991, pp. 170-178. 43 Romano 2006. 44 On Justus’s technique: Galassi 2007. 45 I cite the translation by Baxandall 1964, p. 102. 46 Doehaerd, Kerremans 1952, pp. 409, 423-424, 428-429. 47 On Antonio da Fabriano, see Caporaletti 2011, with previous bibliography. 48 On the connection between the Lomellini Tryptich and Antonio da Fabriano’s St Jerome, see Challéat 2012, pp.56-61. 49 Bologna 1977, pp. 82-86. 50 For Alfonso’s collection of Flemish paintings and tapestries, see Cornudella 2009-2010; Challéat 2012, pp.22-23. 51 Di Fabio 2011.

Fig. 33.1 Town Hall, Bruges (photo from 1952)

33

Jan van Eyck, Polychromer (and Designer?) of Statues Ingrid Geelen

ABSTRACT: In 1435 Jan van Eyck was paid to gild and paint six stone statues and canopies for the magnificent facade of Bruges Town Hall. A distinctive symbol of the ancien régime, the Town Hall was mutilated by the French in 1792, when all the monumental portraits of the counts of Flanders were pulled down and publicly destroyed. This paper sheds light on Van Eyck’s involvement in the Town Hall’s exterior decoration and explores the question of whether other commissions entailing polychroming and/or designing sculpture can be related to him and his workshop. In this respect, the choir stalls of St Saviour’s Cathedral in Bruges seem to be of particular interest. The impressive terminal figures on the armrests show parallels with Van Eyck’s work while the misericords suggest a designer who drew his inspiration from the Brussels Annunciation by the Master of Flémalle, the Merode Altarpiece or a common model.

—o— In 1435 Jan van Eyck was paid to polychrome six statues adorning the facade of Bruges Town Hall (fig. 33.1). Built between around 1376 and 1420, by 1400 the Town Hall had already been decorated with monumental sculptures of prophets, the Virgin Mary, and Baldwin Iron Arm, the first Count of Flanders, followed by a further ten counts in turret niches. The Flemish dynastic lineage was expanded in the fifteenth century with figures of the dukes of Burgundy, and systematically supplemented with the genealogy of the Habsburg rulers until 1786. On 13 December 1792, forty-seven statues embodying over 700 years of political history were torn from their niches and smashed by soldiers and

supporters of the French Revolutionary army. Two weeks later, the remains – which had been stored in the crypt of the St Basil Chapel – were destroyed on the marketplace.1 Of the original facade decoration a mere sixteen consoles survived. Three paintings by Jan Baptiste van Meunincxhove (1672, 1696, second half of the seventeenth century, Bruges, Groeningemuseum) representing a View of the Burg give a slight impression of what the facade looked like before the destruction.2 The engraving by Jakobus Danckaert published in Het nieuw Brugsch herstelde stadhuys (1711, Bruges, Municipal Library) is also helpful as visual evidence. Moreover, Pieter Le Doulx, the Bruges painter and historian, offers the last and fairly reliable evidence of the original facade before its mutilation, as he painted a View of the Burg (c.1751, Bruges, Groeningemuseum) and made a drawing of almost every statue (ms 448, c.1751?, Bruges, Municipal Library).3 The accounts for August 1435 inform us about the circumstances of Van Eyck’s contract: Item es te wetene dat jnt jaer verleden voorwaerde ghemaect was met Jacob van Oost, Gheeraerde Metter tee ende Janne van Cutseghem steenhauwers, van viij steenen beilden te makene ende te stellene an scepenen huus, omme v lb x s. grote van elken beilde van den steenen ende van hauwene ende snidene, comt xl iiij lb gr.

496

ingrid geelen

Fig. 33.2 Compilation of the statues probably designed by Van Eyck, after the drawings by P. Le Doulx, c.1751 (?), ink and watercolour on paper, 33.6 x 20.5 cm, Bruges, Municipal Library

Item, ghegheuen meester Janne van Eick den scildere van vj van den vorseiden beilden met den tabernaclen te vergoudene ende te stofferene, van den sticke v lb gro. comt xxx lb. gr. Item, den zelven ghegheuen van overwerke ende jn hoofscheden te verdrinkene iij lb. xij s. gro. Item ghegheuen Willeme van Tonghere ende Janne van den Driessche scilders van den ij van den voorseiden viij beilden metten tabernaclen te vergoudene ende te stofferene, van der sticke v lb.gro., comt x lb.gro. Item, ghegeuen van den patronen van den voorseiden viij beilden te makene, xx s. gro. Item, ghegheuen van den huere van zekeren zeyle die ghehangen waren an de stellinghe bin der tydt dat men de voorseiden beilden stoffeirde, iij lb. x s. gro. Comt al iiij xx xj lb. x s. gro. Dus afghesleghen L lb. gro. die die van der Sluus daer toe ghegheven hebben van beteringhe, daer zij bi scepenen van Brugghe bi bliverscepe in ghewyst waren, alsoo blyct bi der rekeninghe van den jaer verleden. So hebben tresoriers betaelt xlij lb. ij s. gro.4

From these entries we learn that in the previous year, 1434, the sculptors Jacob van Oost, Gerard Mettertee and Jan van Cutsegem agreed to make eight stone statues, for five pounds groat and ten shillings a piece. Presumably they represented Margaret of Constantinople, Guy of Dampierre and Robert of Bethune, Louis of Nevers and Louis of Male, Philip the Bold and Margaret of Male, and John the Fearless (fig. 33.2).5 Six of these white stone statues, each about 170-180 centimetres tall,6 were polychromed by master Jan van Eyck and his workshop, together with the canopies. The other two were for one reason or another not finished by Van Eyck, but by William van Tongeren and Jan van den Driessche. The painters received five pounds apiece – almost as much as the cost of the statue itself. Overtime and gratuities to Van Eyck cost the city administration an additional three pounds and twelve shillings.7 Another twenty shillings groat was paid for the patterns of the eight

jan van eyck, polychromer (and designer?) of statues

sculptures. Five pounds groat is roughly equivalent to what was paid for earlier statues, such as those made in 1423, and more or less corresponds to 120 working days by a skilled master craftsman, not counting the cost of the material.8 The accounts also mention the rent paid for canvas to drape over scaffolding as a shelter while the polychroming was being executed. Considering the price of the canvas hire (three pounds, ten shillings groat), it presumably remained in place for quite a long time. The finishing of the canopies and the columns of the niches was probably carried out on the spot. But where exactly the statues were polychromed can only be conjectured. To avoid needless hazardous transports, it may well have been in the neighbourhood of the Town Hall, perhaps even in the nearby St Basil Chapel, which was the crypt of the stonemasons and was situated beneath the Chapel of the Holy Blood. For several years in the 1420s, that space had also been used

497

to store four other Town Hall statues that had not been immediately installed on the facade when they were finished. Part of the cost of the commission was defrayed by a fine of fifty pounds groat that Bruges had imposed on its port at Sluis. It was not the first time that Sluis had involuntarily contributed to the decoration of Bruges Town Hall: in 1375, due to a violation of the statues of the silversmiths’ guild, Sluis was sentenced to deliver ten finished stone statues for the facade. The sentence was eventually commuted to a fine of fifty pounds groat tornoois, which seems very cheap compared to the cost of the statues. The sculptors Jacob van Oost, Gerard Mettertee and Jan van Cutsegem had each already worked for the city of Bruges, but are not really known for numerous or prestigious commissions. Both Jacob van Oost and Jan van Cutsegem came from families of stonemasons and stone merchants.9 The sculptor

498

ingrid geelen

Gerard Mettertee carved a canopy for the statue of the Virgin on the Town Hall facade in the late 1420s, but he is not mentioned anywhere else.10 Both the painters Willem van Tongeren and Jan van den Driessche worked on several occasions as polychromers for the city council.11 Unfortunately, there is no information about the colours these painters used for the (repeatedly overpainted) Town Hall statues. In fact, except for a reference to the need for decoration with fynen goud (gold leaf) and azur (probably blue azurite), there is no specific data about the finishing of a single statue. Of the entire facade, the only thing that can be at least partly reconstructed is the polychromy of some of the oldest consoles (1376-1378; Gruuthusemuseum and Town Hall, Bruges). The accounts of 13781379 that record payment to Gillis de Man for the polychromy of twenty-eight consoles include a list of painting materials, the quantity required, and the price: gold leaf, lead white, red lead, sinople, ochres, vermilion, azure and Spanish green, along with oil, and leather for making glue.12 An examination of the remaining traces of paint shows that the appearance of the consoles was especially determined by the glittering of gold leaf.13 The 1435 accounts are not specific enough to tell us who supplied the patterns for the statues. But it is clear that Van Eyck is distinguished from the others by the title of ‘master’. His career at the ducal court and his experience in portraiture undoubtedly made him the perfect person to produce the patterns for the figures of the counts and countesses. In any case, he seems a more likely candidate than the other artists named in the contract. Important historical subjects such as series of counts or funerary monuments were outsourced only to excellent artists. Even when they were fairly standardized types, the comital statues had to look like real historical characters with a certain individuality. Illustrated chronicles or existing series of counts were important sources of inspiration. In 1365 the court painter Jan van der Asselt depicted the counts of Flanders in his murals in the chapel of the Hof ter Walle in Ghent. Subse-

quently, he painted the series of counts in Louis of Male’s chapel in the Church of Our Lady in Kortrijk (Courtrai) (1371-1382). His successor Melchior Broederlam completed these in 1406.14 Between 1390 and 1404 Broederlam also kept the series of counts in the magistrates’ chamber at Ypres up to date. Generally, patterns were carefully scrutinized. The next sculpture to be added to the facade of the Bruges Town Hall was the figure of the Duke of Burgundy, Philip the Good. In 1474 the design was entrusted to his court painter Pieter Coustain. The archives mention that the burgomasters of Bruges went with the town treasurers and other members of the town council to the painter’s house to hear his advice on een patroon te makene van den personage van hertoghe Phelips in pourtrature, omme daernaer te makene een van steene ende dat te stellene voor ‘t scepenhuus – on making a design for the portrait of Duke Philip, from which to make a stone statue intended for the Town Hall facade.15 After the statue was carved, it was presumably Pieter Coustain himself who polychromed and finished it. In about the same period (1479-1480) he repainted and regilded another five of the Town Hall statues and also the west portal. It is worth remembering that in July 1432 the burgomasters of Bruges and several members of the council had also paid a visit to the workshop of Jan van Eyck to see certain works – Ghegheven te Iohannes van Heyck ’tscilders, daer de borchmeesters ende eenighe van der wet ghinghen besien zeker weerken, den cnapen aldaer in hoofscheden, v s. gr., somme iij lib (17 July 1432).16 Van Eyck’s apprentices were given three pounds parisis in tips. Though the reason for the visit is not mentioned, it is not inconceivable that it was related to the Town Hall statues. Even if the contract was drawn up in 1434, preparations could have started earlier. The next sculptures in the sequence were the figures of Charles the Bold, Maximilian of Austria and Mary of Burgundy, who were added to the facade along with a new statue of the Virgin Mary. The patterns for these were designed by Frans van

jan van eyck, polychromer (and designer?) of statues

den Pitte, who was paid for them in 1485.17 As far as we know, Van den Pitte did not work for the court, but he was a renowned painter with a particularly long list of commissions carried out for the Bruges authorities.18 After his designs were presented to the burgomaster and treasurers, they were carved by the Brussels master Lieven van Beughem.19 Frans van den Pitte polychromed them himself in 1487.20 Van Eyck’s surviving oeuvre offers no definite evidence of his activity as a designer of sculpture. But there are some sculptures where it is tempting to see Van Eyck’s hand in the design.21 The impressive choir stalls of St Saviour’s Cathedral in Bruges are of particular interest in this respect (fig. 33.3). Two wings are preserved, and originally there was a third wing just in front of the rood screen. The magnificent figures that terminate the armrests, symbolizing the creed, show some striking similarities with Van Eyck’s oeuvre.22 The apostles, saints and prophets are all strongly portrayed with individual features, spontaneous graceful gestures and bulky, heavy garments (figs 33.4, 33.5). Their attributes and the instruments of their martyrdom are integrated in an original manner. The detailed carving, executed with the utmost care, is manifested in elements like the slender hands, the but-

499

tons on the clothes, St Catherine’s dragon and the flowers in St Dorothy’s basket. These figures are quite reminiscent of the contemplative figures on the lectern of the Singing Angels in the Ghent Altarpiece, especially in their general composition, compact pose and facial expression. The small hands and feet of the philosophers on the lectern barely emerge from garments that almost completely envelop their bodies.23 The drapery of their attire seems somewhat archaic. In comparison, the choir stall sculptures are like a modernized version of those philosophers. Some of the choir stall apostles tug their cloaks across their chest to prevent them sliding off their shoulder. This gesture, which is very clear in the figure of James the Less, occurs in the Eyckian pen drawings in the Albertina in Vienna (figs 33.6, 33.7).24 That series of twelve seated and standing apostles is regarded as a copy of work by Jan van Eyck and is dated around 14301440. Stylistically, some of the figures on the armrests can be better compared when seen in reverse. The arrangement of Peter’s arm, for example, who shows his key instead of clutching at his cloak, is comparable in composition. The female saints in the choir stalls also make elegant, sensitive gestures and are gowned in ample, carefully structured costumes with angular folds. Their introverted mood is similar to that of the apostles in the Vienna

Fig. 33.3 Choir stalls, 1435-1440, north wing, Bruges, St Saviour’s Cathedral

500

ingrid geelen

Fig. 33.4 Choir stalls, arm rest with prophet

jan van eyck, polychromer (and designer?) of statues

Fig. 33.5 Choir stalls, arm rest with Mary Magdalene

501

502

ingrid geelen

Fig. 33.6 Choir stalls, arm rest with James the Less

jan van eyck, polychromer (and designer?) of statues

503

Fig. 33.7 Circle of Jan van Eyck (?), Simon the Apostle, c.1440, pen drawing, 204 x 139 mm, Vienna, Albertina (inv. 3041)

504

ingrid geelen

drawings. An abundance of nuances can be noticed in the modelling of the sculpted heads. There is the same sort of lifelike rendering that is found in certain types in the Adoration of the Lamb. Each character is given his own expression with subtly observed features. One is typified by heavy eyelids, sunken cheeks and jutting chin; another by a wide jaw and loose skin; yet another has chubby cheeks and a downturned mouth… The media are of course difficult to compare; these apostles, saints and prophets seem a sculpted reflection of Van Eyck’s painted figures and present a close relationship in concept and approach. Some of the misericords also show interesting parallels with Van Eyck’s work. Certain of the characters depicted in them come across as more compact and schematic than the figures on the armrests, but they are very rich in detail. They include scenes from the Bible and daily life, occupations and proverbs. The woman with the severe expression, fashionably dressed in a bourrelet and a highwaisted gown with wide sleeves and a fur-lined collar, recalls the Portrait of Margaret van Eyck (1439, Bruges, Groeningemuseum) and the Portrait of Jacqueline of Bavaria, which is assumed to be a copy after Van Eyck (1450-1460, Frankfurt-am-Main, Städelsches Kunstinstitut) (fig. 33.8).25 Lively scenes of men in knee-length houppelandes recall the assiduous workmen in St Barbara. One remarkable misericord depicts the Annunciation, with Mary in a domestic setting, sitting sideways on the step next to the bench, which has lion feet and a vase on top (fig. 33.9). Although the Virgin’s features are largely eroded by wear, she seems to look up from her reading. The kneeling Archangel Gabriel leans with his left hand on his knee and blesses her with his right hand. The Brussels Annunciation (1415-1425?, Brussels, kmskb-mrbab) by the Master of Flémalle, the Merode Altarpiece (1427-1432, New York, Metropolitan Museum of Art) or a common model evidently served as a source of inspiration, as the main motifs in those paintings are reprised in the Bruges Annunciation misericord. The carved composition, however, is

Fig. 33.8 Choir stalls, misericord with woman wearing a bourrelet

more compact and reduced to its essence. Mary wears a different gown; the cushion on the bench is absent; there is no place for the central table. The Bruges misericord is the earliest preserved sculpture that adopted the Flemallesque composition, at least indirectly. Several works which copy the basic composition of the Brussels Annunciation date from the third quarter of the fifteenth century. Small, simplified bas-relief replicas appear in the Rhineland-Westphalia region and in Utrecht and testify to the influence of the Brussels type.26 Given the number of sculpted copies, Lorne Campbell has speculated that both painted Annunciations were based on a lost carved prototype.27 This idea was linked to a carved Annunciation that was installed

jan van eyck, polychromer (and designer?) of statues

505

Fig. 33.9 Choir stalls, misericord with the Annunciation

on the chimneypiece in the Town Hall in Tournai in 1424. No documents exist to tell us about the designer of this sculpture, but Maurice Houtart has suggested it could have been made from a design supplied by Robert Campin.28 Jan van Eyck certainly saw the Merode Altarpiece, from which he derived the niche with the basin and towel in the Annunciation on the exterior of the Ghent Altarpiece.29 He could have seen the painting when he, Johannes, pointre, was welcomed in Tournai, either in 1427 when he was served with wine at the Town Hall (18 October) or else five months later, on 23 March 1428.30 In the Low Countries, the impact of the Brussels Annunciation appears to have been limited. Apart from some miniatures from the workshop of the

Master of Catherine of Cleves, the only reminder of it is Petrus Christus’s Archangel Gabriel in the Berlin Annunciation.31 It is well known that Petrus Christus took over motifs previously used by Van Eyck. It could be coincidence but it is intriguing that precisely in his work there is a possible reminiscence of the Brussels Annunciation. Of course, he could have taken his inspiration directly from the original source, but perhaps he had also seen a sketch or drawing of these Annunciations in Van Eyck’s workshop. Taking all these elements into account, would it be too bold to think of Van Eyck (or his workshop) as the designer of the Bruges choir stalls? No contract or account of the commission, payment or execution of the choir stalls has sur-

506

ingrid geelen

vived. They were first mentioned in 1478, when the thirteenth Chapter of the Order of the Golden Fleece was held in the choir of St Saviour’s. The Excellente Cronike van Vlaenderen mentions the removal of the original canopies of the stalls to make way for the painted coats of arms of the knights of the Golden Fleece.32 This suggests that the choir stalls must already have been there for a while. The terminus post quem is the installation of the rood screen in 1414, which backed the third wing of the choir stalls. On grounds of costume and style, the stalls can be dated to the 1430s. Moreover, several of the subjects and motifs depicted in the misericords, such as the Wilsnacker pilgrimage sign with the three consecrated holy wafers, the moresco dance, and of course the Annunciation, are very difficult to date to any earlier than the 1430s. Fifteen oak panels from the north stalls underwent dendrochronological analysis by Pascale Fraiture (kik-irpa) working together with David Houbrechts (formerly University of Liège, currently Association du Patrimoine artistique).33 The youngest tree rings were dated between 1387 and 1408. Considering a minimum of six sapwood rings, the estimated felling period of the wood from the panels with the youngest tree rings cannot be earlier than 1414. Taking into account more sapwood rings and the time between the felling date and the actual manufacture,34 a date in the 1430s indeed seems very plausible. This exploration has brought to light important data regarding the origin of the Bruges choir stalls. The high-quality figures, their poses and expressions, as well as subject matter such as the Annunciation, all refer to an artistic context of innovation and genius. There is no definite evidence that Jan van Eyck was the inventor of the Bruges choir stall figures. However, if he was not involved, it must at least have been an artist from his close circle, perhaps even someone from his workshop, who collaborated with very skilled sculptors.

NOTES * With thanks to Lee Preedy and Alistair Watkins for the English translation and revision. 1 Jaerboeck van Buyck 1898, p. 361; Van Hese 1931, p. 83. 2 For pictures see online collection database: http:// www.vlaamsekunstcollectie.be (flemishartcollection Jan Baptiste Van Meunincxhove). 3 For pictures see online photo library: http://www.kikirpa.be (Le Doulx, Pierre). 4 State Archives Brussels: Bruges Town Archives, 2 September 1434 – 2 September 1435, fol. 61; Gilliodts-Van Severen 1871-1878, V, pp. 337-338; Weale 1908, pp. xliii-xliv; Janssens de Bisthoven 1944, pp. 12, 75. 5 Janssens de Bisthoven 1944, p. 23; Dhanens 1980, p. 151. According to Jozef Duverger the statue of John the Fearless was made earlier. The account of 1428-1429 (fol.14v) mentions the payment of 6 sc. gro. to Henricke van Versenaere omme tzweert ende dagghe te vermakene van den beilde van hertoghe Jan, staende voor tghiselhuus (‘to repair the sword and the dagger of the statue of Duke John standing in front of the town hall’), Duverger 1955, p. 88, n. 27. It seems unlikely that this account is connected to one of the facade statues. If the figure of John the Fearless had already adorned the facade in the 1420s it would have followed Colarde de Cat’s statue of Willem of Dampierre, omitting the other rulers in the dynastic sequence. The statues attributed to Van Eyck, including John the Fearless, present a stylistically coherent series in the Le Doulx drawings. 6 Deduced from the space between the console and canopy. 7 The subtotal of £91, 10 s. implies that the bookkeeper forgot to count these 12 shillings. Yet, the final sum of the account £42, 2 s. is correct. 8 From 1400 to 1487 the wages of skilled master craftsmen remained more or less unchanged, an average of about ten denier Flemish groat per day. Sosson 1977, pp. 226, 300-301. 9 Gilliodts-Van Severen 1871-1878, III, p. 490, n. 7 (Pieter Van Oost); Gilliodts-Van Severen 1871-1878, VI, pp. 158, 474 (Gregoris van Oost, steenhauwere). Jan van Cutsegem appears regularly in the Bruges municipal accounts and the records of St Donatian’s Church. He also worked for the town hall of Alost; Duclos 1910, p. 382. In the fifteenth century other stonemasons with the same name appear in the archives: Heinric van Cutsegem, Guedin van Cutsegem, Govaert and Joos van Cutsegem. 10 Gilliodts-Van Severen 1871-1878, V, pp. 330, 331. 11 Gilliodts-Van Severen 1871-1878, V, pp. 272, 334, n. 1 (Willem van Tongeren); p. 319 (Jan van den Driessche); Janssens de Bisthoven 1944, p. 32. 12 Municipal archives Bruges: Bruges Town Archives 13781379, fol. 38v; Gilliodts-Van Severen 1871-1878, III, p. 488. 13 Based on personal observations. A stringent stratigraphic examination is nonetheless needed to confirm these data. 14 Unfortunately only nineteenth-century copies of the mural paintings remain. 15 Schouteet 1989-2004, vol. 1, pp. 154-155; Campbell 1995, p. 6. 16 Weale 1908, p. xxxviii; Paviot 1990, p. 88. 17 Schouteet 1989-2004, vol. 2, p. 105: Item, betaelt Fransoys vanden Pitte, den scildere, ter causen voor zyn cost ende moyte van ghemaect te hebbene viere patroonen zo daer toebehoort, te wetene: ’t patroon van Onzer Liever Vrauwenbeilde metten reprysen ende tabernakelen der toebehorende; item de patroonen van wilen zaligher ghedinckenesse hertoghe Karel ende mevrauwe Marye van Bourgoignen ende van onzen harde gheduchten heere ende prince nu wesende den conync van den Romeynen (‘Item, paid

jan van eyck, polychromer (and designer?) of statues

to Fransoys vanden Pitte, the painter, for his expense and trouble in having created four patterns, to wit: the pattern for the statue of Our Lady with its console and canopy; item the patterns in memory of the late Duke Charles and Lady Mary of Burgundy and of the fearsome lord and prince, now King of the Romans’). 18 Schouteet 1989-2004, vol. 2, pp. 93-115. 19 Schouteet 1989-2004, vol. 2, p. 105: item, van den voors. patroonen te draghene t’ Andworpen ende aldaer te tooghene den buerchmeester van deser stede ende eenighen tresoriers (‘Item, for bringing the patterns to Antwerp and for showing them to the burgomaster of this town and some treasurers’). It is not known why the patterns were shown in Antwerp. Gilliodts-Van Severen, 1871-1878, VI, p. 483. 20 Schouteet 1989-2004, vol. 2, p. 107. 21 Dhanens 1980, p. 153. 22 The choir stalls of St Saviour’s Cathedral were extensively studied as part of my PhD, under the direction of Prof. Dr. M. P. J. Martens, University of Ghent.

507

23 For pictures see http://closertovaneyck.kikirpa.be/. 24 Benesh 1928, pl. 1-3, esp. pl. 3, no. 11. 25 Antwerp 2002, pp. 50-52. 26 Campbell 1974, p. 644, n. 79; Dijkstra 1990, pp. 172-177; Kemperdick 1997, pp. 88-92; Leeuwenberg 1965, pp. 160-161. 27 Campbell 1974, p. 644, n. 79. 28 Houtart n.d., p. 10. 29 Panofsky 1953, I, 165. 30 Paviot 1990, p. 86. 31 Dijkstra 1990, p. 177; Dijkstra 1996, pp. 95-104; Ainsworth 1994, pp. 25-65. 32 Steppe, Smeyers, Lauwerys 1973, p. 59; Devliegher 1979, p. 22. 33 Pascale Fraiture, KIK-IRPA, internal report, 2011. 34 See also Fraiture in this volume.

34

Les copies de la Vierge à l’Enfant dans une église de Jan van Eyck et le rôle de la version dessinée du Grand Curtius Valentine Henderiks

ABSTRACT: The Copies of the Virgin and Child in a Church by Jan van Eyck and the Role of the Drawn Version in the Grand Curtius Museum The drawing of the Virgin and Child in a Church is one of the most important works in the fifteenth and sixteenth century collections of the Grand Curtius museum in Liège. It is directly inspired by the small painting by Jan van Eyck in the Gemäldegalerie in Berlin, faithfully reproducing the prototype. The drawing shows a high level of finish. A second drawing and five paintings after Eyckian models are also known. These late works illustrate a characteristic archaizing phenomenon at the turn of the fifteenth century whereby artists harked back to the founders of Flemish painting. Although paintings by Van Eyck often served as sources of inspiration, few of them were wholly copied. Compared with the celebrated Ghent Altarpiece, the Berlin painting presents something of an exception. Analysing different copies of this work helps place them in relation to the prototype, and allows us to work out the role of the Curtius version. Although opinions differ on the subject, this anonymous drawing could well have served as model for artists, and in particular for the Master of 1499.

—o— La Vierge à l’Enfant dans une église (fig. 34.1) constitue une des œuvres majeures de la collection des XVe et XVIe siècles du Grand Curtius.1 Depuis 2009, le visiteur peut découvrir ce dessin d’une grande qualité d’exécution, prêté à long terme grâce à la générosité d’un collectionneur privé. Mesurant 450 ≈ 230 mm et cintré dans la partie

supérieure, le dessin est exécuté à la plume avec de l’encre bistre et certains traits sont repris à l’encre noire. L’examen du papier, en bon état de conservation malgré quelques taches d’humidité, a révélé la présence d’un filigrane représentant un pot d’étain à une anse qui permet de déterminer l’origine champenoise du support et de le dater de la fin du XVe siècle. Dans la partie inférieure droite, est apposé le monogramme JCR en référence à la collection J. C. Robinson dans laquelle il se trouvait jusqu’en 1882. L’année suivante, il est vendu à Amsterdam et passe ensuite à plusieurs reprises sur le marché de l’art pour se retrouver dans une collection privée à Bruges avant d’être finalement acquis par un collectionneur bruxellois.2 Le dessin de Liège s’inspire directement du célèbre tableau de Jan van Eyck (fig. 34.2) conservé à la Gemäldegalerie de Berlin.3 Ce précieux petit panneau, aux dimensions inférieures à celles du dessin, a été longtemps considéré comme une œuvre de jeunesse du maître. Aujourd’hui, la majorité des historiens de l’art s’accorde à le situer tardivement dans sa carrière, soit vers 1437.4 Il s’agit donc d’une œuvre de maturité qui surprend le spectateur par son caractère monumental, malgré son petit format, et par le soin extrême de son exécution picturale. Chaque détail est situé de manière différenciée dans la lumière pour rendre les textures

Fig. 34.1 Anonyme, Vierge à l’Enfant dans une église, dessin à la plume à l’encre bistre et noire, 450 x 230 mm, Bruxelles, collection privée, en prêt au Musée du Grand Curtius à Liège

les copies de la vierge à l’enfant dans une église de jan van eyck

de surface notamment dans le sol et les différents éléments de l’architecture ou encore dans la couronne de la Vierge. Parallèlement, la composition est baignée d’un éclairage enveloppant qui participe à la sensation d’unité tonale qui se dégage de l’ensemble du tableau. Cette extraordinaire acuité du détail au service d’une composition unitaire est obtenue, chez Van Eyck, par une parfaite maîtrise de la structure stratigraphique des couches picturales. Le panneau constituait sans doute, à l’origine, le volet gauche d’un diptyque présentant un portrait de donateur sur le droit aujourd’hui disparu, comme le suggère le regard de la Vierge dirigé vers la droite. Le cadre, malheureusement non original, ne permet pas de déceler la trace éventuelle de charnières. Cependant, le revers semble ne jamais avoir été peint, ce qui pourrait suggérer que l’œuvre ne constituait pas à l’origine le volet gauche d’un diptyque « pliable », mais qu’elle était plutôt destinée à être suspendue, en pendant avec un portrait de donateur.5 Cinq copies peintes de la Vierge à l’Enfant dans une église de Van Eyck sont actuellement répertoriées. Les deux plus célèbres sont celles figurant sur le volet gauche du Diptyque de Christiaan de Hondt (fig. 34.3)6 et du Diptyque Doria Pamphilj.7 La troisième est datée de 1623 et conservée dans la collection Rodríguez Bauzá à Madrid,8 tandis que la quatrième appartenait anciennement à la collection du Marquis Victoire de Heredia (fig. 34.4), laquelle fut vendue à Paris en 1912.9 La dernière d’entre elles est une enluminure de Simon Bening, conservée à Dublin et exécutée vers 1530,10 ne représentant que la Vierge à l’Enfant à mi-corps. Il faut encore noter que Jan Provost s’est également inspiré du modèle eyckien, mais en l’interprétant assez librement, dans un tableau conservé au Museo Civico de Crémone,11 Enfin, à côté du dessin de Liège, il existe une seconde version dessinée, anciennement dans la collection Paar à Vienne.12 Celle-ci fut vendue en février 1896, et sa trace en est depuis perdue. Revenons au dessin liégeois dont l’auteur a reproduit, avec beaucoup d’exactitude, le prototype eyckien. L’artiste s’intéresse aux jeux d’ombre et de

511

lumière, mais, en raison du médium utilisé, il les retranscrit avec d’autres moyens, en particulier le rendu des ombres par des zones de hachures plus ou moins appuyées selon l’intensité recherchée. Ainsi, dans le visage de la Vierge il rend avec précision l’ombre portée des cheveux sur le cou (fig. 34.5). Malgré l’attention et le soin accordé au traitement des ombres dans le dessin, les jeux de lumière, si variés dans le panneau eyckien, sont ici remplacés par des effets beaucoup plus uniformes. Le dessin du Grand Curtius présente, en outre, une série de petites différences. Ses dimensions sont plus grandes, de quatorze centimètres en hauteur et neuf en largeur, ce qui permet à son auteur de représenter une plus grande fraction de la colonne fasciculée à l’avant-plan gauche de la composition. D’autre part, la figure de la Vierge est plus trapue, son visage plus large et d’aspect moins « gothique ». Les plis des drapés sont aussi plus complexes et sa robe est désormais ornée de bijoux sur le buste. Ces différences stylistiques s’expliquent par une exécution plus tardive, à la fin du XVe siècle, comme le révèle l’examen du filigrane. Ces quelques différences relevées dans le dessin se retrouvent dans la copie peinte sur le volet gauche du diptyque conservé au Koninklijk Museum voor Schone Kunsten d’Anvers (fig. 34.3). Sur le volet droit, le donateur Christiaan de Hondt, abbé des Dunes, près de Bruges, entre 1495 et 1509, est représenté agenouillé en prière face à la Vierge à l’Enfant.13 L’œuvre porte la date de 1499 au revers du volet droit, à l’origine du nom d’emprunt du maître, créé par Max J. Friedländer sur la base du diptyque autour duquel il regroupe cinq autres tableaux.14 Pour l’historien de l’art allemand, le Maître de 1499 était actif à Bruges. Aujourd’hui, la plupart des spécialistes étendent son activité à la ville de Gand, en particulier à cause de l’influence très marquée d’Hugo van der Goes. Tous s’accordent pour reconnaître dans cet anonyme un copiste scrupuleux, dénué d’imagination et marqué d’un intérêt significatif pour l’ajout de détails décoratifs.15 Cette caractéristique se perçoit particulièrement bien sur le volet gauche. En effet, par rapport au tableau de

Fig. 34.2 Jan van Eyck, Vierge à l’Enfant dans une église, c. 1437, huile sur panneau, 31 x 14 cm, Berlin, Staatliche Museen zu Berlin, Gemäldegalerie (cat. n°525C)

les copies de la vierge à l’enfant dans une église de jan van eyck

Van Eyck et comme dans le dessin liégeois, le peintre a ajouté une série de motifs ornementaux qui confèrent un caractère plus anecdotique à la composition. Il a ainsi rehaussé le pavement du sol d’éléments décoratifs, posé au premier plan un vase avec des fleurs à la symbolique mariale et disposé bien en évidence la prière, sans doute sur parchemin et désormais encadrée, qui se trouve ici sur la première colonne et non plus sur la suivante comme chez Van Eyck et dans la copie dessinée. Le Maître de 1499 a également peint les petits anges dans le chœur de manière plus détaillée et adapté l’iconographie des vitraux. Il a en outre profondément modifié la gamme chromatique, en particulier dans les vêtements de la Vierge et dans le Calvaire du chœur désormais monochrome et présentant une plus grande fraction de la figure de saint Jean (fig. 34.6). Mais, plus fondamentalement, il a perdu toute la subtilité du tableau eyckien dans la variété du rendu des effets lumineux sur les différents éléments de la composition, ne proposant ici qu’un traitement uniforme de la lumière. Ainsi, les taches de lumière miroitant sur le sol, les modulations d’éclairage dans les voûtes du transept et du cœur, de même que les jeux de réflexion sur la couronne orfévrée de la Vierge ont disparu. La Vierge à l’Enfant d’Anvers est certes d’une exécution serrée et assez soignée, mais le style est plus linéaire et la stratigraphie picturale moins subtile dans les modulations d’épaisseur des couches de glacis. L’addition d’une quantité plus importante de blanc de plomb au ton de fond, caractéristique des peintures de la fin du XVe siècle, lui confère un aspect plus opaque. Il en résulte que le tableau n’apparaît que comme un reflet plat et édulcoré de son prototype. La différence de sensibilité entre les deux œuvres s’explique certainement par une moins grande maîtrise du Maître de 1499, mais sans doute aussi par le modèle qui a servi à l’exécution de la copie anversoise. Il est généralement admis que le Maître de 1499 s’est directement inspiré de l’original de Van Eyck. Ainsi, Yvonne Yiu suggère, sans argument véritable à l’appui, que, vers 1500, soit environ septante ans

513

Fig. 34.4 Anonyme, Vierge à l’Enfant dans une église, huile sur panneau, 40,25 x 23,5 cm, anciennement Madrid, collection Marquis de Heredia; Paris, vente Hôtel Drouot, 15.06.1912, n°22, localisation inconnue

après son exécution, le tableau eyckien se trouvait encore à Bruges ou était basé l’atelier et la clientèle principale du maître.16 Ce serait donc là que Christian de Hondt aurait vu le petit panneau, uniquement accessible à un petit nombre. C’est dans ce contexte que s’inscrirait la commande du diptyque au Maître de 1499 qui réalisa la copie en la retranscrivant dans le style plus décoratif et linéaire qui lui est propre.17 Or, une autre hypothèse, plus convaincante, a été récemment avancée par les auteurs du catalogue de l’exposition sur les diptyques à Washington et Anvers. Ces derniers

Fig. 34.3 Maître de 1499, Diptyque de Christiaan de Hondt, volet gauche, Vierge à l’Enfant dans une église, 1499, huile sur panneau, 31 x 14,5 cm, Anvers, KMSKA (inv. 255)

valentine henderiks

Fig. 34.5 Anonyme, Vierge à l’Enfant dans une église, dessin à la plume à l’encre bistre et noire, 450 x 230 mm, détail de la Vierge du Calvaire et des Marches, Bruxelles, collection privée, en prêt au Musée du Grand Curtius à Liège

les copies de la vierge à l’enfant dans une église de jan van eyck

515

dans la peinture eyckienne. Ainsi, comme dans le dessin, le panneau d’Anvers présente une plus large partie de la colonne fasciculée  à l’avant-plan. De même, le personnage de saint Jean est plus clairement visible au sein du Calvaire (fig. 34.6). Comme nous l’avons déjà souligné, la figure trapue et le visage plus large de Marie, de même que le traitement complexe des plis des drapés par rapport au tableau eyckien sont très comparables dans la copie peinte d’Anvers et le dessin. Enfin, dans ces deux versions, la robe de la Vierge est décorée sur le buste d’un collier à triple pendentif.

Fig. 34.6 Maître de 1499, Diptyque de Christiaan de Hondt, volet gauche, Vierge à l’Enfant dans une église, 1499, huile sur panneau, 31 x 14,5 cm, détail du Calvaire, Anvers, KMSKA (inv. 255)

justifient la différence de sensibilité entre les deux œuvres par le fait que l’anonyme se serait plutôt inspiré d’un modèle graphique, peut-être du dessin conservé au Grand Curtius. Ils relèvent en outre, à juste titre, que le dessin sous-jacent du panneau d’Anvers présente un aspect linéaire très comparable à celui de la version dessinée de Liège.18 Toutefois, une observation attentive du traitement des ombres suggère que l’auteur du dessin sous-jacent du panneau anversois ne peut être le même que celui du dessin liégeois. En effet, la comparaison des jeux de hachures parallèles pour indiquer les zones d’ombres, en particulier dans les marches menant au chœur, (figs 34.5, 34.7) montre qu’elles sont dessinées dans des sens opposés, suggérant que l’auteur du dessin sous-jacent travaille de la main gauche.19 Malgré cette différence, il faut cependant ajouter que les dimensions du tableau anversois se rapprochent davantage de celles du dessin que de celles du panneau de Van Eyck. En outre, on retrouve dans la copie du Maître de 1499 certains éléments déjà présents dans le dessin et absents

Que peut-on déduire de ces comparaisons ? D’une part que la copie peinte d’Anvers se rapproche davantage de la version dessinée de Liège et que cette dernière est plus proche du modèle original de Van Eyck. La présence du parchemin sur la seconde colonne et le carrelage uniforme en témoigne. Ainsi, le Maître de 1499 se serait plutôt inspiré d’une version dessinée pour réaliser sa copie, peut être celle du Curtius, ce qui pourrait expliquer les différences dans la gamme chromatique, notamment dans les vêtements, mais aussi dans le Calvaire désormais monochrome. Le volet gauche du Diptyque Doria Pamphilj constitue la seconde copie du prototype eyckien, exécutée très probablement pour Antonio Siciliano vers 1513. Ce panneau, longtemps donné à Gossart, a été récemment attribué avec beaucoup de clairvoyance à Gérard David par Maryan Ainsworth.20 Le volet droit présentant le donateur reste attribué à Gossart, sans doute avec l’aide de Simon Bening pour le paysage. La copie de David pourrait avoir été directement inspirée de l’original. En effet, la gamme chromatique y est beaucoup plus proche des autres copies, en particulier le manteau bleu de Marie ou encore le Calvaire surmontant le jubé. La composition trahit toutefois un langage déjà plus renaissant, associant un effet plastique prononcé pour la figure de la Vierge à des contours plus estompés dans le modelé de son visage aux traits assez doux. Le rapport de la figure à l’architecture se trouve « modernisé » par le recul

516

valentine henderiks

Fig. 34.7 Maître de 1499, Diptyque de Christiaan de Hondt, volet gauche, Vierge à l’Enfant dans une église, 1499, huile sur panneau, 31 x 14,5 cm, IRR, détail des marches, Anvers, KMSKA (inv. 255)

des personnages divins et un élargissement des proportions de l’église, suite à l’ajout d’une rangée de colonnes à droite. Ce faisant, le peintre perd la sensation d’espace non limité et, surtout, les effets d’éclairages propres au panneau eyckien. Antonio Siciliano, qui se trouvait à Bruges vers 1513 pour l’acquisition du Bréviaire Grimani, pourrait avoir acheté la copie du tableau eyckien peinte par Gérard David, peut être d’après l’original qui pouvait encore être présent dans la ville. Sciliano aurait ensuite commandé le panneau du donateur à Gossart, lequel se trouvait également à Bruges à cette époque ou il travaillait en collaboration avec David. Ainsi s’expliqueraient les différences entre la composition des deux volets du diptyque, entre leurs dessins sous-jacents, ainsi que les résultats de l’examen dendrochronologique suggérant que le bois des deux panneaux ne provient pas du même lot. Enfin, les cadres originaux ont disparu, mais l’absence de peinture aux revers indique que les deux œuvres étaient sans doute associées à l’origine en tant que pendants et non sous la forme d’un diptyque pliable, à l’instar du modèle eyckien.21 À défaut de document d’archives et d’indications concernant la provenance de la Vierge à l’Enfant dans une église de Berlin, il est difficile de confirmer que le Maître de 1499, Gérard David et leurs commanditaires aient vu la peinture originale. Ce que l’on peut affirmer, par contre, c’est que ce sont leurs deux copies qui ont servi de modèles aux autres répliques connues. Ainsi, celle datée de 1623 et conservée à Madrid, reproduit directement le tableau d’Anvers, comme le prouve la représentation des motifs identiques sur le carrelage ou encore dans les vitraux, même si le peintre a ici supprimé certains éléments décoratifs, tels le parchemin sur la colonne ou encore le vase fleuri dans l’angle inférieur. Quant à la miniature de Simon Bening, le style et la morphologie des visages, de même que la présence des bijoux sur la robe de Marie sont autant d’éléments tendant à indiquer que la source d’inspiration serait le modèle dessiné ou celui peint par le Maître de 1499. Enfin, la copie de la collection du Marquis Victoire de Heredia

les copies de la vierge à l’enfant dans une église de jan van eyck

(fig. 34.4) découle directement de la version légèrement modifiée du Diptyque Doria Pamphilj. Pour conclure, le dessin du musée du Grand Curtius, d’une exécution très soignée, reste, parmi l’ensemble des copies, celle qui reproduit le plus fidèlement la composition du prototype eyckien. Son caractère abouti indique qu’il s’agit sans doute d’un dessin autonome qui pourrait avoir servi de modèle à la copie peinte du Maître de 1499, même si son auteur, nous l’avons vu, ne peut être celui du diptyque d’Anvers. Ainsi, son attribution et sa datation précise demeurent, à ce jour, difficiles à déterminer. In memoriam Albert Lemeunier

NOTES 1 Anonyme, Vierge à l’Enfant dans une église, dessin à la plume à l’encre bistre et noire, 450 ≈ 230 mm, Bruxelles, collection privée, en prêt au Musée du Grand Curtius à Liège. Voir Kaemmerer 1898, p. 77; Gand 1923, cat. exp., n° 93, p. 241; Fierens-Gevaert 1905, p. 163, note 1; Weale, Brockwell 1912, p. 169; Lejeune 1956, p. 30, note 2, p. 46, fig. 18 et p. 77, fig. 35; Philippe 1960, p. 117; Friedländer 1967, p. 72; Bruges 1969, p. 211; Bruxelles 1983, n° 6, pp. 22-23; Rotterdam 2002, n° 5, pp. 22-23. 2 Dans la collection J. C. Robinson à Londres en 1882; mis en vente à Amsterdam, chez Frederic Muller, en 1883; dans la collection Sylva en 1914 et Delacre à Gand; vendu à Cologne, chez Lempertz, le 29.04.1961; marché de l’art à Bruxelles; collection privée à Bruges et enfin, collection privée à Bruxelles. 3 Jan van Eyck, Vierge à l’Enfant dans une église, c. 1437, huile sur panneau, 31 ≈ 14 cm, Berlin, Staatliche Museen zu Berlin, Gemäldegalerie (cat. n° 525C), voir : Grosshans 1996, p. 45 et p. 237, n°678.

517

4 Voir Purtle 2003, pp. 253-259. 5 Hand, Metzger, Spronk 2006-2007, pp. 1-13. 6 Maître de 1499, Diptyque de Christiaan de Hondt, 1499, huile sur panneaux, 31 ≈ 14,5 cm, Anvers, Koninklijk Museum voor Schone Kunsten (inv. 255-256-530-531), voir Vandenbroeck 1985, pp. 125-130. 7 Gérard David, Jan Gossart et Simon Bening, Diptyque Doria Pamphilj, c. 1510-1515, huile sur panneaux, 45,9 ≈ 27,5 cm, Rome, Galleria Doria Pamphilj, voir New York 2010-2011, n°7 A,B, pp. 140144. 8 Anonyme, Vierge à l’Enfant dans une église avec un donateur et des saints, 1623, huile sur panneau, 35,5 ≈ 28 cm, Madrid, collection Rodríguez Bauzá, voir Bermejo Martinez 1980, n° 7, pp. 56-57 et fig. 21, p. 253. 9 Anonyme, Vierge à l’Enfant dans une église, huile sur panneau, 40,25 ≈ 23,5 cm, anciennement Madrid, collection Marquis de Heredia; Paris, vente Hôtel Drouot, 15.06.1912, n°22, localisation inconnue, voir Bermejo Martinez 1980, n° 8, p. 57 et fig. 22, p. 253. 10 Simon Bening, Vierge à l’Enfant, Rosarium, c. 1530, 123 ≈ 85 mm, Dublin, Chester Beatty Library (W 99, f. 44v), voir Smeyers 1998, p. 442, fig. 31. 11 Jan Provost, Vierge à l’Enfant dans une église, huile sur panneau, 100 ≈ 80 cm, Crémone, Museo Civico (n° 240), voir : Friedländer 1973, n°166, p. 115 et pl. 178. 12 Anonyme, Vierge à l’Enfant dans une église, dessin à la plume, 300 ≈ 140 mm, anciennement Vienne, collection du Comte Louis Paar; Vienne, vente du 21 février 1896, n° 326; localisation actuelle inconnue, voir Weale, Brockwell 1912, p. 169. 13 Pour une étude de l’œuvre, voir Washington/Antwerp 2006, n° 21, pp. 140-149 et pp. 287-288. 14 Friedländer 1969, n° 35, p. 75 et pl. 44. 15 Eeckhout 2001, pp. 521-522. 16 Yiu 2006, pp. 115. 17 Ibid., p. 116. 18 Washington/Antwerp 2006, p. 140. 19 Nous remercions très sincèrement le Prof. Dr. Fritz Koreny de nous avoir fait part de cette observation lors du colloque. Voir à ce sujet Anvers 2002, pp. 164, 192 et 196. 20 Voir New York 2010, n° 7 A,B, pp. 140-144. 21 Ibid., p. 144.

Fig. 35.1 Turin-Milan Hours, fol. 24v, illumination on parchment, approx. 280 x 190 mm, Turin, Biblioteca nazionale universitaria, K IV 29 (destroyed)

35

Jan Van Eyck as Illuminator? Hand G of the Turin-Milan Hours Catherine Reynolds

ABSTRACT: The miniatures in the Turin-Milan Hours attributed by Hulin de Loo to ‘Hand G’ need to be considered in the context of the whole manuscript, particularly the texts written earlier for the Duke of Berry, which determined their subjects, and the constraints of the layout and illumination previously executed. The miniatures are now widely attributed to Jan van Eyck but stylistic analysis supports Hulin de Loo’s conclusion that they were not by Jan despite their Eyckian connections. They differ in figure style, in the construction and treatment of interior and exterior space, and in overall compositional principles and intentions. Hand G’s predominant role in the New York Crucifixion and Last Judgement, in oil and originally on panel, facilitates comparison with Jan’s certain works.

—o— In her masterly account of the Turin-Milan Hours for the exhibition of Flemish miniatures in Brussels and Paris, Dominique Deneffe signalled that the debate over Hubert and Jan van Eyck’s possible participation would doubtless be refuelled as the conservation project reopens the question of their contributions to the Ghent Altarpiece.1 To encourage that debate, this paper considers Jan van Eyck and the Hand G miniatures of the Turin-Milan Hours, as designated by Hulin de Loo when he allocated a letter of the alphabet to each of the hands or stylistic groups he discerned in the manuscript; H denoted a second Eyckian hand.2 The focus is on the miniatures within the manuscript, leaving aside deductions about possible historical contexts, and on style. Since the Colloquium the exhibition

curated by Stephan Kemperdick and Friso Lammertse in Rotterdam, The Road to Van Eyck, has provided an exceptional opportunity to consider the works of Hand G and Jan van Eyck together.3 A simplified summary of the Turin-Milan Hours’ complex history may be helpful. Around 1410 one part of an elaborate book of hours was completed for the Duke of Berry: this part, with the texts central to a book of hours is now in Paris (BnF, nal 3093). The other part, still unfinished, migrated to the Netherlands where it was eventually completed in the 1450s. Localization to the Netherlands rests on style, on coats of arms within two miniatures, and on the addition of a calendar indicative of the Southern Netherlands. At an unknown date this second part was divided. One section, with the hours of the Virgin’s Lamentations and prayers, was by 1713 in the library of the Duke of Savoy, with which it passed to Turin University (Biblioteca Nazionale Universitaria, k iv 29). There it was virtually destroyed in the fire of 1904, fortunately after the miniature pages had been photographed for Durrieu’s publication of 1902.4 Five of the leaves cut from the Turin volume are known, four in the Musée du Louvre, Paris (Cabinet des dessins, rf 2022-2025) and one in the J. Paul Getty Museum, Los Angeles (ms 67). The second section, containing masses, also surfaced in Italy in the eighteenth century. Published by Hulin de Loo when in the Trivulzio collection

520

catherine reynolds

in Milan, it is still known as the Milan section despite having been since 1935 in the Museo Civico d’Arte Antica in Turin (Palazzo Madama, inv. 467/M).5 The miniatures Hulin de Loo assigned to Hand G in the Turin section are: the Arrest of Christ with initial and bas-de-page (fol. 24, fig. 35.1), Christ and St Martha in St Julian’s ferry (fol. 55v, fig. 35.2), the Virgo inter virgines with bas-de-page and possibly initial (fol. 59, fig. 35.3) and the Prayer on the Shore, with a banner identifying a Bavarian count of Hainault and Holland, with initial and bas-depage (fol. 59v, fig. 35.4); in the Milan section: the Birth of the Baptist with initial and bas-de-page (fol. 93v, fig. 35.5), the Funeral Mass, with the arms of Hainault and probably Hainault quartering Holland, with initial and bas-de-page (fol. 116, fig. 35.6) and the Invention of the Cross (fol. 118).6 This last miniature will not be discussed despite its Hand G characteristics because it is apparently partly repainted, with some materials and techniques differing from those of Milan fols 93v and 116.7 Hulin de Loo also attributed to Hand G the Crucifixion and Last Judgement in New York, originally on panel (fig. 14.1).8 Figurative illumination should not be considered in isolation. While ‘the archaeology of the book’ is a simple, self-explanatory phrase, the archaeology of the Turin-Milan Hours is complex, disturbed and, for a major section, destroyed after only partial recording. Relying on strata, where the layer below will predate the layer above, archaeology has limitations when applied to books. Even when elements on the page overlap, stylistic evidence remains crucial for determining different phases of illumination and their relative and absolute dating. The material context of the Hand G miniatures can provide neither certain dates nor attributions but merits examination for what information it can supply. Hand G worked within the confines of a layout already partly executed. Of the area ruled for the text (166/168 ≈ 109/112.5 mm), the scribe had left sixteen of the twenty lines empty for each minia-

ture. Around the ruling, the bars and the borders of thick-stemmed vine leaves had been completed throughout the manuscript but only on pages with completed miniatures had the spaces left between the stems been filled with small angels, birds and butterflies. The ruling and the bars with their encroaching embellishments defined the area of the bas-de-page below the text; on Turin fol. 59 with only two lines of text, Hand G could overlap the ruling (fig. 35.3). The large initials intended for historiation had probably been designed and at least partly executed with the borders: the lines of the letter forms, which defined the pictorial field, and the gilded framing are often continuous with the stems or gilding of the adjacent bar. The miniature frames, defined by the ruling, were probably gilded at the same time, since in many instances they were left incomplete without their final ink outlining and any coloured inner framing, presumably because later miniaturists overlooked this part of their task.9 It was thus an extant frame that Hand G incorporated into his subtle interplay between threedimensional illusion and the effectively two-dimensional reality of the page, when he overlapped the lower frame and continued the half-built church for the Funeral Mass into the upper margin (fig. 35.6). He also positively exploited the potentially negative limitations of the layout to complete the Trinity manifest at Christ’s Baptism: through gesture and linking gold rays, God in the initial sends down the Dove to the bas-de-page below (fig. 35.5). Where earlier illumination had not progressed beyond drawing, it was easily altered. For the Birth of the Baptist and possibly the Funeral Mass, Hand G ignored a drawn framing arch like those in miniatures from the final stages of work for the Duke of Berry.10 He was possibly unaware of the figures planned for the borders, since they appear on only five leaves of the Turin-Milan section;11 they were added to just one leaf in the Netherlands, attributed by Hulin de Loo to Hand H (Turin, fol. 14).12 In just one instance, around the Virgo inter virgines (fig. 35.3), the original border was completely

jan van eyck as illuminator?

Fig. 35.2 Turin-Milan Hours, fol. 55v, illumination on parchment, approx. 280 x 190 mm, Turin, Biblioteca nazionale universitaria, K IV 29 (destroyed)

521

522

catherine reynolds

Fig. 35.3 Turin-Milan Hours, fol.59, illumination on parchment, approx. 280 x 190 mm, Turin, Biblioteca nazionale universitaria, K IV 29 (destroyed)

jan van eyck as illuminator?

scraped away and the bar erased except at the bottom, leaving on the right a remnant of the protruding embellishment and on the left a more substantial section of curving stems and leaves on a gold ground. Unless the adjacent bas-de-page had already been painted and so was in danger of being damaged, there would have been no reason to leave these sections. The date of the replacement border would, therefore, provide a terminus ante quem for Hand G’s work on this leaf. Described as in pinks, greens and blues,13 the new border of densely entwined acanthus stems has been compared with Middle-Rhine illumination circa 1470/1480, partly from the distinctive combination of painted motifs with line flourishing, possibly in gold, most evident in the lower margin.14 The German connection merits further investigation, although the combination, found earlier in the Rhineland, also occurs in borders localized to Utrecht and dated to the 1460s.15 These borders, however, were designed to straight edges, as had become usual throughout the Netherlands by around 1450, and given framing outlines, a marked difference from the irregular boundaries of the untrimmed replacement border. Borders designed to straight boundaries are already evident in Jean Chevrot’s Cité de Dieu, copied in 1445 (Brussels, kbr, 9015-9016).16 Often cited as Southern Netherlandish comparisons for the added border, these borders are not close in motif and lack the flourished elements.17 It is the miniature on kbr 9015, fol. 1, that connects with Turin fol. 59 through its repetition of landscape motifs from the bas-de-page; if the motifs were originated for the bas-de-page, their re-use would provide a terminus ante quem of 1445 or soon after.18 The border may have been replaced some decades later but its date and localization remain uncertain in the absence of closer parallels. The careful reproduction of the original borders in the new calendar provided in the Netherlands (Turin fols 1-12), with bas-de-pages datable to the early 1450s, reveals a very different attitude that is unlikely to be contemporaneous. The bas-de-page of Turin fol. 59 also provides a terminus post quem, though an imprecise one. Evi-

523

dently inspired by the Ghent Altarpiece where virgin saints approach the Lamb raised on an altar, the illumination has the Lamb on a hillock and the virgins stationary, their robes now trailing before their feet and with open books instead of attributes, presumably to evoke Revelation where male virgins sing to the Lamb on Mount Sion (Rev 14: 1-3). Although the Lamb is not mentioned in the brief invocation on the page, Revelation continues with the male virgins following ‘the Lamb withersoever he goeth’ (Rev 14: 4), a passage absorbed and re-gendered into a hymn much used for the office for several virgin saints;19 female virgins are accordingly shown in movement following the Lamb in the Breviary of Philippe le Bel.20 The rarity of images of virgin saints alone with the Lamb indicates that the Ghent Altarpiece both suggested the scene and provided the model, explaining why the virgins sing to the Lamb’s rump. Although the Adoration of the Lamb was most probably designed, at the very least, before Hubert’s death in 1426, it is impossible to pinpoint its date or to know whether Hand G saw the original (in Ghent?) or a pattern drawing. It is beyond the scope of this paper to pursue the wider context of the repetitions of motifs and compositions.21 Hand G worked within existing iconographical and compositional traditions, which can be traced most successfully for more usual scenes like the Arrest of Christ (Turin fol. 24).22 Even with familiar subjects, problems of survival and dating make it difficult to assess whether similarities derive from direct borrowing or from independent knowledge of a common model, complicated by the possibility that the model may have circulated in variant forms produced either by its originator or by the intermediaries responsible for its transmission. A further difficulty lies in assessing the significance of similarities rather than true repetitions: similarities may arise from shared experience of the contemporary world as well as from shared pattern books. Hand G’s subjects in the Turin-Milan Hours were defined by the elaborate compilation written

524

catherine reynolds

Fig. 35.4 Turin-Milan Hours, fol. 59v, illumination on parchment, approx. 280 x 190 mm, Turin, Biblioteca nazionale universitaria, K IV 29 (destroyed)

jan van eyck as illuminator?

for the Duke of Berry that required far more pictorial embellishment than a conventional book of hours and so went beyond the conventional for its 267 subjects.23 The texts largely followed Berry’s Petites heures (BnF, lat. 18014), itself partly modelled for texts and layout on the elaborate Savoy Hours as extended in the 1370s for Berry’s brother, Charles V.24 The requirement to provide miniatures appropriate for Berry’s texts outlasted Berry’s ownership, which needs to be remembered when making deductions about patronage and date from the content of later miniatures. That content, moreover, may not have been originated for the manuscript: illuminators were just as likely as painters to turn to existing models. The Virgo inter virgines (fig. 35.3) is not the expected miniature for an invocation of female virgin saints, a category that does not contain the Virgin Mary herself. A group of virgines without the Virgo was the obvious illustration, as in the miniature to the same invocation in Berry’s Petites heures.25 It was not, however, often required. Only lavishly decorated breviaries illustrated the office for more than one virgin saint;26 a suffrage to virgin saints as a category was not a standard inclusion in books of hours: it was one of Charles V’s additions to the Savoy Hours with a miniature of virgines alone.27 The Virgo inter virgines seems an even rarer subject in breviaries and books of hours and then accompanies Marian devotions, as befits the central focus of the subject.28 It was apparently more popular on panel, with an unquantifiable presence on cloth, at least in Germany to the east of the Netherlands, where survival rates of independent paintings from the decades round 1400 are higher.29 It is likely that Hand G, himself a panel painter, looked beyond the illuminators’ usual repertoire for a rarely required subject and found a possible, if not ideal, solution, just as the Ghent Altarpiece provided the bas-de-page. The initial, by a related hand, is also not entirely apposite, since St Ursula shelters the Pope and her betrothed along with three of her 11,000 virgins. The necessity of finding additional subjects for the bas-de-page and historiated initial

525

was doubtless a further stimulus to widen the search for models. On the verso of the leaf, an even more unusual subject was required. The miniature by Hand G known as the Prayer on the Shore (fig. 35.4) precedes a prayer that was by its nature exceptional since it is to be said by a ruler. The prayer, which also appears in the Petites heures, may have originated with Charles V’s additions to the Savoy Hours, from which very few texts were recorded. Among the few was a prayer to be said by a king of France, which was repeated in both the Petites heures and the Turin-Milan Hours (Turin fol. 77v), suggesting that the prayer on Turin fol. 59v also came from the Savoy Hours.30 Whether devised for Charles V or his brother, the prayer needed a miniature showing a ruler, irrespective of the book’s current owner, just as a king of France needed to accompany the prayer on Turin fol. 77v. This miniature space was also filled in the Netherlands: the appearance of a king of France is not sufficient to establish Charles VII of France as the patron, just as the appearance of a Bavarian count of Hainault and Holland on Turin fol. 59v does not establish his patronage.31 Unlike the careful correspondence between the French king’s miniature, bas-de-page and prayer, the miniature with the count does not obviously relate to the words below. As has been noted, the count seems to have halted for the ladies and kneeling man; the vision of God and the count’s praying hands seem superimposed on a scene invented for a different context, perhaps not unrelated to the Fishing Party (fig. 30.1).32 The uncertain focus of the composition contrasts with the coherence of the Birth of the Baptist (fig. 35.5), where everyone is in concerted movement except, appropriately, Zechariah, and with the close-knit animation of the Arrest of Christ (fig. 35.1) or the crowd in the New York Crucifixion (fig.14.1). If this is also a composition re-used as an imprecise illustration of the text, the miniature need not have been executed for, or in the lifetime of, a Bavarian Count of Hainault and Holland, although it was presumably done where this visualization of a ruler

Fig. 35.5 Turin-Milan Hours, fol. 93v, illumination on parchment, 264 x 204 mm, Turin, Museo civico d’arte antica, Palazzo Madama (inv. no. 467/M)

jan van eyck as illuminator?

was acceptable and accessible in some form. The subsequent decision to erase the border on the recto shows that by then the miniature was not thought especially important.33 Because both Durrieu and Hulin de Loo assumed that the miniature was indeed painted for the count depicted, whom they identified as William of Bavaria, they dated the Eyckian miniatures before his death in 1417. The obvious links to Eyckian works and Jan’s documented employment by John of Bavaria (1422-1425), William’s brother and successor in Holland, made attribution to the Van Eycks themselves very appealing. When panel painting was thought to have developed from manuscript illumination, it was even more appealing to find the Van Eycks’ earliest works in a manuscript: the Van Eycks’ careers could then neatly embody a vital phase in the history of art, as then understood.34 Durrieu proposed the authorship of Hubert and Jan van Eyck for the Eyckian miniatures in rather undefined terms. Hulin de Loo, more rigorous in his division of hands, proposed that Hand G was either one of the Van Eycks or their unknown Master. Having thus narrowed the field, he opted for the Van Eycks and considered whether Jan was Hand G.35 The answer was a clear ‘no’, even though Hulin do Loo was very conscious that the conclusions he had reached from prolonged study in Turin could no longer be checked against the originals and that he should be precise about any doubts or uncertainties. He found a major obstacle to identifying Hand G with Jan in the weakness of the figures and their lack of structure, something disguised by their tiny size; the larger scale of the Virgo inter virgines made their idiosyncrasies more evident and explained its apparent difference from the rest of the group.36 In general, as Hulin de Loo remarked, the faces of Hand G’s figures are unlike Jan’s as they tend to the triangular with broad cheekbones, small mouths and accentuated chins, as seen for Christ (figs 35.1, 35.2, 35.5) and in the New York Last Judgement; the slumped head of Christ in the Crucifixion is difficult to compare (fig. 14.1). Although the type

527

of Christ is not unlike the Heads of Christ by Jan known from copies, the triangular effect is lacking in the proportionately narrower Eyckian heads with wider mouths.37 The triangular faces are particularly evident for females when veils define the point; otherwise the chin is seen against a rather square jaw line. The nose tends to run into the brow to create a straightened, almost Grecian, profile, typical of Hand G’s simplified contours. This distinct female type does not appear in Jan’s panels, where a broader mouth is placed above a neatly rounded little chin, defined against a full rounded double chin that masks the jaw line. Jan’s different female type is evident even for the Virgin by a Fountain who comes closest to Hand G’s women. The contours of Hand G’s figures with their sloping shoulders are simplified into a series of often long straight lines or continuous curves. Sharply defined angles are usually avoided and so attract the eye when they appear: the angular red shapes created by the ladies and child against Elizabeth’s bed draw attention to the infant Baptist (fig. 35.5). In the Funeral Mass (fig. 35.6) the obviously elegant, simplified silhouettes of the draped women differ markedly from the more broken, particularized shape of, for example, the Virgin by a Fountain. Hand G here subsumes people and building into his play of verticals, whereas in Jan’s panels the definition of specific shapes takes precedence over geometric abstractions. The contrasting attitudes to shape and contour are exemplified by the St Michael of the Dresden Triptych, with his broken somewhat fussy shapes (fig. 22.1), and the St Michael of the New York Last Judgement, with his boldly simplified lines, predominantly curves from the smooth arcs of his body, circular shield and the rounded shapes that soften the diagonals of his wings (fig. 14.1). The angles centred on his sword, partly created by the elision of his right sleeve and wing, here accent Michael’s role as wielder of divine retribution. The abstracted, fluid contours of Hand G’s figures convey movement very effectively. Lacking clear articulation, his figures are not fixed in any

Fig. 35.6 Turin-Milan Hours, fol. 116, illumination on parchment, 264 x 204 mm, Turin, Museo civico d’arte antica, Palazzo Madama (inv. no. 467/M)

jan van eyck as illuminator?

one anatomically convincing pose and so suggest a continuing momentum; the thin arms, when freed from enveloping drapery, appear rubbery and even stretched. The ambitious but not very successful group of Peter stepping over Malchus (fig. 35.1) contrasts with the far more plausible Cain and Abel from the Van der Paele Virgin’s throne. The elastic quality of Hand G’s figures, over-evident in Peter’s foreshortened head with sideways spreading face, is shared by his dogs, notably that stretching up the church at the Funeral Mass (fig. 35.6). The different approach evident in the particularity of the Arnolfinis’ static pet is not explained by its larger size (fig. 16.1), since a similar difference emerges when Hand G’s figures are compared with the builders of St Barbara’s tower (fig. 35.7). The builders have stiffer, more angular outlines, with sharply splayed legs, suggesting not movement but a frozen pose with the next position often difficult to deduce. Their heavier awkwardness is partly due to their overlarge heads, proportions they share with the man in the red chaperon on the parapet in the Virgin and Child with Chancellor Rolin but not with Jan’s bigger non-portrait figures. Hand G’s figures are consistently differently proportioned, with a more normal height of at least seven heads. Their characteristically sloping shoulders are found on Jan’s portrait sitters as part of the fashionable silhouette of the 1420s and 1430s but among Jan’s builders only on the two whose shoulders are dragged down by the load they share. Hand G’s ability to suggest figures in movement is matched by his ability to convey the fleeting play of light, still evident in the tonal gradations and different light sources of Elizabeth’s rooms (fig. 35.5) and the church (fig. 35.6). Movements of air and water are implied by the candle flames flickering over the coffin and the highlights on the ripples of the Jordan (fig. 35.2). Hulin de Loo and Durrieu described the superb light effects of the lost miniatures: natural, around St Julian, for instance, with the setting sun catching the hilltops and castle above the spume of the breaking waves (fig. 35.2), and artificial, with, for instance, the sparks flying

529

from Malchus’s dropped torch and the lamplit windows of Jerusalem (fig. 35.1).38 The transitory phenomena, the impressions of the passing moment, surround (or surrounded) figures imbued with activity to an overall effect totally unlike Jan’s certain works. Although Hulin de Loo found strong similarities between Hand G and Jan in their treatment of landscape, there are also strong differences.39 Jan, uninterested in transitory weather effects, used high viewpoints for the landscapes behind St Barbara (fig. 35.7) and Chancellor Rolin: Hand G preferred lower viewpoints that enhance the immediacy of the depiction by bringing it closer to the viewer’s usual experience of actual landscape, a glimpse of a world continuing in space and time as waves break, the sun sets or clouds scud by. The difference is obvious between the river landscapes of the Virgin and Child with Chancellor Rolin and the Birth of the Baptist, where the bas-de-page format arguably imposed the lower viewpoint (fig. 35.5). It is, however, just as obvious when comparing the arrangements of castle, wooded landscape and river to the left of St Barbara and to the left of the miniature of St Julian, free of the restrictions of the bas-de-page (figs 35.2, 35.7). The miniatures are further distinguished by their greater consistency of viewpoint, shared with the New York Crucifixion (fig. 14.1) where the steep summit of Golgotha rationalizes the tiered figures and avoids the higher viewpoint of the Virgin and Child with Chancellor Rolin for the background river and mountains. Although both painters tend to use lower viewpoints for their figures than for their landscapes, the discrepancy is less for Hand G, who also comes closer to a single viewpoint for all his figures. We look slightly down on the seated, and so lower, Christ and Martha in Julian’s boat, whereas Jan shows both St Barbara and the little figures below her as if seen straight on (figs 35.2, 35.7). Hand G’s distinct traits in his treatment of interior space were overlooked by Hulin de Loo.40 The perspective in the Birth of the Baptist and the Funeral Mass (figs 35.5, 35.6) is much further from

530

catherine reynolds

approaching a single vanishing point construction than that of the Arnolfini Portrait, the Virgin in a Church and the Washington Annunciation with their ‘areas’ where lines converge (figs 34.2, 19.1).41 A more fluid approach apparently suited Hand G’s interiors, which are more ambitious than Jan’s. The Birth of the Baptist occupies a more complex succession of spaces than those reflected in the Arnolfinis’ mirror, conceivably in a conscious challenge to the Van Eyck composition, a possibility also suggested by a comparison of the Funeral Mass and the Virgin in a Church. In the miniature, the depth of the church is less, since building has not progressed beyond the chancel, still unscreened, yet more of its width and height are shown. The structure is clarified, and complicated, by including both crossing piers at an angle to the picture plane, emphasized by the foremost pier’s projecting over the frame. Above, in an intriguing interplay of illusions, the incomplete vaulting ribs abut the frame, while the chancel arch and unfinished brickwork continue ‘behind’ it.42 The impression of a building erected in instalments is reinforced by the shift in design between the bays of the choir and apse. The miniature could again be seen as intentionally outdoing the panel in the complexity of what is portrayed. Attributing the Virgin in a Church to Hubert van Eyck and overestimating its superficial similarity to the Funeral Mass, Hulin de Loo proposed identifying Hand G with Hubert. Other considerations apart, the costume depicted makes it doubtful that Hand G’s works were all painted in Hubert’s lifetime and undermines Hulin de Loo’s basic premise that the miniatures dated from before Jan’s certain works. Leaving aside the Prayer on the Shore, the costume in the miniatures is comparable, and presumably broadly similar in date, to that in Jan’s panels. At the Funeral Mass, the chaperons with narrow bourrelets resemble that of the Portrait of a Man (‘Léal Souvenir’) (1432; see fig. 15.7), while the chaperon tied round with twisted cornette resembles that worn by Arnolfini in Berlin (probably a few years after 1434) and the man at the

parapet in the Virgin and Child with Chancellor Rolin.43 Zechariah’s hat (fig. 35.5) is like those of some of the grander men by St Barbara’s tower (1437; fig. 35.7). Two of the female visitors to the tower and the Arnolfini wife (1434; fig. 16.1) wear robes with no central join to the bodice, parallels for those in figures 35.4 and 35.5. While fashions typical of the 1440s do not appear in the miniatures, the man below the Bad Thief in the New York Crucifixion wears a broad chaperon with projecting bourrelet of a width unlikely to predate Jan’s death in 1441 (fig. 14.1). Resembling the chaperon worn by Philip the Good in a portrait type originated by Rogier van der Weyden around 1445 and found into the 1450s, it was enlarged in painting from the underdrawn version.44 Although the bourrelet of the man to the left is not so pronounced, it is still significantly wider than those in the Funeral Mass; dating the New York paintings after the miniatures is borne out by the more successful foreshortening of the ambitiously angled heads. Even if the New York paintings predate 1441, it is difficult to integrate the Hand G works into Jan’s certain oeuvre. If they are placed early, Jan has to develop by abandoning a distinctively abstracted figure type for more individualized, convincingly structured figures of a different type. Even less plausibly, his development would also require him to lose not only the sense of movement from his figures but also the animation of his superb light and weather effects, while limiting his depictions of interior space and constructing landscapes to more obviously contrived viewpoints. If the Hand G works are deemed contemporary with Jan’s panels, Jan has to oscillate between different conceptions of the figure, different skills with landscape and interiors, and deploying or not deploying his mastery of movement and the transitory. Since the New York paintings share the miniatures’ characteristics, such radical shifts cannot be accounted for by medium. Wherever the Hand G works are placed in Jan’s career, their execution would require him to change his entire attitude to picture-making

jan van eyck as illuminator?

Fig. 35.7 Jan van Eyck, St Barbara, detail, 1437, panel, whole with frame 41.4 x 27.8 cm, Antwerp, KMSKA (inv. no. 410)

532

catherine reynolds

in an erratic fashion that defies logical explanation. Jan gave each element in his compositions its own particular appearance, creating the illusion of a reality minutely observed component by component, whereas the Hand G miniatures reveal an artist who more obviously abstracted, simplifying contours and synthesizing through patterns governed by comparatively straightforward shapes. This profound difference underlies Hulin de Loo’s evocative summary of their different approaches: Jan van Eyck had to work from life, devant le modèle posé, to achieve his apparently exactly detailed reproductions of material forms, whereas Hand G had to work essentially from memory, essentiellement de souvenir, to recreate the illusion of the fleeting effect, the transitory moment.45 To Hulin de Loo, who had studied all the originals, these were the hallmarks of two distinct artists. N OTES * This paper is indebted to Christina Currie for help with illustrations and to the many contributors to the extensive debate on the subject, which makes a full bibliography for every point impossible: references are restricted to the more significant, comprehensive or accessible publications. 1 Deneffe 2011, pp. 169-170. 2 Hulin de Loo 1911. 3 Rotterdam 2012, pp. 16-18, 59-65, 98-102, 284-289 4 Durrieu 1902; for attempts to recreate their colour, Châtelet 2005-2006, and König 2007, discussing distortions inherent in the reproductive processes and illustrating surviving fragments. 5 See especially Durrieu 1902 (new edn. Châtelet 1967), 1903 and 1922; Hulin de Loo 1911; Meiss 1967, pp. 107-133, 337-340; Delaissé 1972; Meiss 1974, pp. 108-110, 141-142; König 1998 (with reproductions of all miniatures); Châtelet 1993 (with annotated bibliography); Van Buren, Marrow, Pettenati 1996a; Marrow 2002; VillelaPetit 2011. For facsimiles of the three parts, König 1992; König et al. 1994; Van Buren, Marrow, Pettenati 1996b. Hereafter, BnF, NAL 3093 (accessible at http://gallica.bnf.fr) is referred to as Paris; Turin, BNU, K.IV.29 as Turin; Turin, Palazzo Madama, 467/M, as Milan. 6 Hulin de Loo 1911, pp. 30-31. 7 Van Buren, Marrow, Pettenati 1996a, pp. 331, 334-335, 377, 399; retouching of other Milan miniatures noted by Durrieu 1910, pp. 21-22, and Hulin de Loo 1911, pp. 77-79. 8 Hulin de Loo 1911, p. 35; discussed with the Eyckian miniatures by Durrieu 1903; see Ainsworth and Wallert respectively in this volume; for an assistant’s contribution to the Last Judgement, Buck 1995, pp. 65-72. 9 A completed gilding campaign is substantiated by the miniatures extended into the upper margin in a later campaign for Berry, obviously by the Limbourgs, Paris pp. 225, 240, and more subtly on Louvre RF 2023 (followed Turin fol. 58), associated with the Baptist Master, where the black outlining of the gilding integrates addition and original frame, Villela-Petit 2011, p. 152; for details König 1998, pp. 76, 80, 113.

10 Marijnissen 1969; Butler, Van Asperen de Boer 1989; Van Buren, Marrow, Pettenati 1996a, pp. 372-373, 376-377, 397-398; for completed arches, see Turin fols 58, 80v and Louvre RF 2023; underdrawing as evidence for attribution is beyond the scope of this paper. 11 Turin fol. 39v, RF 2024 (followed Turin fol. 75) and Milan fols 4v, 87, 120. 12 Hulin de Loo 1911, p. 35. 13 Durrieu 1910, p. 33. 14 König 1998, p. 137. 15 E.g. the Bible of Evert Zouldenbach, Vienna, Österreichische Nationalbibliothek, Cod. 2771-2772, and the Hours of Gijsbrecht van Brederode, Liège, Bibliothèque de l’Université, Wittert 13; knowledge in these circles of the miniature composition on Turin fol. 24 need not have come directly from the manuscript, Utrecht/New York 19891990, nos 6, 65. 16 Deneffe 2011, pp. 171-173. 17 König 1998, p.137. 18 For the motifs and the hand of the KBR 9015-9016 frontispieces in the Turin-Milan Hours, see also Hulin de Loo 1911, pp. 41-43, Smeyers 1989, p. 62; Châtelet 1993 pp. 77-79; Van Buren, Marrow, Pettenati 1996a, pp. 332-333; König 1998 pp.137, 266, commenting on the positioning of the virgins behind the Lamb. 19 The hymn Jesu corona virginum, ‘O Jesus, crown of virgins’, does not name the Lamb but evokes him with Qui pascis inter lilia, ‘who is pastured/fed among lilies’ (Songs 2: 16), while Quocumque pergis virgines sequuntur, ‘the virgins follow withersoever you advance’ echoes Rev 14: 4: Hi [the male virgins] sequuntur agnum quocumque abierit, Walsh 2012, pp. 194-195, 454. 20 BnF lat. 1023, fol. 516v, Paris, 1290s, accessible at http://mandragore.bnf.fr; Leroquais 1934, II, p. 473; the virgins and Lamb on fol. 40 of the Spinola Hours belong to an All Saints miniature extending across the opening (Los Angeles, J. Paul Getty Museum, Ludwig IX 18, Bruges and Ghent c.1510-1520), von Euw, Plotzek 1982, II pp. 256285. 21 On the repetition of compositions, see especially Lyna 1955; Smeyers 1988; Smeyers 1989; Marrow 1991, pp. 58-60; Van Buren, Marrow, Pettenati 1996a, pp. 315-317; Cardon 2002, pp. 328-338; Rotterdam 2012, pp. 98-102. 22 König 2005, pp. 121-123. 23 As calculated by Durrieu 1902, p.10. 24 The Savoy Hours, begun for Blanche, Countess of Savoy, c.1335-1340, with numerous small miniatures of less than the width, and under half the height, of the text column, was BNU, E.V.49, and also perished in the 1904 fire; some detached gatherings are New Haven, Yale University, Beinecke Rare Book and Manuscript Library (Beinecke), MS 390; the Petites heures has some large but mostly similarly small miniatures; the Paris-Turin-Milan miniatures occupy the full width and 4/5 of the text column. For the relationships between these and Philip the Bold’s Grandes heures (Cambridge, Fitzwilliam Museum, MS 3-1954; KBR 11035-37) and prayer book (KBR 10392) see Durrieu 1902, 1911a, 1911b, 1922; De Winter 1985, pp. 186-189; Avril, Dunlop, Yapp 1989, pp. 72-4; Wieck 1991 and 2005; Van Buren, Marrow, Pettenati 1996a, pp. 262-268; Bousmanne, Van Hoorebeeck 2000, pp. 229-242, 264-272; for the designs of Berry’s books of hours, Korteweg 2005. 25 BnF, lat. 18014, fol. 105v, illustrated Meiss 1967, fig. 133; accessible at http://gallica.bnf.fr. 26 E.g BnF, lat. 1023, fol. 516v (see n. 20); not illustrated in e.g. the Breviary of Philip the Good, c.1460 (KBR 9026 and 9511), Bousmanne, Van Hoorebeeck 2000, pp. 87-89,155-159, it occurs in some later lavish Netherlandish breviaries, e.g. the Grimani Breviary, c.15151520 (Venice, Biblioteca Nazionale Marciana, lat. I. 99, fol. 432v), Ferrari, Salmi, Mellini 1972, pl. 56. 27 For Netherlandish examples, see Baltimore Walters Art Gallery, W. 215, fol. 68v, from Flanders or Artois, c.1400-1415, and W. 164,

jan van eyck as illuminator?

fol. 158, ? Brabant, c.1430-1440, Randall 1997, I pp. 92-93, 133-134, 136, and the later Spinola Hours, fol. 269v (see n. 20); for the Savoy Hours, Beinecke, MS 390, fol. 9v, Blanchard 1910, pl. XVIII, accessible at http://brbl-dl.library.yale.edu; for a group of mid-fifteenth-century French Hours, König 1982, pp. 157, 171, 294. 28 For Netherlandish examples, see Carpentras, Bibliothèque Inguimbertine, MS 57, fol. 55v (with two saints), Bruges, c.1400-1410, Leuven 1993, pp. 97-99; accessible at http://www.enluminures.culture. fr/documentation/enlumine/fr/; The Hague, Koninklijke Bibliotheek, 135 K 40, fol. 175v, Delft, c.1480, Utrecht/New York 1989-1990, pp. 255, 277-278, accessible at http://manuscripts.kb.nl/; the Spinola Hours, fol. 64v (see n. 20); the mismatch with the text on Turin fol. 59 noted in Van Buren, Marrow, Pettenati 1996a, p. 303. 29 E.g in the Gemäldegalerie, Berlin, the Cologne triptych, c.1420, inv. 1238, and the Middle Rhine panel, c.1430, inv. 1920, Kemperdick 2010, pp.194-201, 222-226; the much repeated composition originated by Hugo van der Goes, presumably for a panel painting (Campbell 1985, pp. 47-51) was reused in the Grimani Breviary, fol. 719v, Ferrari, Salmi, Mellini 1972, pl. 91; the miniature, New York, Pierpont Morgan Library, M. 659, is attributed to Gerard David, principally a panel painter, Los Angeles/London, 2003, pp. 363-365. 30 For the texts of the prayers: Turin fol. 59v and 77v, Delisle 1884, pp. 290-291; BnF lat. 18014, fols 106 and 121v accessible at http://gallica.bnf.fr; BNU, E.V.49, p. 485/fol. 176 and its repetition at Turin fol. 77v, Delisle 1907, I, p. 210.

533

31 As stressed by Smeyers 1989, p. 61. 32 See Chavannes-Mazel in this volume; a new interpretation of the bas-de-page of Turin fol. 59v by Hugo van der Velden is forthcoming. 33 Scraping away the border had holed the parchment; the damage to the initial on the verso was unexplained, Durrieu, 1910, p. 36, and 1922, p. 109. 34 E.g. Friedländer 1967-1976, I, p. 47; for further comment, Reynolds 2000, pp. 10-12. 35 Durrieu 1903; Hulin de Loo 1911, pp. 8-9, 33-34. 36 Hulin de Loo 1911, p. iv, 34. 37 Dhanens 1980, pp. 292-294. 38 Durrieu 1903, p. 21, and 1910, pp. 42, 48, 53; Hulin de Loo 1911, p. 32. 39 Hulin de Loo 1911, p. 33. 40 Hulin de Loo 1911, pp. 33-34. 41 See Leonardi in this volume. 42 The Duke of Berry, Turin fol. 78v, provides a precedent for the image continuing beyond the frame. 43 Scott 1986, p.75; for similar twisted cornettes in manuscripts written in Brussels 1437 and Arras 1438, Van Buren 2011, B.43 and B.46, pp. 159, 163, 365. 44 Leuven 2009, pp. 280-293; Buck 1995, p.71. 45 Hulin de Loo 1911, p. 34.

Fig. 36.1 Rothschild Prayerbook, fol. 134v, Coronation of the Virgin, Gerard Horenbout, 1510s, (F. Unterkircher, Das Rothschild-Gebetbuch, die schönsten Miniaturen, Graz, 1984)

36

The Tomb-Slab of Hubert Van Eyck and Gerard Horenbout. A Tribute to the Great Ghent Master Ronald Van Belle

ABSTRACT: In the Rothschild Prayerbook is a Coronation of the Virgin miniature in typical Eyckian style by the hand of the Ghent illuminator Gerard Horenbout (c. 1465-1541). The margin shows a church interior with three people kneeling in prayer on a floor with two incised tomb-slabs with brass inlay. One of the slabs shows an emaciated rotting body partially covered with a rectangular brass plate. The similarity to the incised slab identified as that of Hubert van Eyck on display in the Museum van Stenen Voorwerpen (Museum of Stone Objects) housed in St Bavo’s Abbey in Ghent is striking. Horenbout has, for reason of space, somewhat simplified the design but the slab in the miniature essentially corresponds to the Van Eyck slab. An incised slab dating from 1490 in Soignies with the partially-preserved inlay of the deceased represented as a verminous, almost skeletal cadaver cleaned by worms gives an idea of the appearance of Hubert van Eyck’s grave slab before the loss of its inlays.

—o— A Miniature by Horenbout The Ghent painter and illuminator Gerard Horenbout (c.1465-1541) became a master of the city’s painters’ guild in 14871 and made his fame in the art of illumination. As early as 1498 he was leading a very active workshop. An illumination by his hand on fol. 134v of the Rothschild Prayerbook represents the Coronation of the Virgin in typical Eyckian style (fig. 36.1).2 The margin shows a

church interior. Steps lead to a private chapel, as appears from the uncommon pavement in blue majolica tiles and other tiles bearing the inscriptions ave and mat[er]’. The floor tiles can be compared to those on the Singing Angels and Musician Angels panels of the Ghent Altarpiece.3 Three people kneel in prayer before what we can conjecture to be an altar with a retable.4 Burning wax torches and two incised slabs inlaid with brass confirm the importance of this private chapel. Strangely enough, up to now no one seems to have paid much attention to an extraordinary detail: set into the floor on the kneeling woman’s proper left is an incised slab which shows an emaciated rotting corpse partially covered with a rectangular brass plate (fig. 36.2). The similarity to the incised slab identified as that of Hubert van Eyck, now on display in the lapidary museum of St Bavo’s Abbey in Ghent, is stunning (fig. 36.3).5 For reasons of space Horenbout has somewhat simplified the design but the miniature essentially corresponds to the slab. Scholars date the Rothschild Prayerbook to around the second decade of the sixteenth century.6 At that time Horenbout was painter to the court of Margaret of Austria in Mechelen. From 1521 until his death in 1541 he lived in England.

536

ronald van belle

Fig. 36.2 Rothschild Prayerbook, fol. 134v, Coronation of the Virgin, detail, slab with brass plate corresponding to the slab of Hubert van Eyck

Facts about the Hubert van Eyck Slab What do we know about Hubert’s tomb-slab? In 1495 the German traveller Hieronymus Münzer visited the Vijd Chapel in St John’s Church (the later St Bavo’s Cathedral) and in his diary he penned a short description of the Ghent Altarpiece, mentioning that ‘…. the master of the painting is buried in front of the altar’.7 The master in question could of course only be Hubert van Eyck (d. 18 September 1426) since Jan van Eyck was buried in St Donatian’s in Bruges. On the occasion of the twenty-third Chapter of the Order of the Golden Fleece in 1559, Lucas de Heere, in his tribute to the Ghent Altarpiece installed in the Vijd Chapel, said of Hubert’s slab, ‘He is buried there, as well as his sister’.8 In 1568 the chronicler Marcus van Vaernewyck described the slab in the Vijd Chapel as […] een witte steenen dode / in eenen Zaercsteen / die een metalen Tafelkin voor haer houdt /die dit (na die oude vlemsche carmina) in ghegraveert state […] (‘a corpse in white stone [set into] a slab, holding a metal plate before it on which old Flemish verses are engraved’).9 Karel van Mander gives additional details in his Schilder-boeck (1604): […] men in de muur een zarksteen gemetseld, en op dezelve in marmer een geraamte verbeeld ziet, houdende eenen koperen plaat waarop het volgend oud Vlaamsch Vers gelezen word […] (‘the slab has been mortared into the wall, and on it, inlaid in marble, is a skeleton holding a brass plate on which the following old

Flemish verses can be read […]’).10 The brass inlays for the central plate, the border inscription and the scrolls of the slab were lost during the religious disturbances of 1578, a fate that befell nearly all the monumental brasses in Ghent. In 1599 the church authorities asked the rightful families or descendents to restore or move the damaged tombs. In Hubert’s case there was no family left, so the slab was disposed of and reused in the north portal.11 Recovered in 1892, it was only in 1895 that the link was made with Vaernewyck’s description of Hubert van Eyck’s tomb-slab.12 The slab in the St Bavo’s Abbey lapidary museum is made of Tournai stone and is quite worn. It had a brass border fixed by means of brass dowels inserted in lead. Most of those dowels are still in place. An inscription which is not recorded was probably engraved upon this border. As the central plate or tablet was engraved with an ode to Hubert and a short epitaph, we can assume that the text on the border included the traditional epitaph with name and the date of death as well as the request to pray for the soul.13 The brass plate on the lower breast (also fixed with dowels) is unique.14 According to James Weale (1832-1917), ‘The brass tablet commemorating Hubert may possibly have been let into this slab in 1533, but the slab itself is certainly not earlier than the sixteenth century.’15 I cannot agree with Weale, because in that case there would have been a deeper recess under the tablet for the skeleton figure (as marble inlay requires a deeper excavation) which is not the case. The initial concept of the slab comprised a brass tablet on the breast. As to the dating, Weale seems to have been misled by the date of the transept’s reconstruction in 1533.16 Another unusual feature on the Hubert van Eyck slab is that each corner was decorated with a brass lozenge rather than a quatrefoil. This appears on only one other Ghent incised slab, that of Bussart van Munte and his wife Mergriete Sersanders (both d. 1414), now also in the St Bavo’s Abbey lapidary museum.17 Just such a

Fig. 36.3 Tomb-slab of Hubert van Eyck (d. 1426) with brass plate (missing), Ghent, Museum van Stenen Voorwerpen, St Bavo’s Abbey

538

ronald van belle

fifteenth century.21 It presents rhetorician characteristics; the flowing rhythm and the word play of the verses indicate a certain skill. But as the poet wanted to touch a wider audience, he did not make use of intricate language. He used simple rhyming couplets and a relatively simple vocabulary, without the ‘flowery’ expressions of rhetorician intimates. As to its content the ode is of course a typical memento mori text.

Fig. 36.4 Brass lozenge from the corner of a tomb-slab, excavated from a location once occupied by the Rich Clares in a suburb of Ghent, c.1414

brass lozenge, with the evangelist symbol of St Luke, has recently been unearthed in Gentbrugge, a suburb of Ghent, in a location occupied in the past by the Rich Clares (fig. 36.4).18 The corner lozenges on Hubert’s slab were probably also engraved with the evangelist symbols, while the quatrefoils in the middle of the long sides had the coat of arms of the painter and perhaps that of the painters guild, as was observed on the now lost slab of Jan van Eyck in Bruges.19 The mortuary figure was set into in a deeper recess, proving that it was of stone – marble, according to Van Mander. It was secured to the matrix with pitch instead of being fixed with dowels, and represented a rotting, almost skeletal body, or, as Hubert’s ode puts it, ‘food for the worms’. The Ode on Hubert’s Slab What can we learn from the text of the ode on Hubert’s slab? Clearly it was not composed by clerics, who would have used Latin and a different choice of words, but by poets, who used the vernacular – in other words, rhetoricians.20 The ode is written in Flemish in the authentic form of the

Where was Hubert van Eyck Buried? Hubert van Eyck died on 18 September 1426 but the Ghent Altarpiece was only finished in 1432. The Vijd foundation was registered on 13 May 1435.22 Joos Vijd died between March and December 143923 and his wife Elisabeth Borluut on 5 May 1443.24 We do not know where Joos was buried, probably in a family tomb in another church.25 Elisabeth had her tomb covered with a monumental brass in the church of the Augustinians where many members of her family had their graves.26 Weale suggested that when Hubert van Eyck died he was originally laid to rest in the lower church (crypt) beneath the chapel for which he had painted the altarpiece. The crypt was destroyed in 1533 in order to make place for the new aisle.27 According to others, Hubert was originally buried in the churchyard at the west end of St John’s Church. However, that cemetery was cleared around 1460 in order to make space for the construction of the high tower.28 It is very unlikely that a slab with a brass inlay would ever have been located in the cemetery. It must have been laid inside the church. According to Napoleon de Pauw, Hubert’s tomb was situated in the present transept, on the south side.29 His tomb was probably opened in 1533 due to the extension of the church. But what about the transfer of the slab to the Vijd Chapel, which must have occurred before Münzer’s visit in 1495? We probably need to make a distinction between Hubert’s actual grave, in which his bones rested, and his slab, which could have been moved quite soon to the Vijd Chapel.30

the tomb-slab of hubert van eyck and gerard horenbout

The Iconography of Shrouds and Verminous Cadavers on Slabs and Brasses Represented on Hubert van Eyck’s slab is a verminous cadaver.31 This iconography occurs in miniatures such as ‘The Three Living and the Three Dead’ in the Psalter of Queen Bonne of Luxemburg, c.1345-1350.32 Miniatures probably provided the inspiration for the representation of shrouds and verminous cadavers on slabs and brasses. The writings of the influential theologian Jean Gerson (1363-1429), author of the poem Danse macabre (1425), must also have influenced the intellectuals and the chambers of rhetoric.33 The ‘triumph of death’ and ‘dance of death’ were performed on stages and in pageants and will undoubtedly have contributed to the long-lasting success of shroud and verminous cadaver iconography.34 With the Wouter Copman brass (d. 1387) the typology of the shrouded figure was introduced on brasses and slabs in Flanders. The representation of the deceased as a corpse in a shroud, with scrolls sprouting from his mouth or held in his hands, would enjoy great success in Flanders in the following centuries.35 The verminous cadaver, a rotting body crawling with worms, seems to have been somewhat less popular. The iconography of the deceased as a dead body is present on the incised slab of Jacques Langheeraertsone, who in his will of 1394 ordered Eenen saercke ghepingeert with eenen dooden gherase (‘an engraved slab [with] an emaciated body [skeleton]’), to be placed in St Michael’s Church in Ghent.36 The corpse on Hubert’s slab (d. 1426) holds two scrolls that were once inlaid with brass. Such scrolls often occur on incised slabs and shroud brasses. They usually indicate that the deceased is praying to God or the Virgin Mary to save his soul, while in other cases a moralizing message is expressed, such as the one on the lost brass of Joannes Graumans (d. 1360) in Ghent: Si quis eat qui transieris, saepe plora sum quod eris, fueramque quod es pro me precor ora (‘If you are passing here, weep a lot, as I am what you will be, I was what you are. I ask you to pray for me’).37

539

In Bruges and its hinterland such verminous cadavers and skeleton figures were also less in favour than the shroud figures.38 An exceptional slab with the representation of two verminous cadavers partially wrapped in shrouds, dating from the second half of the fifteenth century, is preserved beneath the tower of the Church of Our Lady in Damme (fig. 36.5).39 The shrouded cadavers address moralizing messages to the viewer; on one of the very worn scrolls we can read vanitas…40 Similar

Fig. 36.5 Tomb-slab of an unknown deceased (second half of the 15th century), Damme, Church of Our Lady (rubbing: Ronald Van Belle)

ronald van belle

Fig. 36.6 Tomb-slab of an unknown deceased representing a decaying cadaver, 1490, Soignies, Church of St Vincent (rubbing: Ronald Van Belle)

the tomb-slab of hubert van eyck and gerard horenbout

Fig. 36.7 The Soignies tomb-slab, positive image of the brass inlay

moralizing texts referring to human frailty and mortality, penitential psalms or intercession requests to God or the Virgin were probably engraved on the brass scrolls of Hubert van Eyck’s slab. The incised slab in hard blue Ecaussinnes stone in St Vincent’s Church in Soignies (Hainault), memorializing an unknown person but dated 1490 and with partially preserved inlay, gives an idea of the appearance of Hubert’s slab (figs 36.6, 36.7).41 The deceased is represented as a verminous cadaver, cleaned by worms. The inlay here is of brass instead of marble. The slab of Jacob du Buisson (d. 25 June 1508) and his wife Lievyne van der Leersene (?) (d. 15…)

541

in St Bavo’s Abbey in Ghent42 could have been inspired by the Hubert van Eyck slab (fig. 36.8). Made of Tournai stone, the slab has suffered extensively following a reuse and is also badly weathered. Although the epitaph alludes to two persons only one is represented, shown as a verminous cadaver with a scroll above the head. The belly has burst open showing the intestines on which worms are feasting. The epitaph engraved along the border is interrupted at the corners by evangelist symbols.43 The text of the scroll reads: O mensche spieghelt u an my allen (‘O man, all mirror now thou to me’), which is a variant on the first verse of the ode on the brass plate of Hubert van Eyck’s slab which read: Spieghelt u an my die op my tredens (‘mirror thou on me who on me treads’). Another Ghent slab, this one commemorating Niclaus Bennis (d. 1511) and his wife Lysbette Bunnens (d. 15…), also shows two verminous cadavers (fig. 36.9).44 The figures are slightly turned towards each other and there are plants at their feet. The man is naked save for a shroud that covers his loins. His body is eaten by worms. The scroll sprouting out of his mouth has nearly disappeared and only one letter of the text remains. He holds another scroll in his right hand which reads Ansiet [den] mensch (‘behold the man’). His wife, also represented as a rotting body, is wrapped in a shroud. Both slabs confirm the lasting use of verminous cadaver iconography in Ghent and its suburbs, perhaps under influence of the Hubert van Eyck slab. So many questions remain unanswered and will probably remain as such. But the figure of Hubert van Eyck is still subject of passionate debate and the miniature by Horenbout representing his slab – or one uncannily like it – is certainly a fascinating piece to add to the file.

ronald van belle

Fig. 36.8 Tomb-slab of Jacob du Buisson (d. 25 June 1508) and his wife Lievyne van der Leersene (?) (d. 15..), previously in the Charterhouse at Rooigem near Ghent, Ghent, Museum of Stone Objects, St Bavo’s Abbey

the tomb-slab of hubert van eyck and gerard horenbout

Fig. 36.9 Tomb-slab of Niclaus Bennis (d. 1511) and his wife Lysbette Bunnens (d. 15..), Ghent, Museum of Stone Objects, St Bavo’s Abbey

544

ronald van belle

NOTES 1 Campbell, Foister 1986, pp. 719-727; Dogaer 1987, pp. 161167, with further bibliography; Los Angeles/London 2003, pp. 427-433. 2 Formerly Vienna ÖNB, Codex Vindobonensis S.N. 2844. Sold by Christie’s London on 8 July 1999 for £8,580,000. It was again offered for sale at Christie’s Rockefeller Center saleroom on 29 January 2014. For a facsimile see Unterkircher 1984. 3 See http://closertovaneyck.kikirpa.be. 4 One man wearing a robe and with a chaperon on his shoulder reads a prayer book as he kneels on a slab inlaid with a monumental brass engraved with shrouded figures. The woman on his right is not a nun, as stated by Unterkircher (1984, p. 96), but a widow wrapped in a typical Flemish cloak called a falie in Flemish (faille in French). She reads a long litany while the man on her right reads a prayer book. 5 On this slab see Van Werveke 1895-1896, pp. 1-8; Duverger 1945; Dhanens 1980, pp. 29-32; Van den Kerkhove, Baldewijns 1986, no. 79; Despodt 2000-2001, vol. 2, no. 83 and 1.3./014. 6 Unterkircher 1984, p.12. 7 Dhanens 1965, p.102. 8 Dhanens 1965, p.106. 9 Van Vaernewyck 1568, fols cxvij- cxix; Dhanens 1965, p. 113. 10 Hubertus van Eyck ligt, gelijk hier boven gezegd is, in de St. Jankerk te Gent begraven, in welke Kerk men in de muur een zarksteen gemetseld, en op dezelve in marmer een geraamte verbeeld ziet, houdende eenen koperen plaat waarop het volgend oud Vlaamsch Vers gelezen word… Van Mander 1604, 1, p. 27. 11 Mortier 1894, pp. 239-246; Van Werveke 1898, no. 73. 12 De Smet 1912, p. 504. 13 A comparable ode seem to have been present on another Ghent slab, that of Bauduin uter Volderstraten. The border inscription recorded his name and (incomplete) date of death ‘14[…]’ and further syn epitaphie was ghestelt in carmina, maer bijna verdonckert’ (‘his epitaph was set in carmina, but it was almost erased’). This must have been an additional epitaph drafted in rhyme, perhaps like Hubert van Eyck’s, see De Pauw 1896, p. 91. 14 The other example of brass inlay in a stone slab (ingelegd taverneel) from 1445-1446 to which Dhanens refers (Dhanens 1965, p. 23) is in fact a memorial set into the wall in front of the tomb (see Roggen 1938, pp. 56, 72-73, doc. XX). 15 Weale 1908, p. 8. 16 Dhanens 1965, p. 22. 17 Originally in the ambulatory of the Charterhouse church in Rooigem near Ghent. 18 Laleman 1984, pp. 61-62. 19 Dhanens 1977, p. 15. 20 On Dhanens’s opinion regarding the author of the ode and the quatrain, see Dhanens 1965, pp.13-15. By a deed of 24 April 1469 the Rethorykkamer van Sinte-Agnete (the St Agnes Chamber of Rhetoric), also known as De Bodemlooze Mande (The Bottomless Basket) or the ‘Mandisten’ for short, obtained the use of a chapel in the crypt of St John’s Church (De Schryver, Marijnissen 1952, p. 7 n. 3). In a deed dated 19 May 1475 the heirs of Elisabeth Borluut transferred the use of their chapel (the St Agnes Chapel in the crypt beneath the Vijd Chapel) to De Bodemlooze Mande (De Schryver, Marijnissen 1952, p. 7 and appendix 1; Dhanens 1965, pp. 99-101). Could the ode have been composed by the Mandisten, who since 1469 had held their

offices in St John’s Church but who had probably existed as a unofficial chamber long before (the Mandisten were only officially recognized by the city authorities on 22 October 1471, see Blommaert 1837, p. 419). 21 According to Professor Dirk Coigneau, Ghent University (email of 10 September 2012). 22 Dhanens 1965, pp. 89-93; Dhanens 1976. 23 Dequeker 2011, p. 203. 24 Dhanens 1965, p. 96. 25 His father Nicolas Vijd (d. 1412) was buried in the Charterhouse church at Rooigem, under a monumental brass engraved with his effigy. For his epitaph see University of Ghent, MS G 11767, fol. 47. 26 Dhanens 1965, p. 96 with citation of the epitaph. Despodt 2000-2001, 2, p. 41 (ref. 2.3/010). 27 Weale 1908, pp. 5-6. This was also Van Werveke’s opinion (Van Werveke, 1898). 28 Dequeker 2011, p. 185; Desmet 1935, pp. 199-201. 29 De Pauw 1896, pp. 96-97. 30 Very few tombstones remain in their original position in church floors. 31 On this iconography see for instance Oosterwijk 2005, pp. 40-80 and 133-140. 32 New York, Metropolitan Museum, inv. 69.86, fol. 322. 33 Gerson was appointed dean of the St Donatian’s in Bruges, where he remained for four years (1397-1401) (Van Steenberghe 1935, pp. 5-52). 34 In ‘t Spil van de levende ende doode (the play of the living and the dead) performed in 1429 in Veurne, actors dressed as skeletons stood up from their coffins, which were carried by members of their sodality, while exhibiting skulls. A stage play with a dance of death was performed for Philip the Good in Bruges in 1449 (Van de Velde 1855, p. 117; Viaene 1975-1976, pp. 101-105; Viaene 1977, p. 119; Enklaar 1950, p. 75; Clark 1950, p. 92). 35 For an illustration see Vermeersch 1976, 2, no. 87; Van Belle 2006, pp.148-149. 36 Coomans 1894, p. 106. 37 Despodt 2000-2001, 2, 1.2.006: brass of Joannes Graumans (d. 1360) previously in the Church of Our Lady in Ghent. 38 On shroud slabs and brasses in Bruges see Vermeersch 1976, vol.1, pp. 151-154. 39 Van Belle 2006, pp. 243-244. Unfortunately the epitaph on the border is completely worn away. 40 Inspired by Ecclesiastes 1: 2: ‘Vanity of vanities, says the Teacher, vanity of vanities! All is vanity’. Other popular warnings: ‘What you are, I once was, What I am, You will be’ or ‘Today you… tomorrow me’. Speaking corpses on slabs and brasses are typical of the second half of the fourteenth century and the fifteenth century in Flanders. 41 Belonje and Greenhill 1962, pp. 499-501. 42 From the Charterhouse in Rooigem. 43 Hier ligghen begraven /Jacob d [u Bui]sson van Valenchienne die staerf int jaer XVC achte de(n) XX dach in junis / ende joncvr(ouwe [Lievy] ne vander / lee[?]rsene zijne ghezelnede die staerf int jaer XV en(de)… see Despodt 2000-2001, 2, p. 53, no. 118. 44 Slab in the collection of the Museum of Stone Objects, St Bavo’s Abbey, Ghent, but the place of origin is unknown, see Despodt 2000-2001, 2, p. 55, no. 122.

Fig. 37.1 Jan van Eyck, The Crucifixion, metalpoint on prepared paper, 254 x 187 mm, Private collection

37

A Note on Gold and Silver in a Metalpoint Drawing by Jan van Eyck Arie Wallert

ABSTRACT: A technical study was done on a hitherto unknown Eyckian metalpoint drawing. This drawing is clearly related to the Crucifixion panel in New York. The metalpoint drawing, measuring 254 x 187 mm, must have had a specific function in the genesis of the painting. The characteristics of the present metalpoint drawing are briefly discussed in comparison with the underdrawing of Van Eyck’s St Francis Receiving the Stigmata. Our study was aimed at gaining a better understanding of the working methods used to produce the drawing. It was examined using a wide range of analytical methods, which revealed that most of the drawing was done with at least two different styluses, silverpoint and goldpoint. In both cases the metalpoint lines were not made with the pure metals, but rather with alloys. Efforts were made to quantify the metal compositions and also to visualize the different metalpoint applications on the paper by means of macro-X-ray fluorescence scanning and infrared reflectography. Hopefully the results will not only provide material for a proper assessment of the drawing itself, but also help our understanding of the relation between painting and drawing and shed more light on the process of transmission of motifs and ideas in Jan van Eyck’s workshop.

—o— Recently, a peculiar metalpoint drawing has come to light. It has convincingly been attributed to Jan van Eyck or his immediate circle.1 Examination of this drawing may shed some light on working processes in Jan van Eyck’s workshop. Measuring 254 ≈ 187 mm, it shows a large crowd of bystanders around a Crucifixion. (fig. 37.1) It bears an obvious relationship to the left wing of a diptych by Jan

van Eyck, now in the Metropolitan Museum in New York.2 The drawing has recently been the subject of technical investigation.3 In our analyses of the drawing, we had sufficient evidence to identify it as a metalpoint drawing on a linen rag paper primed with a white coating. This coating appeared to consist of hydroxyapatite, i.e. the basic structure of bone white.4 The bone white provided the substrate for the metalpoint drawing. This finding is entirely consistent with the methods described in historical sources for the preparation of paper for metalpoint drawings. Such texts prescribe the calcination of animal bones to a white powder. Mixed to a thin watery suspension with dilute gelatinous animal glue, the bone white powder was then used to give the paper a very thin coating. Under the microscope, this material on the present drawing appeared indeed as a relatively smooth layer, effectively covering the paper fibres. Once the surface of the coating was burnished to a perfect ivory-white smoothness, the papers were ready to receive the tracing with a metalpoint.5 A medieval manuscript included by Jehan le Begue in a compilation of technical treatises6 indicates that several types of metal and alloys could be used on such coatings: […] quo possis super ipsas protrahere nigro cum grossio, seu stilo auri, argenti, latonis, vel aeris (‘…so that you may be able to draw on them in black with a pencil, or with a stylus of gold, silver, bronze, or brass’).7

548

arie wallert

In this case, the drawing was made with at least two different styluses. There are lines in silverpoint and lines in goldpoint. The silverpoint lines are not made with pure silver, but rather an alloy of silver and copper. The ratio of silver to copper resembles that of the composition of one of the styluses used for the famous study for the Portrait of Cardinal Niccolò Albergati.8 The silverpoint lines on the present drawing are marked by significant amounts of mercury. Also, this element has recently been identified in other metalpoint drawings.9 Other parts of the drawing were done in goldpoint. As with the silverpoint lines, the goldpoint lines are not pure metal, but appear to consist of an alloy with fairly large amounts of copper and zinc and smaller amounts of silver.10 It was not immediately clear why Van Eyck would have used these two different styluses. Both metalpoints initially produce fairly similar grey lines. Attempts were made to map out the distribution of these elements by using a recently developed macro-X-ray fluorescence scanning technique.11 This action was not very successful. The scanning tests were performed with a collimated X-ray source rather than by a focused beam in which all the available energy is employed by capillary optics. A strong impediment to successful scanning of the drawings was that, by necessity, the dwell time in the xrf scanning experiment was very short. Other measurements have shown that a long duration is necessary for reliable detection of such nanometer-thin silver and gold lines. Furthermore, the scanning experiment was done with the use of a Rhodium anode rather than a Mo anode, which is better for detection of silver. Measurements were made on a small selected area of the drawing. From those measurements we could not draw any conclusions about the distributions of gold or silver lines over the whole drawing, or their possible function. The use of either silver or gold did not seem to be reserved for specific figures in the drawing, nor could we deduce that the use of gold represented a specific stage in the making process.12

The examination of the drawing in the infrared wavelengths turned out to be more useful.13 In doing this, a peculiar change in colour over time of both metal lines appeared to be of significance. Initially, both metalpoints would have produced similar cool grey marks on the abrasive bone white ground. It is estimated, however, that the total amount of metal particles deposited onto the bone white preparation would add up to some tens of μg per square centimetre. This would amount to just several nanometres of thickness of material. The metal deposited, therefore, has a large surface relative to its volume. Such large exposed surfaces leave the metal particles invitingly open to influences from the environment. In particular, silverpoint lines tend to react with even the smallest amounts of sulphides in the air. This results in a relatively fast change in colour. Fresh silverpoint lines on a test panel showed a notable change in colour from a cool bluish-grey to warmer purplishgrey in just two weeks. Lines of gold or gold alloys, however, remain unchanged. These changes are too subtle to be distinguished on the present drawing with the unaided eye, but can be observed with infrared reflectography. Metals – and therefore metallic lines of a metalpoint drawing – are usually not transparent to IR radiation. They tend to absorb the IR radiation, and therefore they often appear as black in the infrared reflectograms. Reconstructions in the conservation studios of the Rijksmuseum in Amsterdam showed that this was indeed the case. Initially, all metalpoint lines showed up quite nicely. If, however, the metallic silverpoint lines were exposed to H2S gases and thus effectively converted into AgS lines, the lines, no longer being metallic, became increasingly translucent to IR radiation.14 In the reflectogram of the Eyckian drawing, the lines in silverpoint, no longer purely metallic and thus no longer fully IR absorbent, are invisible. Therefore, the whole landscape and the city of Jerusalem behind the scene of the Crucifixion all but disappear in the infrared image (fig. 37.2). The lines in goldpoint, however, remain quite

a note on gold and silver in a metalpoint drawing by jan van eyck

prominently visible in the reflectogram. In fact, the lines in gold are more visible in the infrared reflectogram than in normal light. This is particularly prominent in the face of St John and figure of Mary Magdalene (figs 37.3a-b). In the reflectogram it becomes evident that the drawing has a stronger emphasis on the dramatic aspects of the event than the painting in the Metropolitan Museum. The thief on the proper right of the drawing reveals all the signs of suffering. His body – drawn in goldpoint – shows more twisting and contortion and that in his agony his left leg has come loose from his cross (figs 37.4a-b). This is

549

clearer in the reflectogram than in the drawing itself. In the painting, the bodies of the thieves to the left and right of Christ attract less attention. This dramatic style may link the drawing closer to a Crucifixion in a book of hours, now in the Österreichische Nationalbibliothek in Vienna, datable to around 1420, by the so-called Bedford Master. It lends further support to the suggestion that some of Jan van Eyck’s images could be influenced by workshop patterns from French manuscripts.15 It is not immediately clear whether goldpoint and silverpoint were used in different stages in the execution of the drawing. It is hard to imagine that

Figs 37.3a-b The Crucifixion, detail, St John and Mary Magdalene: (a) IRR; (b) normal light

550

arie wallert

Fig. 37.2 The Crucifixion, metalpoint drawing, IRR, goldpoint lines very prominently visible, silverpoint lines nearly invisible

a note on gold and silver in a metalpoint drawing by jan van eyck

551

Figs 37.4a-b The Crucifixion, detail, the bad thief on Christ’s left: (a) IRR; (b) normal light

the clubs and spears sticking out behind the group of horsemen could have been drawn first, without having the shape of the crowd below to support them. Here, it is logical to expect the silverpoint to precede the use of goldpoint. However, the position of the rider with the fur-lined bonnet, drawn primarily in silverpoint, is determined by the two figures drawn in goldpoint in front of and behind the horse’s head (figs 37.5a-b). There, one would expect the goldpoint lines to precede the silverpoint ones. The use of either silver or gold does not seem to be reserved for specific figures in the drawing. We could not deduce whether or not the use of gold represents a specific stage in the making process, i.e., that the first lines would have been done

in gold and the drawing would then be further worked up in silver, or vice versa. One possible yet unlikely explanation for the different metals is based on an intentional colour development on the part of Van Eyck. The colour would have evolved after the drawing was exposed to hydrogen sulphide gases in the environment. Thus the drawing would be marked by subtly different areas of cool, almost greenish grey and warm purplish grey. These colour differences would develop, however, after only a few days. The problem with this hypothesis is that during the act of drawing it would have been impossible to see or even predict the full extent of these future differences.

552

arie wallert

Figs 37.5a-b The Crucifixion, detail, the crowd in the background: (a) IRR; (b) normal light

The infrared reflectograms of other parts of the drawing seem to offer more plausible explanations. The clothing of the riders to the left, for instance, is densely hatched, the hatching running at acute angles almost parallel to the main contours. The changes of movement and planes in the angular pleats and folds of the gambeson worn by the rider at the foot of Christ’s Cross are executed in a dense network of subtle strokes supported by heavier but soft contours (figs 37.6a-b). The manner in which the horses and the group of women around Mary are set off against the background in the metalpoint drawing is remarkably similar to that in which – in the underdrawing – the shape of Brother Leo is set off with dense hatching against the rocky background of St Francis Receiving the Stigmata.16 The horses – one urinating in gold, discharging in silver – are similarly modelled with forceful parallel hatching indicating shadows around the bodies. The contour lines of the horses are long and powerful and filled with brisk, shorter modelling

strokes that convincingly suggest the movement of strong muscles under a smooth velvety skin. The outlines of the horses, indicated in long continuous strokes, are largely in silverpoint; the more sturdy parallel hatchings and cross-hatchings, placed rectangularly around these contours to bring the figures into relief and to indicate shadow, are primarily done in goldpoint. The lines in silverpoint tend to be longer and thinner, whereas the lines in goldpoint are shorter, broader and situated in darker, shadow-filled areas. This may be due to the characteristics of the drawing material. Using metalpoint it is not possible to obtain darker areas just by applying more pressure on the stylus. Pressing harder does not give darker lines. It merely makes grooves in the paper.17 The only way of obtaining darker lines in metalpoint is by repeatedly marking the same place, and thus depositing more metal. By its nature, a silverpoint is harder than a goldpoint. Therefore, the softer goldpoints tend to deposit more material in one single stroke than silverpoint.

a note on gold and silver in a metalpoint drawing by jan van eyck

553

Figs 37.6a-b The Crucifixion, detail, horses: (a) IRR; (b) normal light

Therefore it does not come as a surprise to find that most of the goldpoint is present in areas where more contrast is needed, namely the figure groups around the Cross and the foreground. There is no goldpoint in the low-contrast part of the hazy background with the city of Jerusalem. The use of different metals in one drawing may not be exceptional. There is convincing historical evidence that most fifteenth-century artists worked with more than one type of metal at the same time. A painting from the studio of Dieric Bouts, St Luke Painting the Virgin, shows the evangelist in the act of drawing.18 He is handling a brass stylus with different metalpoints at either end (fig. 37.7). A similar two-tipped stylus can be seen in Hugo van der Goes’s St Luke Drawing the Madonna (c.1480), in the Museu Nacional de Arte Antigua in Lisbon.19

Conclusion Infrared reflectography is a useful tool in the study of underdrawings in paintings. It may prove equally useful in the study of metalpoint drawings on paper, particularly in the differentiation between goldpoint and silverpoint. irr of metalpoint drawings may provide valuable information about workshop procedures, drawing techniques and the relationship between drawings and paintings. The study of this particular drawing, showing a – by now – familiar language of strokes, contours and hatchings, also taught us much more about problems of attribution in the complexities of Jan van Eyck’s workshop. The characteristics of the metalpoint drawing agree very well with those of underdrawings of accepted Eyckian paintings. This is particularly evident in the reflectograms of the New York painting that are discussed by Ainsworth in the present volume.

554

arie wallert

Fig. 37.7 Dieric Bouts the Elder (workshop of), St Luke Painting the Virgin, detail, oil on canvas transferred from panel, 109 x 86.5 cm, Coll. National Trust, Penrhyn Castle

a note on gold and silver in a metalpoint drawing by jan van eyck

NOTES *

For Prof. Dr J.R.J. van Asperen de Boer.

I am beholden to several people: Dr M. Ainsworth of the Metropolitan Museum New York, Dr I. Kneepkens of Amsterdam University and Prof. Dr J.P. Filedt Kok. I thank them for their helpful discussions on the function of this drawing and its different materials, as well as their opinions on attribution. I am also grateful to Dr G. Tauber, senior conservator of the Rijksmuseum, for correcting my mistakes and helping strengthen my arguments. 1 Messling 2012; Wallert 2013; see Ainsworth in this volume. 2 Jan van Eyck, Crucifixion and Last Judgement, c.1430, oil on canvas, transferred from panel, each 56.5 ≈ 19.7 cm, Metropolitan Museum, New York, Fletcher Fund, 1933 (33.92ab). Wehle, Salinger 1947; Belting, Eichberger 1983; Ainsworth 1998, pp. 86-89; see Ainsworth in this volume. 3 Wallert 2013. Polarized light microscopy (PLM) was done in plain transmitted, and polarized light, with a Zeiss Standard 17 microscope (mag. 200x). Elemental analyses with energy dispersive X-ray fluorescence spectrometry (ED-XRF) were done with an ARTAX μ-XRF spectrometer, 50kV, 600μA, Mo-anode, using a 0.060μm capillary lens. Measurements usually took 120 seconds. To increase detection of low Z elements a He-flush (1.7L/m) was applied. Light microscopy of the drawing’s surface was done with a HIROX KH-7700 digital microscope with 2.11 megapixel CCD sensor, res. max. 10.000 ≈ 10.000 pixel, magnifications 35≈ – 2000≈. X-ray diffraction (XRD) measurements were carried out on the drawing with a Siemens D8, GADDS (general area detection diffraction system) instrument with CuKα radiation and with monocapillary optics (2 Theta from 18.00º to 56.50º with 100s steptime). Other XRD analyses were carried out with 57.3 mm Gandolfi cameras with CuKα radiation, with the voltage set at 40kV and current at 30mA; exposure times varied from 1½ to 5½ hours. 4 Kay, Young, Posner 1964.

555

5 Watrous 1957, pp. 12-23. 6 Paris, Bibliotheque Nationale, MS 6741. 7 Merrifield 1849, pp. 274-279. 8 Jan van Eyck, Portrait of Cardinal Niccolò Albergati, metalpoint on prepared paper, 214 ≈ 181 mm, Kupferstichkabinett Dresden, inv. no. C 775. Reiche et al., 2005, pp. 8-13; Ketelsen et al., 2005. 9 Reiche et al., 2004. 10 Wallert 2013. 11 The instrument was a Bruker M6 Jetstream scanning instrument. The surface of the entire drawing was scanned in three cycles, with the excitation radiation of 35kV and 500μA from a Rh anode focused onto the 356 μm measuring spots with a monocapillary lens. Step size was 350 μm with mechanical res. 267. Dwell time was 50 msec. Elemental distribution maps were created for Ag, Pb, Fe, K, Au, and Ag. 12 Wallert 2013. 13 Infrared reflectography (IRR) was done with an Osiris 512 ≈ 512 infrared camera, equipped with a Hamamatsu (G11135-512DE), InGaAs image sensor allowing a 4096 ≈ 4096 pixel capture area. The sensitivity in the NIR region extended to approximately 1700nm. The instrument is equipped with a Rodagon 1:5.6 lens, f=150 mm IR. For close-up examinations, the instrument was fitted with a dedicated f=75 mm macro lens. Visible light was filtered out at 875nm with a Schott RG830 filter. 14 Tanimoto, Verri 2009. 15 Buck 1995, pp. 70-72, fig. 5. 16 Jan van Eyck, St Francis of Assisi Receiving the Stigmata, oil on panel 29.2 ≈ 33.4 cm, Turin Galleria Sabauda. Van Asperen de Boer 1997, pp. 54-55. 17 Watrous 1957, p. 23. 18 Workshop of Dieric Bouts, St Luke Painting the Virgin, oil on canvas transferred from panel 109 ≈ 86.5 cm, Coll. National Trust, Penrhyn Castle. 19 Filedt Kok 2006, p. 9.

Bibliography

ARCHIVAL SOURCES, MANUSCRIPTS AND INCUNABULA

Oxford

Bruges

Biblia pauperum, blockbook made in Germany (c.1470). The Bodleian Library, University of Oxford: Arch. G c.14.

Anonymous, Collection of sermons and meditations (fifteenth century). Openbare Bibliotheek Brugge: MS. 408.

The Hague

Anonymous, Prayer book made in West Flanders (fifteenth century). Grootseminarie ter Brugge: MS. 72-175. Anselm of Bec, collection of writings (thirteenth century). Openbare Bibliotheek Brugge: MS. 104. Bruges Tanners Guild chapel register (1479). Rijksarchief te Brugge. Bruges Tanners Guild chapel register (1504). Rijksarchief te Brugge. Member list of Our Lady of the Dry Tree (20 June 1469). Stadsarchief Brugge: Gilde Droogenboom, no. 505.

Biblia pauperum, made in Hesdin or Amiens, Rambures Master, illuminator (c.1470). Koninklijke Bibliotheek: RMMW, 10 A 15. Book of Hours made in Utrecht (c.1455-1460). Koninklijke Bibliotheek: MS. 135 E 40. Festal Missal, made in Amiens, Petrus de Raimbaucourt, illuminator (1323). Koninklijke Bibliotheek: MS. 78 D 40. Missal, use of Utrecht, made in Utrecht (c.1425-1450). Koninklijke Bibliotheek: MS. 76 D 14.

Missal from a church in Bruges (fifteenth century). Openbare Bibliotheek Brugge: MS. 314. Missal from the Convent of the Leprosy House of Mary Magdalene in Bruges (1454). Grootseminarie te Brugge: MS. 48-3.

PUBLISHED SOURCES

Missal from Ter Doest Abbey in Lissewege (1478-1492). Grootseminarie te Brugge: 49-18.

Adams 1988 Percy G. Adams, Travel Literature Through the Ages: an Anthology, New York, 1988.

Dresden, Staatliche Kunstsammlungen Dresden Archiv ,Acten directorielle Aufzeichnungen über von der Galerie=Commission gepflogene Verhandlungen und gefaßte Beschlüsse enthaltend’, 1855-1869, fols 37/38, Archive of the Staatliche Kunstsammlungen Dresden 01/ GG 8, vol. 3. Eduard Bendemann, letter to Hugo Bürkner dated 3 April 1888, ‚Tätigkeit der Galerie=Commission und die Restauration der Gemälde’, 18831891, fol. 258/259, Archive of the Staatliche Kunstsammlungen Dresden 01/GG 8, vol. 9. Mechelen, Stadsarchief Mechelen (SAM) Gyseleers-Thijs J.-F.-C. (1802-1843): Recueil chronologique d’extraits tirés tant des actes et registres reposant aux archives, que de différents ouvrages imprimés et manuscrits concernant l’histoire de la Ville et province de Malines, 99 vols. Rymenans J.B. (1748-1840): Extracten der Stadsrekeningen, van 1311 tot 1791, 5 vols. Registers van de stadsrekeningen, 1386-1387-1416-1417.

Ainsworth 1989 Maryan W. Ainsworth, ‘Northern Renaissance Drawings. A Proposed Method of Study’, in Master Drawings, 27, 1989, pp. 5-38. Ainsworth 1994 Maryan W. Ainsworth, ‘The Art of Petrus Christus’, in Maryan W. Ainsworth (ed.), Petrus Christus: Renaissance Master of Bruges, exh. cat., New York (The Metropolitan Museum of Art), 1994, pp. 25-65. Ainsworth 1995 Maryan W. Ainsworth (ed.), Petrus Christus in Renaissance Bruges: An Interdisciplinary Approach, New York/Turnhout, 1995. Ainsworth 1998 Maryan W. Ainsworth, ‘Religious Painting from about 1420 to 1500; In the Eye of the Beholder’, in Maryan W. Ainsworth and Keith Christiansen (eds), From Van Eyck to Bruegel, Early Netherlandish Painting in The Metropolitan Museum of Art, New York (The Metropolitan Museum of Art), 1998, pp. 86-89. Ainsworth 2001 Maryan W. Ainsworth (ed.), Early Netherlandish Painting at the Crossroads: A Critical Look at Current Methodologies: The Metropolitan Museum of Art Symposia, New York, 2001.

558 Ainsworth 2003 Maryan W. Ainsworth, ‘Revelations about Jan van Eyck’s “Virgin and Child with Saint Donatian and George, and the Canon van der Paele”‘, in Hélène Verougstraete and Roger Van Schoute (eds), Le dessin sousjacent et la technologie dans la peinture; Jérôme Bosch et son entourage et autres études, Colloque XIV Bruges/Rotterdam, 13-15 septembre 2001, Louvain/Paris/Dudley, MA, 2003, pp. 273-285. Ainsworth 2004 Maryan W. Ainsworth, ‘’Á La façon Grèque’ The Encounter of Northern Renaissance Artists with Byzantine Icons’, in Helen Evans (ed.), Byzantium: Faith and Power (1261 - 1557), exh. cat., New York (The Metropolitan Museum of Art), 2004, pp. 545-555. Ainsworth, Christiansen 1999 Maryan W. Ainsworth and Keith Christiansen (eds), From Van Eyck to Bruegel . Early Netherlandish Painting in the Metropolitan Museum of Art, New York, 1999. Algeri 1991 Giuliana Algeri, ‘1440-1460: la Liguria, il mondo fiammingo e la “cultura mediterranea”‘, in Giuliana Algeri and Anna De Floriani (eds), La pittura in Ligiria. Il Quattrocento, Genoa, 1991, pp. 161-226. Almudín 2007 Eduard Mira and An Delva (eds), A la búsqueda del Toisón de Oro. La Europa de los Príncipes. La Europa de las ciudades, vol. 2 (Obras en exposición), exh. cat., Almudín (Museo de la ciudad), 2007. Anselm of Canterbury 1956 Anselm of Canterbury [Bec], ‘Why God Became Man’, in A Scholastic Miscellany: Anselm to Ockham, E. Fairweather (ed.), Philadelphia, 1956, pp. 100-183.

Baert 2003 Barbara Baert, ‘Schaamte en lichamelijkheid in de beeldvorming omtrent Maria Magdalena’, in Speling: Tijdschrift voor Bezinning, 55, 2003, pp. 46-60. Baert 2009 Barbara Baert, ‘Between Technique and Symbolism: Notes on the Meaning of the Use of Gold in Pre-Eyckian Panel Painting. A Contribution to the Comparative History of Art North and South of the Alps’, in Cyriel Stroo (ed.), Pre-Eyckian Panel Painting in the Low Countries, 2, Essays (Contributions to Fifteenth-Century Painting in the Southern Netherlands and the Principality of Liège, 9), Brussels, 2009, pp. 7-23. Baert 2010 Barbara Baert, ‘“Who Touched my Clothes?” The Healing of the Woman with the Haemorrhage (Mark 5: 24-34; Luke 8: 42-48 and Matthew 9: 19-22) in Early Medieval Visual Culture’, in Konsthistorisk Tidskrift - Journal of Art History, 79 (2), 2010, pp. 65-90. Baert, Fraeters 2005 Barbara Baert and Veerle Fraeters (eds), Het wellende water: De bron in tekst en beeld in de middeleeuwse Nederlanden en het Rijnland, Leuven, 2005. Bagley-Young 2007 Abbie Bagley-Young, ‘St.-Antonius behandeld: Een schilderij van Jan Mandijn onderzocht en gerestaureerd’, in Halszaken, 17, 2007, pp. 2-4. Baillie 1984 Michael G.L. Baillie, ‘Some Thoughts on the Art-historical Dendrochronology’, in Journal of Archaeological Science, 11, 1984, pp. 371-393.

Antwerp 2002 Fritz Koreny (ed.), Meestertekeningen van Jan van Eyck tot Hiëronymus Bosch, exh. cat., Antwerp (Rubenshuis), 2002.

Baillie, Hillam, Briffa 1985 Michael G.L. Baillie, Jennifer Hillam and Keith R. Briffa, ‘Re-Dating the English Art-historical Tree-ring Chronologies’, in Nature, 315, 1985, pp. 317-319.

Anvers 2002 Fritz Koreny (ed.), Dessins de Jan van Eyck à Hieronymus Bosch, exh. cat., Anvers (Rubenshuis), 2002.

Bain 1982 D.M. Bain, ‘Katwnakhn ton coiron apotetilmenax’, in Liverpool Classical Monthly, 7, 1982.

Aquinas 1920 Thomas Aquinas, The Summa Theologica of St. Thomas Aquinas, 5 vols, London, 1920.

Balbi 1962 Giovanna Balbi, ‘Uomini d’arme e cultura nel Quattrocento genovese: Biagio Assereto’, in Atti della Società ligure di storia patria, II, 1962, 1, pp. 99-147.

Aru, de Geradon 1952 Carlo Aru and Étienne de Geradon, La Galerie Sabauda de Turin (Corpus de la peinture des anciens Pays-Bas méridionaux au quinzième siècle, 5), Antwerp, 1952. Atkins Underwood 1950 Paul Atkins Underwood, The Fountain of Life in Manuscripts of the Gospels, Cambridge, MA, 1950. Augustine 1888 Augustine, Tractate 120 (John 19:31-20:9), 2. Nicene and Post-Nicene Fathers, First Series, vol. 7, Philip Schaff (ed.), Buffalo, 1888. Avril, Dunlop, Yapp 1989 François Avril, Louisa Dunlop and Brunsdon Yapp, Les Petites Heures de Jean de France, duc de Berry, Introduction/Kommentar, Lucerne, 1989. Baert 1997 Barbara Baert, ‘The Allegory with a Virgin. Contributions to the Solution of an Iconographical Enigma’, in Roger Van Schoute and Anne Dubois (eds), Memling Studies, Leuven, 1997, pp. 195-210.

Baltimore 1998 Roger S. Wieck, Time Sanctified. The Book of Hours in Medieval Art and Life, exh. cat., Baltimore (Walters Art Museum), 1998. Barcelona 1992 Xavier Barral i Altet (ed.), Prefiguración del Museu Nacional d’Art de Catalunya, exh. cat., Barcelona (Museo Nacional d’Art de Catalunya), 1992. Barcelona/Bilbao 2003 La pintura gòtica hispanoflamenca. Bartolomé Bermejo i la seva època, exh. cat., Barcelona (Museo Nacional d’Art de Catalunya)/Bilbao (Museo de Bellas Artes de Bilbao), 2003. Barney et al. 2006 Stephen A. Barney et al., The Etymologies of Isidore of Seville, New York, 2006. Barocchi 1978 Paola Barocchi (ed.), Scritti d’arte. Pittura e scultura, 3, Turin, 1978.

559 Bartl et al. 2005 Anna Bartl, Christoph Krekel, Manfred Lautenschlager and Doris Oltrogge, Der „Liber illuministarum” aus Kloster Tegernsee, Stuttgart, 2005, pp. 90-91.

Belonje, Greenhill 1962 Johan Belonje and Frank Allan Greenhill, ‘Some Brasses in Germany and the Low Countries (V)’, in Transactions of the Monumental Brass Society, vol. IX, part IX, 1962, pp. 499-501.

Battilana 1833 Natale Battilana, Genealogie delle famiglie nobili di Genova, 3 vols, Genoa, 1833.

Belting 1990 Hans Belting, The Image and Its Public in the Middle Ages: Form and Function in Early Paintings of the Passion, M. Bertusis and R. Meyer (trans.), New Rochelle, 1990.

Bauch 1961-1962 Kurt Bauch, ‘Bildnisse des Jan van Eyck’, in Jahresheft 1961/62 der Heidelberger Akademie der Wissenschaften, pp. 96-142. Bauch et al. 1978 Josef Bauch, Dieter Eckstein and G. Brauner, ‘Dendrochronologische Untersuchungen an Einchenholztafeln von Rubens-Gemälden’, in Jahrbuch der Berliner Museen, 20, 1978, pp. 209-221. Bauch, Eckstein 1970 Josef Bauch and Dieter Eckstein, ‘Dendrochronological Dating of Oak Panels of Dutch Seventeenth-Century Paintings’, in Studies in Conservation, 15, 1970, pp. 45-50. Bauch, Eckstein 1981 Josef Bauch and Dieter Eckstein, ‘Woodbiological Investigations on Panels of Rembrandt Paintings’, in Wood Science and Technology, 15, 1981, pp. 251-263. Bax 1956 Dirk Bax, Beschrijving en poging tot verklaring van het Tuin der Onkuisheiddrieluik van Jeroen Bosch: gevolgt door kritiek op Fraenger (Verhandelingen der Koninklijke Nederlandse Akademie van Wetenschappen Afdeling Letterkunde), Amsterdam, 1956. Baxandall 1963 Michael Baxandall, ‘A Dialogue on Art from the Court of Leonello d’Este’, in Journal of the Warburg and Courtauld Institutes, 26, 1963, pp. 304-326. Baxandall 1964 Michael Baxandall, ‘Bartholomaeus Facius on Painting: a FifteenthCentury Manuscript of the De Viris Illustribus’, in Journal of the Warburg and Courtauld Institutes, 27, 1964, pp. 90-107. Bé 1997 Kenneth Bé, ‘Geological Aspects of Jan van Eyck’s “Saint Francis Receiving the Stigmata”’, in Jane Watkins (ed.), Jan van Eyck: Two Paintings of Saint Francis Receiving the Stigmata, Philadelphia, 1997, pp. 88-95. Beenken 1933 Hermann Beenken, ‘The Ghent Van Eyck Re-Examined’, in The Burlington Magazine, 58, 1933, pp. 64-72. Beenken 1933-1934 Hermann Beenken, ‘Zur Entstehungsgeschichte des Genter Altars Hubert und Jan von Eyck’, in Wallraf-Richartz Jahrbuch, n.F. 2-3, 19331934, pp. 176-232.

Belting 1994 Hans Belting, Likeness and Presence: A History of the Image before the Era of Art, E. Jephcott (trans.), Chicago, 1994. Belting, Eichberger 1983 Hans Belting and Dagmar Eichberger, Jan van Eyck als Erzähler: Frühe Tafelbilder in Umkreis der New Yorker Doppeltafel, Worms, 1983. Belting, Kruse 1994 Hans Belting and Christiane Kruse, Die Erfindung des Gemäldes. Das erste Jahrhundert der Altniederländischen Malerei, Munich, 1994. Bénédictins de Bouveret, 1965 Bénédictins du Bouveret, Colophons de Manuscrits Occidentaux des Origines au XVIe siècle, Tome I, Colophons signés A-D (1-3561), in G.G. Meersseman, A. Hänggi and P. Ladner (eds), Spicilegii Friburgensis Subsidia, vol. 2, Fribourg, 1965. Benesch 1928 Otto Benesch, Die Zeichnungen der niederländischen Schulen des XV. und XVI. Jahrhunderts (Beschreibender Katalog der Handzeichnungen der Graphischen Sammlung Albertina, 2), Alfred Stix (ed.), Vienna, 1928. Benesovská, Hlobil 1999 Klára Benesovská, Ivo Hlobil et al., Peter Parler & St Vitus Cathedral, 1356-1399, Prague, 1999. Benito Doménech 2001 Fernando Benito Doménech, ‘Evocaciones flamencas en los Primitivos Valencianos’, in Fernando Benito Doménech and José Gómez Frechina (eds), La Clave Flamenca en los Primitivos Valencianos, exh. cat., Valencia (Museo de Bellas Artes), 2001, pp. 23-62. Berg Sobré 1989 Judith Berg Sobré, Behind the Altar Table. The Development of the Painted Retable in Spain, 1350-1500, Columbia, 1989. Bergmans 1907 Paul Bergmans, ‘Note sur la représentation du retable de l’Agneau Mystique des Van Eyck, en tableau vivant, à Gand, en 1458’, in Fédération archéologique et historique de Belgique. Annales du XXe congrès, Ghent, 1907, pp. 530-535. Bergmans 1988 Anna Bergmans, ‘Muurschilderingen met internationale allure in de parochiekerk Sint-Catharina van Duisburg’, in M&L Monuments, Landscapes & Archaeology in Flanders, 7.3, 1988, pp. 48-59.

Beenken 1936 Hermann Beenken, ‘Der Stand des Hubert van Eyck-Problems. Fragen um den Genter Altar’, in Oud Holland, 53, 1936, pp. 26-53.

Bergmans 1994 Anna Bergmans, ‘Tervuren, Sint-Catharinakerk te Duisburg’, in Marjan Buyle and Anna Bergmans, Middeleeuwse muurschilderingen in Vlaanderen (M&L Cahier, 2), Brussels, 1994, pp. 180-183.

Bellaert 1485 Jacop Bellaert (ed.), Engelsman, Bartholomeus. Van den proprieteyten der dinghen. 1485, Editie Werkgroep Middelnederlandse Artesliteratuur, Digitale Bibliotheek Nederlandse Letteren (DBNL), 2010.

Bergmans 1998 Anna Bergmans, Middeleeuwse muurschilderingen in de 19de eeuw. Studie en inventaris van middeleeuwse muurschilderingen in Belgische kerkgebouwen (KADOC Artes, 2), Louvain, 1998.

560 Bergmans, Buyle 2012 Anna Bergmans and Marjan Buyle, ‘International Style in Mechelen. The Wall Paintings from c. 1400 in the Tower of St-John’s Church’, in Antwerp Royal Museum Annual, 2011.

Billinge, Campbell 1995 Rachel Billinge and Lorne Campbell, ‘The Infra-red Reflectograms of Jan van Eyck’s Portrait of Giovanni(?) Arnolfini and his wife Giovanna Cenami(?)’, in National Gallery Technical Bulletin, 16, 1995, pp. 47-60.

Berlin 1883 Gemäldegalerie Berlin, Beschreibendes Verzeichnis der Gemälde, 2 vols, Berlin, 1883.

Billinge, Verougstraete, Van Schoute 2000 Rachel Billinge, Hélène Verougstraete and Roger Van Schoute, ‘The Saint Barbara’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 41-47.

Berman 1999 Judith C. Berman, ‘Bad Hair Days in the Paleolithic. Modern (Re)Constructions of the Cave Man’, in American Anthropologist, 101, 1999, pp. 288-304.

Bischoff 1951 Bernhard Bischoff, ‘Das griechische Element in der abendländischen Bildung des Mittelalters’, in Byzantinische Zeitschrift, 44, 1951, pp. 27-55.

Bermejo 1990 Elisa Bermejo, ‘Influencia de Van Eyck en la pintura española’, in Archivo Español de Arte, 252, 1990, pp. 555-569. Bermejo Martinez 1980 Elisa Bermejo Martinez, La Pintura de los Primitivos Flamencos en España, Madrid, 1980. Berschin 1988 Walter Berschin, Greek Letters and the Latin Middle Ages. From Jerome to Nicholas of Cusa, revised and expanded edn, Jerold C. Frakes (trans.), Washington, 1988. Beuting 2004 Micha Beuting, Holzkundliche und dendrochronologische Untersuchungen an Resonanzholz als Beitrag zur Organologie, Remagen-Oberwinter, 2004. Beuting 2011 Micha Beuting, ‘Dendro-Organology? The Dendrochronological Method Applied to Musical Instruments’, in Pascale Fraiture (ed.), Tree Rings, Art, Archaeology. Proceedings of the conference, Brussels, Royal Institute for Cultural Heritage, 10-12 February 2010 (Scientia Artis, 7), Brussels, 2011, pp. 273-283. Białostocki 1966 Jan Białostocki, Les musées de Pologne (Gdańsk, Kraków, Warszawa) (Corpus de la peinture des anciens Pays-Bas méridionaux au quinzième siècle, 9), Brussels, 1966. Biedenfeld 1837 Ferdinand Frhr. Von Biedenfeld, ‘Ursprung, Auftreten, Größe, Herrschaft, Verfall und jetzige Zustände sämmtlicher Mönchs- und Klosterfrauen-Orden’, in Orient und Occident, 2, Weimar, 1837. Billamboz 2011 André Billamboz, ‘Applying Dendro-Typology to Large Timber Series’, in Pascale Fraiture (ed.), Tree Rings, Art, Archaeology. Proceedings of the conference, Brussels, Royal Institute for Cultural Heritage, 10-12 February 2010 (Scientia Artis, 7), Brussels, 2011, pp. 177-188. Billinge 2000 Rachel Billinge, ‘Examining Jan van Eyck’s underdrawing’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 83-96. Billinge et al. 1997 Rachel Billinge, Lorne Campbell, Jill Dunkerton, Susan Foister, Jo Kirby, Jennie Pilc, Ashok Roy, Marika Spring and Raymond White, ‘Methods and Materials of Northern European Painting in the National Gallery, 1400-1550’, in National Gallery Technical Bulletin, 18, 1997, pp. 6-55.

Blacksberg 2005 Leslie Ann Blacksberg, ‘Between Salvation and Damnation. The Role of the Fountain in the Fountain of Life’, in Barbara Baert and Veerle Fraeters (eds), Het wellende water: de bron in tekst en beeld in de middeleeuwse Nederlanden en het Rijnland, Leuven, 2005, pp. 157-173. Blanc 1997 Odile Blanc, Parades et parures. L’invention du corps de mode à la fin du Moyen Age, Paris, 1997. Blanchard 1910 P. Blanchard, Les Heures de Savoie: Facsimiles of Fifty-Two Pages from the Hours Executed for Blanche of Burgundy, London, 1910. Blasco Vallés, Reche Ontillera 2007 Almudena Blasco Vales and Alberto Reche Ontillera, ‘Documentos referentes a las embajadas de 1426 y 1427 contenidos en la cancillería de los reyes de Aragón’, in Eduard Mira and An Delva (ed.), A la búsqueda del Toisón de Oro. La Europa de los Príncipes. La Europa de las ciudades, vol. 1, exh. cat., Almudín (Museo de la ciudad), 2007, pp. 91-96, 411-418. Blockmans 1987 Wim Blockmans, ‘Het wisselingsproces van de Gentse schepenen tijdens de 15de eeuw’, in Handelingen der Maatschappij voor geschiedenis en oudheidkunde te Gent, Nieuwe Reeks, 41, 1987, pp. 75-96. Blockmans 2010 Wim Blockmans, Metropolen aan de Noordzee. De geschiedenis van Nederland, 1100-1560, Amsterdam, 2010. Blockmans, Prevenier 1997 Wim Blockmans and Walter Prevenier, De Bourgondiërs. De Nederlanden op weg naar eenheid 1384-1530, Amsterdam/Louvain, 1997. Blommaert 1837 Philip Blommaert, ‘Beknopte geschiedenis der kamers van rhetorica te Gent’, in Belgisch museum, 1, 1837, pp. 417-444. Bluck 2000 Richard S. Bluck (ed.), Plato’s Phaedo: A Translation of Plato’s Phaedo with Introduction, Notes and Appendices, London, 2000. Bodenstedt 1944 Sister Mary Immaculate Bodenstedt, The Vita Christi of Ludolphus the Carthusian, Washington D.C., 1944. Boese 1973 Helmut Boese (ed.), Thomas Cantimpratentsis Liber de natura rerum: editio princeps secundum codices manuscriptos, Berlin/New York, 1973. Bol 2012 Marjolijn Bol, Oil and the Translucent: Varnishing and Glazing in Practice, Recipes and Historiography, 1100-1600, PhD diss., University of Utrecht, 2012.

561 Bol, Lehmann 2012 Marjolijn Bol and Anne-Sophie Lehmann, ‘Painting Skin and Water. Towards a Material Iconography of Translucent Motifs in Early Netherlandish Painting’, in Lorne Campbell, Jan van der Stock, Catherine Reynolds and Lieve Watteeuw (eds), Rogier van der Weyden in Context. Papers presented at the Seventeenth Symposium for the Study of Underdrawing and Technology in Painting held in Leuven, 22-24 October 2009, Leuven, 2012, pp. 214-225.

Eyck/Gossaert/Bruegel. Masterpieces from the Kunsthistorische Museum Vienna, exh. cat., Bruges (Groeningemuseum), 2011, pp. 23-37.

Bologna 1977 Ferdinando Bologna, Napoli e le rotte mediterranee della pittura da Alfonso il Magnanimo a Ferdinando il Cattolico, Naples, 1977.

Borchert (forthcoming) Till-Holger Borchert, Van Eyck: the Complete Works, Cologne/London/ Hongkong, forthcoming.

Bomford, Roy, Smith 1986 David Bomford, Ashok Roy and Alistair Smith, ‘The Techniques of Dieric Bouts: Two Paintings Contrasted’, in National Gallery Technical Bulletin, 10, 1986, pp. 39-57.

Borchert, Beyer 2002 Till-Holger Borchert and Andreas Beyer, The age of van Eyck: the Mediterranean World of Early Netherlandish painting, 1430-1530, London, 2002.

Bonde et al. 1997 Niels Bonde, Ian Tyers and Tomasz Ważny, ‘Where does the Timber come from? Dendrochronological Evidence of the Timber Trade in Northern Europe’, in A. Sinclair, E. Slater and J. Gowlett (eds), Archaeological Sciences 1995, Proceedings of a Conference on the Application of Scientific Methods to Archaeology (Oxford Books Monography Series, 64), Oxford, 1997, pp. 201-204.

Borradaile, Borradaile 1966 Violet Borradaile and Rosamund Borradaile (trans.), The Strasburg Manuscript: A Medieval Painters Handbook, London, 1966.

Boone 1990 Marc Boone, Gent en de Bourgondische hertogen ca. 1384 - ca. 1453. Een sociaal-politieke studie van een staatsvormingsproces (Verhandelingen van de Koninklijke Academie voor Wetenschappen, Letteren en Schone Kunsten van België, Klasse der Letteren, 52), no. 133, Brussels, 1990. Boone 2011 Marc Boone, ‘De Bourgondische Nederlanden. “Terres de promission” and/or “Creative Environments”?’ in Bernard Bousmanne and Thierry Delcourt (eds), Vlaamse miniaturen 1404-1482, Brussels/Paris, 2011, pp. 47-54. Borchert 1997 Till-Holger Borchert, ‘Schreiber am Liebeshof. Wer ist der mysteriöse “Tymotheos” von Jan van Eyck’, in Rainer Flöhl and Henning Ritter (eds), Wissenschaftsjahrbuch, Frankfurt/Main, 1997, pp. 558-562. Borchert 1998 Till-Holger Borchert, ‘The Discovery of Bruges Painting’, in Maximiliaan P.J. Martens (ed.), From Memling to Pourbus. Bruges and the Renaissance, Ghent, 1998, pp. 15-21. Borchert 2000 Till-Holger Borchert, L’Homme au chaperon bleu de Jan van Eyck, Metz, 2000. Borchert 2006 Till-Holger Borchert, ‘Innovation, Reconstruction, and Deconstruction: Early Netherlandish Diptychs in the Mirror of their Reception’, in John Oliver Hand and Ron Spronk (eds), Essays in context: Unfolding the Netherlandish Diptych, Cambridge/New Haven/London, 2006, pp.172-199. Borchert 2007 Till-Holger Borchert, Jan van Eyck. Renaissance Realist, Hong Kong/ Cologne/London/Los Angeles/Madrid/Paris/Tokyo, 2007. Borchert 2008 Till Holger-Borchert, Jan Van Eyck, Cologne, 2008. Borchert 2011 Till-Holger Borchert, ‘Flemish Portraiture’, in Sylvia Ferino, TillHolger Borchert and Manfred Sellink (eds), Imperial Treasures: Van

Borchert 2012 Till-Holger Borchert, ‘Form and Function of van Eyck’s Portraiture’, in Marc De Mey, Maximiliaan P.J. Martens and Cyriel Stroo (eds), Vision and Material: Interaction between Art and Science in Jan van Eyck’s time, Brussels, 2012, pp. 213-233.

Bosshard 1992 Emil Bosshard, ‘Revealing Van Eyck. The Examination of the ThyssenBornemisza Annunciation’, in Apollo, 136, 1992, pp. 4-11. Bousmanne, Van Hoorebeeck 2000 Bernard Bousmanne and Céline van Hoorebeeck (eds), La librairie des ducs de Bourgogne, Manuscrits conservés à la Bibliothèque royale de Belgique, vol. 1, Turnhout, 2000. Brachert 1964 Thomas Brachert, ‘Pressbrokat-Applikationen, ein Hilfsmittel für die Stilkritik, dargestellt an einen Werkstatt der Spätgotik’, in Schweizerisches Institut fur Kunstwissenschaft. Jahresbericht 1963, 1964, pp. 3747. Braekman 1986 Willy Louis Braekman (ed.), ‘Middelnederlandse verfrecepten voor miniaturen en ‘alderhande substancien’, in Scripta, Mediaeval and Renaissance Texts and Studies, 18, Brussels, 1986, pp. 46-47. Brand-Philip 1971 Lotte Brand-Philip, The Ghent Altarpiece and the Art of Jan Van Eyck, Princeton, 1971. Braunfels 1974 Sigrid Braunfels, ‘Georg’, in Werner Braunfels (ed.), Lexikon der christlichen Ikonographie, vol. 6, Rome, 1974, col. 366-390. Braunfels-Essche 1976 Sigrid Braunfels-Esche, Sankt Georg: Legende, Verehrung, Symbol, Munich, 1976. Brettle 1924 Sigmund Brettle, San Vincente Ferrer und sein literarischer Nachlass (Vorreformationsgeschichtliche Forschungen, 10), Münster, 1924. Briels 1980 Jan Briels, ‘Amator Pictoriae Artis. De Antwerpse kunstverzamelaar Peeter Stevens (1590-1668) en zijn Constkamer’, in Jaarboek van het Koninklijke Museum voor Schone Kunsten Antwerpen, 1980, pp. 137226. Bringemeier 1974 Martha Bringemeier, Priester- und Gelehrtenkleidung: Tunika, Sutane, Schaube, Talar. Ein Beitrag zu einer geistesgeschichtlichen Kostümforschung, Münster, 1974.

562 Brinkman 1993 Pim Brinkman, Het geheim van Van Eyck: aantekeningen bij de uitvinding van het olieverven, Zwolle, 1993. Brinkman et al., 1984-1985 Pim Willem, Frederik Brinkman, Léopold Kockaert, Luc Maes, Liliane Masschelein-Kleiner, Francis Robaszynski and Evert Thielen, ‘Het Lam Godsretabel van Van Eyck. Een heronderzoek naar de materialen en schildermethoden. 1. De plamuur, de isolatielaag, de tekening en de grondtonen’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 20, 1984-1985, pp. 137-166. Brinkman et al., 1988-1989 Pim Brinkman, Leopold Kockaert, Luc Maes, Evert Thielen and Jan Wouters,’ Het Lam Godsretabel van Van Eyck - een heronderzoek naar de materialen en schildermethoden,2. de hoofdkleuren blauw, groen, geel en rood’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 22, 1988-1989, pp. 26-49. Brinkman et al., 1990 Pim Willem, Frederik Brinkman, Léopold Kockaert, Luc Maes, Evert Thielen and Jan Wouters, ‘Het Lam Godsretabel van Van Eyck. Een heronderzoek naar de materialen en schildermethoden, 2. De hoofdkleuren. Blauw, groen, geel en rood’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van Koninklijk Instituut voor het Kunstpatrimonium, 22, 1988-1989 (1990), pp. 26-49. Brinkmann 1993 Pim Brinkman, Het geheim van Van Eyck, Zwolle, 1993. Bruges 1969 Primitifs flamands anonymes. Maîtres aux noms d’emprunt des Pays-Bas méridionaux du XVe et du début du XVIe siècle, exh. cat., Bruges (Groeningemuseum), 1969. Bruges 1976 600 jaar Brugs stadhuis 1376-1976, exh. cat., Bruges (Bank Brussel Lambert), 1976. Bruges 2002 Till-Holger Borchert (ed.), The Age of Van Eyck. The Mediterranean World and Early Netherlandish Painting 1430-1530, exh. cat., Bruges (Groeningemuseum), 2002. Bruges 2010 Till-Holger Borchert (ed.), Van Eyck to Dürer. Early Netherlandish Painting and Central Europe 1430-1530, exh. cat., Bruges (Groeningemuseum), 2010. Bruges 2011 Till-Holger Borchert et al., Van Eyck to Dürer: The Influence of Early Netherlandish Painting on European Art, 1430-1530, exh. cat., Bruges (Groeningemuseum), 2011. Bruges 2011 Sylvia Ferino, Till-Holger Borchert and Manfred Sellink (eds), Imperial Treasures: Van Eyck/Gossaert/Bruegel. Masterpieces from the Kunsthistorische Museum Vienna, exh. cat., Tielt, 2011.

Bruxelles 2007 Joost vander Auwera, Sabine van Sprang and Inga Rossi-Schrimpf (eds), Rubens. L’atelier du génie, exh. cat., Brussels (Musées royaux des BeauxArts de Belgique), 2007. Bruyn 2005 Josua Bruyn, Ernst van de Wetering et al., A Corpus of Rembrandt Paintings IV: Self-Portraits, The Hague, 2005. Buck 1995 Stephanie Buck, ‘Petrus Christus’s Berlin Wings and the Metropolitan Museum’s Eyckian Diptych’, in Maryan W. Ainsworth (ed.), Petrus Christus in Renaissance Bruges: An Interdisciplinary Approach, New York/ Turnhout, 1995, pp. 65-83. Buck 2000 Stefanie Buck, ‘An Approach to looking at Eyckian Drawings’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 183-195. Buettner 1992 B. Buettner, ‘Profane Illuminations, Secular Illusions: Manuscripts in Late Medieval Courtly Society’, in The Art Bulletin, 74, 1, 1992, pp. 75-90. Bull 2013 David Bull, ‘The Cleaning and Restoration of “The Annunciation” by Jan van Eyck’, in Facture, 1, 2013, pp. 156-166. Buren 1985 Anna H. van Buren, ‘Un jardin d’amour de Philippe le Bon au parc de Hesdin. Le role de Van Eyck dans une commande ducale’, in Revue du Louvre, 25, 1985, pp. 185-192. Burr 1984 David Burr, ‘Eucharistic Presence and Conversion in Late ThirteenthCentury Franciscan Thought’, in Transactions of the American Philosophical Society, 74, 3, 1984, pp. 1-113. Burroughs 1939 Alan Burroughs, Art Criticism from a Laboratory, London, 1939. Butler 1997 Marigene H. Butler, ‘An investigation of the Philadelphia “Saint Francis Receiving the Stigmata”’, in Jane Watkins (ed.), Jan van Eyck: Two Paintings of Saint Francis Receiving the Stigmata, Philadelphia, 1997, pp. 29-46. Butler, Van Asperen de Boer 1989 Marigen H. Butler and J. R. J. van Asperen de Boer, ‘The Examination of the Turin-Milan Hours with Infrared Reflectography: A Preliminary Report’, in Le dessin sous-jacent dans la peinture. Colloque VII, Géographie et chronologie du dessin sous-jacent, 1987, Louvain-la-Neuve, 1989, pp. 55-70. Buylaert 2011 Frederik Buylaert, Repertorium van de Vlaamse adel (ca. 1350 - ca. 1500) (Historische Monografieën Vlaanderen, 1), Ghent, 2011.

Brussels/Paris 2011 Bernard Bousmanne and Thierry Delcourt (eds), Miniatures flamandes 1404-1482, exh. cat., Brussels (Bibliothèque royale de Belgique)/Paris (Bibliothèque nationale de France), 2011.

Buylaert, De Clerq, Dumolyn 2011 Frederik Buylaert, Wim De Clerq and Jan Dumolyn, ‘Sumptuary Legislation, Material Culture and the Semiotics of “vivre noblement” in the County of Flanders (14th - 16th Centuries)’, in Social History, 36, 2011, pp. 393-417.

Bruxelles 1983 Dessins du XVe au XVIIIe siècle dans les collections privées de Belgique, exh. cat., Brussels (Générale de Banque), 1983.

Buylaert, Verroken 2013 Frederik Buylaert and Erik Verroken, ‘Een adellijk altaarstuk. Een sociaalhistorische studie van de ontstaanscontext van het Lam God-

563 sretabel’, in Danny Praet (ed.), Het Lam Gods. Kunst, Geschiedenis, Wetenschap en Religie, Brussels, 2013. Buyle 1995 Marjan Buyle, ‘Een puzzel op de schouw en draken op de muren. Merkwaardige muurschilderingen uit de late 14de eeuw in een Brugs woonhuis (A puzzle on the mantelpiece and dragons on the wall. Remarkable wall paintings from the late 14th century in a house in Bruges)’, in M&L Monuments, Landscapes & Archaeology in Flanders, 14.1, 1995, pp. 6-27. Buyle 2008 Marjan Buyle, ‘Een reus achter het orgel. Vondst van middeleeuwse muurschilderingen in de Mechelse Sint-Janskerk (A giant behind the organ. Discovery of medieval mural paintings in Saint John’s Church in Mechelen)’ in M&L Monuments, Landscapes & Archaeology in Flanders, 27.6, 2008, pp. 23-33. Buyle, Bergmans 1994 Marjan Buyle and Anna Bergmans, Middeleeuwse muurschilderingen in Vlaanderen (M&L Cahier, 2), Brussels, 1994. Buyle, Bergmans 2012 Marjan Buyle and Anna Bergmans, ‘Découverte, etude et restauration de peintures murales (ca. 1400) en l’église Saint-Jean à Malines’ in CeROArt [http://ceroart.revues.org/2822], August 2012. Buyle, Bergmans 2013 Marjan Buyle and Anna Bergmans, ‘’Internationale stijl’ in Mechelen. Onderzoek en conservatie van de muurschilderingen van rond 1400 in de toren van de Sint-Janskerk, (International Style in Mechelen. Research and conservation of the wall paintings dating from around 1400 in the tower of Saint John’s Church)’ in Relicta, 10, 2013. Calabrese 2006 Omar Calabrese, L’art de l’autoportrait. Histoire et théorie d’un genre pictural, Odile Ménégaux and Reto Morgenthaler (trans.), Paris, 2006. Calkins 1986 Robert Calkins, ‘Liturgical Sequence and Decorative Crescendo in the Drogo Sacramentary’, in Gesta, 25, 1, 1986, pp. 17-23. Camille 1998 Michael Camille, ‘Mimetic Identification and Passion Devotion in the Later Middle Ages: A Double-sided panel by Meister Francke’, in Alasdair MacDonald, Bernhard Ridderbos and Rita Schlusemann (eds), The Broken Body, Passion Devotion in Late-Medieval Culture, Groningen, 1998, pp. 183-210. Campbell 1974 Lorne Campbell, ‘Robert Campin, the Master of Flémalle and the Master of Mérode’, in The Burlington Magazine, 116, 1974, pp. 634646. Campbell 1985 Lorne Campbell, The Early Flemish Pictures in the Collection of Her Majesty The Queen, Cambridge, 1985. Campbell 1990 Lorne Campbell, Renaissance Portraits, European Portrait-Painting in the 14th, 15th and 16th Centuries, New Haven/London, 1990. Campbell 1995 Lorne Campbell, ‘Approaches to Petrus Christus’, in Maryan W. Ainsworth (ed.), Petrus Christus in Renaissance Bruges. An interdisciplinary approach, Turnhout, 1995, pp. 1-10.

Campbell 1998 Lorne Campbell, The Fifteenth Century Netherlandish Schools (National Gallery Catalogues), London, 1998. Campbell 2014 Lorne Campbell, The Sixteenth Century Netherlandish Paintings with French Paintings before 1600 (National Gallery Catalogues), London 2014. Campbell, Foister 1986 Lorne Campbell and Susan Foister, ‘Gerard, Lucas and Susanna Horenbout’, in The Burlington Magazine, 128, 1986, pp. 719-727. Campbell et al. 2012 Lorne Campbell, Jan van der Stock, Catherine Reynolds and Lieve Watteeuw (eds), Rogier van der Weyden in Context. Papers presented at the Seventeenth Symposium for the Study of Underdrawing and Technology in Painting held in Leuven, 22-24 October 2009, Paris/Louvain/Walpole, MA, 2012. Campbell et al. 2008 Lorne Campbell et al. (eds), Renaissance Faces. Van Eyck to Titian, exh. cat., London (National Gallery), 2008. Campbell et al. 1994 Lorne Campbell, David Bomford, Ashok Roy and Raymond White, ‘“The Virgin and Child before a Firescreen”: History, Examination and Treatment’, in National Gallery Technical Bulletin, 15, pp. 20-35. Campbell, Foister, Roy 1997 Lorne Campbell, Susan Foister and Ashok Roy (eds), compiled and written by Rachel Billinge, Lorne Campbell, Jill Dunkerton, Susan Foister, Josephine Kirby, Jennie Pilc, Ashok Roy, Marika Spring and Raymond White, ‘Early Northern European Painting’, in National Gallery Technical Bulletin, 18, 1997. Caporaletti 2011 Silvia Caporaletti, ‘Accenti fiamminghi in Santonio da Fabriano: riscontri documentari e ipotesi’, in Notizie da Palazzo Abani, 39, 2011, pp. 41-56. Cappelli 1990 Adriano Cappelli, Dizionario di abbreviature latine et italiane, 6th edn, Milan, 1990. Cardon 1998. Bert Cardon, ‘Ingekeerde portretten van Bouts’, in Ons Erfdeel, 41, 1998, pp. 515-524. Cardon 2002 Bert Cardon, ‘The Portfolio of a Bruges Miniaturist in the Mid15th Century’, in Bert Cardon, Jan van der Stock and Dominique van Wijnsberghe (eds), ‘Als Ich Can’. Liber Amicorum in Memory of Professor Dr Maurits Smeyers, Leuven, 2002, pp. 317-355. Cardon 2005 Bert Cardon, ‘De Paradijsbron. Een meander in de beeldende kunst’, in Barbara Baert and Veerle Fraeters (eds), Het wellende water: De bron in tekst en beeld in de middeleeuwse Nederlanden en het Rijnland, Leuven, 2005, pp. 145-155. Cariati et al. 2000 Franco Cariati, Laura Rampazzi, Lucia Toniolo and Andrea Pozzi, ‘Calcium oxalate films on stone surfaces: experimental assessment of the chemical formation’, in Studies in Conservation, 45, 2000, pp. 180188.

564 Carleton 1982 David L. Carleton, in The Art Bulletin, 64, no.1, 1982, pp. 118-124.

Châtelet 2011 Albert Châtelet, Hubert et Jan van Eyck, créateurs de l’Agneau Mystique, Dijon, 2011.

Cazelles, Rathofer 1988 Raymond Cazelles and Johannes Rathofer, Illuminations of Heaven and Earth. The Glories of the Très Riches Heures du Duc de Berry, New York, 1988.

Châtelet, Faggin 1969 Albert Châtelet and Giorgio T. Faggin, Tout l’œuvre peint des frères van Eyck, Paris, 1969.

Cennini (ed. Thompson 1960) C. Cennini, The Craftman’s Handbook “Il Libro dell’Arte” by Cennino d’Andrea Cennini, D. Thompson (ed.), New York, 1960.

Clark 1950 James M. Clark, The Dance of Death in the Middle Ages and the Renaissance, Glasgow, 1950.

Challéat 2007 Claire Challéat, ‘Intercambios diplomáticos y artísticos entre la corte de Borgoña y la corte de Nápoles relacionados con el proyecto de cruzada: el mapamundi Borgoñon de Alfonso el Magnánimo’, in Eduard Mira and An Delva (eds), A la búsqueda del Toisón de Oro. La Europa de los Príncipes. La Europa de las ciudades, vol. 1, exh. cat., Almudín (Museo de la ciudad), 2007, pp. 365-379.

Clark 1986 J. E. Clark, ‘Marot and Fabri’s Rhetoric of Letters’, in C.M. Grisé and C.D.E. Tolton (eds), Crossroads and Perspectives: French Literature of the Renaissance. Studies in honour of Victor E. Graham (Travaux d’Humanisme et de Renaissance, 211), Geneva, 1986.

Challéat 2012 Claire Challéat, Dalle Fiandre a Napoli. Committenza artistica, politica, diplomazioa al tempo di Alfonso il Magnanimo e Filippo il Buono, Rome, 2012. Châtelet 1967 Paul Durrieu, Heures de Turin. Quarante-cinq feuillets à peintures provenant de Très Belles Heures de Jean de France, duc de Berry, Paris, 1902, reissue with preface by Albert Châtelet, Turin, 1967. Châtelet 1980 Albert Châtelet, Les primitifs hollandais: la peinture dans les Pays-Bas du Nord au XVe siècle, Fribourg, 1980. Châtelet 1985 Albert Châtelet, ‘Art français et art flamand: les échanges dans le domaine pictural’, in La France de la fin du XVe siècle, Actes du colloque du CNRS, Paris, 1985, p. 232. Châtelet 1993 Albert Châtelet, Jean van Eyck enlumineur, Strasbourg, 1993. Châtelet 2000a Albert Châtelet, ‘Jan van Eyck entre l’Italie et la France’, in Journal des savants, 2000, pp. 73-98. Châtelet 2000b Albert Châtelet, L’âge d’or du manuscrit à peintures en France au temps de Charles VI et Les heures du Maréchal Boucicaut, Paris, 2000. Châtelet 2005-2006 Albert Châtelet, ‘Les miniatures de Jan van Eyck revisitées’, in Art de l’enluminure, 15, 2005-2006, pp. 36-66. Châtelet 2007 Albert Châtelet with the collaboration of Jean Paviot, Visages d’antan. Le receuil d’Arras (XIVe-XVIe s.), Lathuile, 2007. Châtelet 2010a Albert Châtelet, ‘Compte rendu de Pierre Colman, Jan van Eyck et Jean sans Pitié (Académie royale de Belgique. Mémoire de la Classe des Arts. Collection in-8°, 3e série, t. XXVII), Brussels, 2009, pp. 73-78, 130131’, in Art de l’enluminure, 34, 2010, pp. 67-68. Châtelet 2010b Albert Châtelet, ‘Early Dutch Painting: Thirty Years On’, in Oud Holland, 123, 2010, pp. 314-329.

Clark 1990 Kenneth Clark, The Nude. A Study in Ideal Form, Princeton, 1990. Clarke 2001 Mark Clarke, The Art of All Colours. Mediaeval Recipe Books for Painters and Illuminators, London, 2001. Clarke 2011 Mark Clarke, Mediaeval Painters’ Materials and Techniques. The Montpellier ‘Liber diversarum aricum’, London, 2011. Cnops 1986 L. Cnops, ‘De beveiliging van het Lam Gods’, in M&L Monuments, Landscapes & Archaeology in Flanders, 5.4, 1986, pp. 31-36. Collier, Carleton 1983 James M. Collier and David L. Carleton, in The Art Bulletin, 65, no. 4, 1983, pp. 691-692. Colman 2006 Pierre Colman, ‘Mise en scène satirique de préoccupations matrimoniales à la cour de Guillaume VI de Bavière-Hollade dans un dessin eyckien du Musée du Louvre’, in Oud Holland, 119, 2006, pp. 821. Colman 2009 Pierre Colman, Jan van Eyck et Jean sans Pitié (Académie royale de Belgique. Mémoire de la Classe des Arts. Collection in-8°, 3e série, t. XXVII), Brussels, 2009, pp. 73-78, 130-131. Colman 2011 Pierre Colman, in Kunstgeschichte, 2011 [http://www.kunstgeschichteejournal.net/140/1/Lebensbrunnen.pdf]. Colman 2012 Pierre Colman, ‘À la recherche d’un autoportrait de Jan van Eyck’, in Bulletin de la Classe des Arts de l’Académie royale des Sciences, des Lettres et des Beaux-Arts de Belgique (6e série), 23, 2012, pp. 81-96. Cologne 1978 Die Parler und der schöne Stil 1350-1400, exh. cat., Cologne, 1978. Cologne 1987 Joachim Plotzek (ed.), Andachtsbücher des Mittelalters aus Privatbesitz, exh. cat., Cologne, (Schnütgen-Museum), 1987. Colvil, Foister 1996 Howard Colvil and Susan Foister (eds), The Panorama of London circa 1544 by Anthonis van den Wijngaerde, London/Oxford, 1996.

565 Comblen-Sonkes 1970 Micheline Comblen-Sonkes, ‘Le dessin sous-jacent chez les Primitifs flamands’, in Bulletin de l’Institut royal du Patrimoine artistique, 12, 1970, pp. 195-223.

Coremans, Philippot, Sneyers 1951 Paul Coremans, Albert Philippot and René Sneyers, Van Eyck. L’Adoration de l’Agneau. Éléments nouveaux intéressant l’histoire de l’art, Brussels, 1951.

Comblen-Sonkes 1996 Micheline Comblen-Sonkes, The Collegiate Church of Saint Peter, Louvain (Corpus of Fifteenth-Century Painting in the Southern Netherlands and the Principality of Liège, 18), Brussels, 1996.

Corner 1936 George W. Corner, ‘A Thirteenth-Century Medical Cryptogram’, in Bulletin of the Institute of the History of Medicine, 4, No. 9, 1936, pp. 745750.

Comblen-Sonkes, Lorentz 1995 Micheline Comblen-Sonkes and Philippe Lorentz, Musée du Louvre, Paris, 2 (Corpus de la peinture des anciens Pays-Bas méridionaux au quinzième siècle, 17), Brussels, 1995.

Cornudella 2009-2010 Rafael Cornudella, ‘Alfonso el Magánimo y Jan van Eyck. Pintura y tapices flamencos en la corte del rey de Aragón’, in Locus Amoenus, 10, 2009-2010, pp. 39-62.

Conover et al. 2011 Damon M. Conover et al., ‘Towards Automatic Registration of Technical Images of Works of Art’, in Computer Vision and Image Analysis of Art II, Proceedings of SPIE, 7869, 2011, pp. 1-9.

Criminisi, Kemp, Kang 2004 Antonio Criminisi, Martin Kemp and Sing Bing Kang, ‘Reflections of Reality in Jan van Eyck and Robert Campin’, in Historical Methods, 37, no. 3, 2004, pp.109-121.

Coomans 1894 Mathias Coomans, ‘La distinction établie aujourd’hui entre artiste et artisans existait elle au moyen âge (discussion et suite)’ in Bulletijn van de Geschied- en Oudheidkundige kring van Gent, 1, 1894, p. 106.

Crowe, Cavalcaselle 1857 Joseph Archer Crowe and Giovanni Battista Cavalcaselle, The Early Flemish Painters, London, 1857.

Coomans 2009 Thomas Coomans, ‘Entre France et Empire: l’architecture dans le duché de Brabant au temps de Jeanne de Brabant et Wenceslas de Luxembourg (1355-1406)’, in Revue de l’Art, 166, 2009, pp. 9-25. Coremans 1950 Paul Coremans, ‘Cleaning and Restoration of Old Paintings: the Point of View of a Physical and Chemical Laboratory - Le Nettoyage et la restauration des peintures anciennes: le point de vue du laboratoire physico-chimique’, in Museum, III, 3, 1950, pp. 227-232. Coremans 1953 Paul Coremans (ed.), L’Agneau Mystique au laboratoire. Examen et traitement (Contributions à l’étude des Primitifs flamands, 2), Antwerp, 1953. Coremans 1954 Paul Coremans, ‘La technique des “Primitifs flamands”. Étude scientifique des matériaux, de la structure et de la technique picturale. III. Van Eyck: L’Adoration de l’Agneau Mystique (Gand: Cathédrale SaintBavon)’, in Studies in Conservation, 1(4), 1954, pp. 145-161. Coremans 1955 Paul Coremans, ‘Un colloque international sur l’Agneau Mystique’, in Bulletin des Musées royaux des Beaux-Arts, 1-3, Miscellanea Erwin Panofsky, March-June 1955, pp. 171-172. Coremans, Gettens, Thissen 1952 Paul Coremans, Rutherford John Gettens and Jean Thissen, Jr., ‘Étude scientifique des matériaux, de la structure et de la technique picturale. I. Introduction; II. Th. Bouts: Le Retable du Saint Sacrement (Louvain: Église Saint-Pierre)’, in Studies in Conservation, 1(1), 1952, pp. 1-29.

Cunningham 2006 Lawrence S. Cunningham, ‘Francis of Assisi as a Catholic Saint’, in Logos: A Journal of Catholic Thought and Culture, 9.1, 2006, pp. 56-71. Cuypers, Brinkman 1996 Dani Cuypers and Pim Brinkman, ‘Jan van Eyck en de uitvinding van de olieverf’, in Openbaar Kunstbezit, 3, 1996, pp. 1-64. Dalton 1912 Ormonde Maddeck Dalton, Catalogue of Finger Rings bequeathed by Sir Augustus Wollaston Franks, K.C.B., London, 1912. Daniels 1949 De spieghel der menscheliker behoudenisse. De Middelnederlandse vertaling van het Speculum Humanae Salvationis. Naar het handschrift uitgegeven, ingeleid en toegelicht, L.M. Daniels (ed.), Tielt, 1949. Davidson 2009 Arnold I. Davidson, Miracles of the Bodily Transformation, or How St. Francis Received the Stigmata, Chicago, 2009. Davies 1954 Martin Davies, The National Gallery, London, 2 (Corpus de la peinture des anciens Pays-Bas méridionaux au quinzième siècle, 3), Antwerp, 1954. De Antonio 1998 Trinidad De Antonio, ‘Coleccionismo, Devoción y Contrarreforma’, in Fernando Checa Cremades, Felipe II. Un Príncipe del Renacimento, exh. cat., Madrid (Museo del Prado), 1998, pp. 135-141. De Baets 1959 Jozef De Baets, ‘De handschriften van het “Lam Gods”’, in Wetenschappelijke Tijdingen (Orgaan van de Vereeniging van Wetenschap, Gent), 19, Ghent, 1959, cols 101-106.

Coremans, Thissen 1953 Paul Coremans and Jean Thissen, ‘Materiaux originaux et technique picturale eyckienne’, in L’Agneau Mystique au laboratoire. Examen et traitement (Contributions a l’étude des Primitifs flamands, 2), Antwerp, 1953, pp. 69-76.

De Baets 1961 Jozef De Baets, De gewijde teksten van het Lam Gods kritisch onderzocht (Verslagen en Mededelingen van de Koninklijke Vlaamse Academie voor Taal- en Letterkunde), Ghent, 1961.

Coremans, Janssens De Bisthoven 1948 Paul Coremans and Aquilin Janssens de Bisthoven, Van Eyck: The Adoration of the Mystic Lamb, Amsterdam, 1948.

De Barante 1835 Amable G. P. de Barante, Histoire des Ducs de Bourgogne de la maison de Valois, 2, Brussels, 1835.

566 De Clippel, Van Cauteren, Van der Stighelen 2011 Karolien De Clippel, Katharina van Cauteren and Katlijne van der Stighelen (eds), The Nude and the Norm in the Early Modern Low Countries, Turnhout, 2011. De Laborde 1849 De Laborde (comte), Les ducs de Bourgogne. Études sur les lettres, les arts et l’industrie pendant le XVe siècle et plus particulièrement dans les Pays-Bas et le duché de Bourgogne, 2, 1, Paris, 1849. De Maesschalck 1919-1920 P. G. de Maesschalck, ‘Josse Vydt, le donateur de l’Agneau Mystique et quelques membres de sa famille’, in Annalen van den Oudheidkundigen Kring van het Land van Waas, 33, 1919-1920, pp. 13-36. De Melo, Sanyova, Cruz 2011 Helena Pinheiro de Melo, Jana Sanyova and António João Cruz, ‘An Unusual Glazing Technique on a Portuguese Panel Painting from the Second half of the 16th Century: Materials, Technique and Reconstructions’, in Janet Bridgland (ed.), Preprints of the ICOM-CC 16th Triennial Conference, Lisbon, 2011, ATSR group, paper 0105. De Mey 2008 Marc De Mey, ‘Jan van Eyck and the Representation of Glow’, in Maria Clelia Galassi and Anna De Florani (eds), Nord/Sud, ricezioni fiamminghe al di qua delle Alpi, Biblioteca d’arte, Cinisello Balsano, 21, 2008, pp. 19-29.

De Schryver, Marijnissen 1953 Antoine De Schrijver and Roger Marijnissen, ‘Histoire matérielle’, in Paul Coremans (ed.), L’Agneau Mystique au laboratoire. Examen et traitement (Contributions à l’étude des Primitifs flamands, 2), Antwerp, 1953, pp. 21-68. De Simoni, Belgrano 1867 Cornelio De Simoni and Luigi T. Belgrano, ‘Documenti ed estratti inediti o poco noti riguardanti la storia del commercio e della marineria ligure, 1, Brabante, Fiandra e Borgogna’, in Atti della Società Ligure di Storia Patria, V, 1, 1867, pp. 359-547. De Smet 1856 Joseph-Jean De Smet, Recueil des Chroniques de Flandre, 3, Brussels, 1856. De Thöt 1934 Paolo De Töth, Il beato cardinal Nicolò Albergati e I suoi tempi 13751444, Viterbo, 1934. De Tolnay 1938 Charles de Tolnay, Le Retable de l’Agneau Mystique des Frères van Eyck, Brussels, 1938. De Tolnay 1939 Charles de Tolnay, Le Maître de Flémalle et les frères Van Eyck, Brussels, 1939.

De Mey, Martens, Stroo 2012 Marc De Mey, Maximilian P.J. Martens and Cyriel Stroo (eds), Vision & Material. Interaction between Art and Science in Jan van Eyck’s Time, Brussels, 2012.

De Vos 1983 Dirk De Vos, ‘Nogmaals ALS ICH CAN’, in Oud Holland, 97, 1983, pp. 1-4.

Denaux, Van den Berghe 1984 Adelbert Denaux and Eric Van den Berghe, De Duinenabdij en het Grootseminarie te Brugge, Tielt, 1984.

De Vos 2011 Dirk De Vos, ‘Een mysterieuze persoonsverwisseling: Jan van Touraine en niet Filips de Stoute’, in Revue belge d’archéologie et d’histoire de l’art, 80, 2011, pp. 197-200.

Denis 1964 Valentine Denis, The Adoration of the Mystic Lamb, M. Langley (trans.), Milan, 1964. De Patoul, Van Schoute 1998 Brigitte de Patoul and Roger Van Schoute (eds), Les Primitifs flamands et leur temps, Brussels, 1998. De Pauw 1896 Napoleon De Pauw, ‘Enquête instituée relativement à l’authenticité de la pierre tombale présumée d’Hubert van Eyck’, in Bulletijn van de Geschied- en Oudheidkundige Kring van Gent, 3, 1896, pp. 85-103. De Poorter 1931 Alphonse De Poorter, ‘Testament van Anselmus Adornes, 10 Febr. 1470’, in Biekorf, 37, 1931, pp. 225-239. De Poorter 1934 Alphonse De Poorter, Catalogue des Manuscrits de la Bibliothèque Publique de la Ville de Bruges, Catalogue Général des Manuscrits des Bibliothêques de Belgique, vol. 2, Gembloux/Paris, 1934. De Ridder 1989-1991 Juliaan H. A. de Ridder, ‘Vlaamse Primitief op de dool: “Het Laatste Oordeel” van Lieven van den Clite’, in Bulletin van het Koninklijke Musea voor Schone Kunsten van België/Bulletin des Musées royaux des Beaux-Arts de Belgique, Brussels, 38-40, 1989-1991/1-3, pp. 91-121. De Schryver, Marijnissen 1952 Antoine De Schrijver and Roger Marijnissen, De oorspronkelijke plaats van het Lamgodsretabel. Nieuwe gegevens omtrent het Van Eyck-probleem (Contributions à l’étude des Primitifs Flamands, 1), Antwerp, 1952.

De Vries 1844-1848 Matthias de Vries (ed.), Jan van Boendale, Der leken spieghel. Leerdicht van den jare 1330, Leiden, 1844-1848. De Winter 1985 Patrick de Winter, La Bibliothèque de Philippe le Hardi, duc de Bourgogne (1364-1404), Paris, 1985. Del Federico et al. 2006 Eleonora Del Federico, Wolfgang Shöfberger, Johannes Schelvisj, Sofia Kapetanaki, Lindsey Tyne and Alexej Jerschow, ‘Insight into Framework Destruction in Ultramarine Pigments’, in Inorganic Chemistry, 45, 2006, pp. 1270-1276. Delaissé 1972 Léon M.J. Delaissé, ‘The Miniatures added in the Low Countries to the Turin-Milan Hours and their Political Significance’, in Artur Rosenauer and Gerold Weber (eds), Kunsthistorische Forschungen Otto Pächt zu seinem 70. Geburtstag, Vienna, 1972, pp. 135-149. Delisle 1884 Léopold Delisle, ‘Les livres d’heures du duc de Berry II’, in Gazette des Beaux-arts, 2e période, 29, 1884, pp. 281-292. Delisle 1907 Léopold Delisle, Recherches sur la librairie de Charles V, 2 vols, Paris, 1907. Delumeau 1995 Jean Delumeau, History of Paradise: the Garden of Eden in Myth and Tradition, New York, 1995.

567 Demus, Klauner, Schütz 1981 Klaus Demus, Friderike Klauner and Karl Schütz, Kataloge der Gemälde Galerie. Flämische Malerei von Jan van Eyck bis Pieter Bruegel d.Ä, Vienna, 1981. Deneffe 2011 Dominque Deneffe, La miniature eyckienne, in Bernard Bousmanne and Thierry Delcourt (eds), Miniatures flamandes 1404-1482, exh. cat., Brussels/Paris, 2011, pp. 166-173. Deneffe 2011 Dominique Deneffe, De pre-Eyckiaanse miniatuurkunst, in Bernard Bousmanne and Thierry Delcourt (eds), Vlaamse miniaturen 1404-1482, exh. cat., Brussels/Paris, 2011, pp. 128-131. Depuydt-Elbaum 2002 Livia Depuydt-Elbaum, ‘The Madonna at the Fountain by Jan van Eyck’, in Restoration, 2, no. 1, 2002. Depuydt-Elbaum 2003 Lydia Depuydt-Elbaum, ‘La technique picturale de la Vierge à la Fontaine de Jan van Eyck - étude sous binoculaire stéréoscopique/De schildertechniek van de Madonna bij de Fontein van Jan van Eyck - onderzoek onder de binoculaire stereomicroscoop’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 30, 2003, pp. 165-172. Depuydt-Elbaum et al. 2010 Livia Depuydt-Elbaum, Pascale Fraiture and Françoise Rosier, ‘Technologische studie en restauratie van twee panelen van het Mirafloresretabel: “De prediking van Johannes de Doper” en “Het banket van Herodes”‘, in Juan de Flandes en het Mirafloresretabel – Gesignaleerd en opgespoord, exh. cat., Antwerp (Museum Mayer van den Bergh), 2010, pp. 56-83. Dequeker 1984 Luc Dequeker, ‘Jewish Symbolism in the Ghent Altarpiece of Jan van Eyck’, in Dutch Jewish History, 1984, pp. 347-362. Dequeker 1986 Luc Dequeker, ‘Het Lam-Godsretabel van Van Eyck en het Jodenvraagstuk’, in Ter Herkenning. Tijdschrift voor christenen en Joden, 14, 1986, pp. 98-110. Dequeker 1992 Luc Dequeker, ‘Joodse exegese op het Lam-Godsretabel in de SintBaafskathedraal te Gent’, in Cees Verdegaal and Wim Weren (eds), Stromen uit Eden. Genesis 1-11 in Bijbel, Joodse exegese en moderne literatuur, Boxtel/Bruges, 1992, pp.195-211. Dequeker 2011 Luc Dequeker, Het Mysterie van het ‘Lam Gods’. Filips de Goede en de Rechtvaardige Rechters van Van Eyck (Annua Nuntia Lovaniensia, 63), Leuven/Paris/Walpole, 2011. Derbes, Neff 2004 Anne Derbes and Amy Neff, ‘Italy, the Mendicant Orders, and the Byzantine Sphere’, in Helen Evans (ed.), Byzantium. Faith and Power (1261 - 1557), exh. cat., New York (The Metropolitan Museum of Art), 2004, pp. 448-461. Derolez 1966 Albert Derolez, Corpus Catalogorum Belgii. De Middeleeuwse Bibliotheekscatalogi der Zuidelijke Nederlanden. I. Province West-Vlaanderen, Brussels, 1966. Derolez 1984 Albert Derolez, ‘La tranche dorée et ciselée d’après les collections de la Bibliothèque de l’Université de Gand’, in G. Colin (ed.), De libris

compactis miscellanea (Studia Bibliothecae Wittockianae, 1), Brussels, 1984, pp. 251-272. Desmet 1912 Joseph Desmet, ‘Quelques mots à propos d’Hubert van Eyck’, in Bulletijn van de Geschied- en Oudheidkundige kring van Gent, 1912, pp. 502-527. Desmet 1935 Joseph Desmet, ‘La question Hubert van Eyck’, in Bulletin de la Société Royale d’Archéologie de Bruxelles, 23, 1935, pp. 199-201. Desneux 1958 Jules Desneux, ‘Underdrawings and Pentimenti in the Pictures of Jan van Eyck’, in The Art Bulletin, 40, 1958, pp. 13-21. Despodt 2000-2001 Veronique Despodt, ‘Dat du best, was ic ende wat ic bem, dat sal tu werden ...’, Gentse grafmonumenten en grafschriften tot het einde van de Calvinistische Republiek (1584), unpublished thesis, Ghent University, 2000-2001. Devillers 1892 Léopold Devillers, Cartulaire des comtes de Hainaut, de l’avènement de Guillaume II à la mort de Jacqueline de Bavière, 5, Brussels, 1892. Devliegher 1979 Luc Devliegher, De Sint-Salvatorskathedraal te Brugge, Inventaris (Kunstpatrimonium van West-Vlaanderen, 8), Tielt, 1979. Devolder 2009 Bart Devolder, ‘The Representation of Brocaded Silks and Velvets in 15th and early 16th Century Netherlandish Paintings: Methods and Materials’, in AIC Paintings Specialty Group Postprints, 21, Denver, 2009, pp. 61-74. Devliegher 1965 Luc Devliegher, Beeld van het Kunstbezit, Kunstpatrimonium van WestVlaanderen, I, Tielt/The Hague, 1965. Dhanens 1965 Elisabeth Dhanens, Het Retabel van het Lam Gods in de Sint-Baafskathedraal te Gent (Inventaris van het kunstpatrimonium van Oostvlaanderen, 6), Ghent, 1965. Dhanens 1969-1972 Elisabeth Dhanens, ‘De wijze waarop het Lam Godsaltaar was opgesteld’, in Gentse bijdragen tot de kunstgeschiedenis en oudheidkunde, 22, 1969-1972, pp. 109-150. Dhanens 1971 Elisabeth Dhanens, Van Eyck: The Ghent Altarpiece (Art in Context), New York, 1971. Dhanens 1973 Elisabeth Dhanens, Van Eyck: the Ghent Altarpiece (Art in Context), London, 1973. Dhanens 1976 Elisabeth Dhanens, De Vijd-Borluut fundatie en het Lam Godsretabel: 1432-1797 (Mededelingen van de Koninklijke Academie voor Wetenschappen, Letteren en Schone Kunsten. Klasse der Schone Kunsten, 38), 1976, no. 2. Dhanens 1977 Elisabeth Dhanens, De kwartierstaat en het graf van Jan van Eyck (Medelingen van de Koninklijke Académie voor Wetenschappen, Letteren en Schone Kunsten van Belgïe. Klasse der Schone Kunsten, 39), 1977, pp. 7-50.

568 Dhanens 1980 Elisabeth Dhanens, Hubert and Jan van Eyck, Antwerp, 1980. Dhanens 1980 Elisabeth Dhanens, Hubert et Jean van Eyck, Jacques Gengoux (trans.), Antwerp, 1980. Dhanens 1987 Elisabeth Dhanens, De Blijde inkomst van Filips de Goede in 1458 en de plastische kunsten te Gent (Academiae Analecta. Mededelingen van de Koninklijke Academie voor Wetenschappen, Letteren en Schone Kunsten van België, Klasse der Schone Kunsten, 48), 1987, no. 2, pp. 55-89. Dhanis 1857 Michiel C. H. Dhanis, Opkomst en bloei van het christendom in Mechelen, 1, Mechelen, 1857. Di Fabio 2011 Clario Di Fabio, ‘Nascita e rinascita della statuaria celebrativa laica a Genova fra Tre e Quattrocento. Opizzino, Giacomo Spinola di Luccoli e la parte di Domenico Gagini’, in Arturo Carlo Quintavale (ed.), Medioevo: i committenti, Atti del Convegno Internazionale di studi, Parma, 21-26 settembre 2010, Milan, 2011, pp. 623-641. Díaz Padrón 2009 Matías Díaz Padrón, ‘Pourbus y la tradición primitivos flamencos’, in Ars magazine: revista de arte y coleccionismo, 2009, pp. 52-65. Dickinson 1955 Joycelyne G. Dickinson, The Congres of Arras 1435. A Study in Medieval Diplomacy, Oxford, 1955. Didi-Huberman 2006 Georges Didi-Huberman, ‘Opening up Venus: Nudity, Cruelty and the Dream’, in Thomas Frangenberg and Robert Williams (eds), The Beholder. The Experience of Art in Early Modern Europe, Aldershot, 2006, pp. 37-52. Dierick 2000 Alphonse L. Dierick, ‘Jan van Eyck’s Handwriting’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 79-82. Dijkstra 1990 Jeltje Dijkstra, Origineel en kopie. Een onderzoek naar de navolging van de Meester van Flémalle en Rogier van der Weyden, Amsterdam, 1990. Dijkstra 1996 Jeltje Dijkstra, ‘The Brussels and the Merode Annunciation Reconsidered’, in Susan Foister and Susie Nash (eds), Robert Campin. New Directions in Scholarship, Turnhout, 1996, pp. 95-104. Dijon 1992 D’ocre et d’azur. Peintures murales en Bourgogne, exh. cat., Dijon (Musée des Beaux-Arts), 1992. Dijon 2004 L’art à la cour de Bourgogne. Le mécénat de Philippe le Hardi et de Jean sans peur (1364-1419), exh. cat., Dijon (Musée des Beaux-Arts)/Cleveland (The Cleveland Museum of Art), 2004. Dionysius Carthusianus Dionysius Carthusianus, Opera omnia, Monstrolii/Tornaci, 1896-1912. Dodgson 1926 Campbell Dodgson, Albrecht Dürer, London/Boston, 1926.

Doehaerd, Kerremans 1952 Renée Doehaert and Charles Kerremans, Les relations commerciales entre Gê nes, la Belgique et l’Outremont d’après les archives notariales génoises. 1400-1440, Brussels/Rome, 1952. Doehlemann 1905 Karl Doehlemann, in Zeitschrift für Mathematik und Physik, 52, 1905, pp. 419-425. Doerner 1921 (1935) Max Doerner, The Materials of the Artist and their Use in Painting with Notes on the Techniques of the Old Masters, Eugen Neuhaus (trans.), London, 1935. Dogaer 1987 Georges Dogaer, Flemish Miniature Painting in the 15th and 16th Centuries, Amsterdam, 1987. Dresden 2005 Thomas Ketelsen and Uta Neidhardt (eds), Das Geheimnis des Jan van Eyck. Die frühen niederländischen Zeichnungen und Gemälde in Dresden, exh. cat., Dresden (Gemäldegalerie Alte Meister, Kupferstich-Kabinet Staatliche Kunstsammlungen), Munich/Berlin, 2005. Dubois, Van Grevenstein 2011 Hélène Dubois and Anne Van Grevenstein, ‘Bridging Controversy: the Contribution of an International Commission of Experts to the 195051 Study and Treatment of Van Eyck’s Ghent Altarpiece’, in Proceedings of the ICOM-CC triennial conference, Lisbon, 18-23 September 2011 (Publication CD-Rom). Duclos 1910 Adolf Duclos, Bruges, Histoire et souvenirs, Bruges, 1910. Duffy 1992 Eamon Duffy, The Stripping of the Altars. Traditional Religion in England c. 1400-c. 1580, New Haven/London, 1992. Duits 2008 Rembrandt Duits, Gold Brocade and Renaissance Painting. A Study in Material Culture, London, 2008. Dülberg 1990 Angelika Dülberg, Privatporträts. Geschichte und Ikonologie einer Gattung im 15. und 16. Jahrhundert, Berlin, 1990. Dunkerton 2000 Jill Dunkerton, ‘Observations on the Handling Properties of Binding Media Identified in European Painting from the Fifteenth to the Seventeenth Centuries’, in Bulletin de l’Institut royal du Patrimoine artistique/ Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium 1996-98, 27, 2000, pp. 287- 292. Dunkerton 2008a Jill Dunkerton, ‘The restoration of Margaret, the Artist’s Wife’ [http:// www.nationalgallery.org.uk/paintings/research/the-restoration-ofmargaret-the-artists-wife/introduction]. Dunkerton 2008b Jill Dunkerton, ‘The Technique and Restoration of “The Virgin and Child Enthroned, with Four Angels” by Quinten Massys’, in National Gallery Technical Bulletin, 29, 2008, pp. 60-75. Dunkerton, Billinge 2005 Jill Dunkerton and Rachel Billinge, Beyond the Naked Eye. Details from the National Gallery, London, 2005.

569 Dunkerton, Spring 2013 Jill Dunkerton and Marika Spring, with contributions from Rachel Billinge, Kamilla Kalinina, Rachel Morrison, Gabriella Macaro, David Peggie and Ashok Roy, ‘Titian’s Painting Technique to c.1540’, in National Gallery Technical Bulletin, 34, 2013.

Eeckhout 2001 Paul Eeckhout, ‘Le Maître de 1499’, in Brigitte de Patoul and Roger Van Schoute (eds), Les Primitifs flamands et leur temps, Louvain-la-Neuve, 2001, pp. 521-522.

Durand de Mende 1854 Guillaume Durand de Mende (ed.) Charles Barthélemy, Rational ou manuel des divins offices, Paris, 1854.

Effman 2006 Elise Effman, ‘Theories about the Eyckian Painting Medium from the Late-Eighteenth to the Mid-Twentieth Centuries’, in Reviews in Conservation, 7, 2006, pp. 17-26.

Durrieu 1902 Paul Durrieu, Heures de Turin. Quarante-cinq feuillets à peintures provenant de Très Belles Heures de Jean de France, duc de Berry, Paris, 1902.

Eichberger 1987 Dagmar Eichberger, Bildkonzeption und Weltdeutung im New Yorker Diptychin des Jan van Eyck, Wiesbaden, 1987.

Durrieu 1903 Paul Durrieu, ‘Les débuts des Van Eyck’, in Gazette des Beaux-Arts, 3e période, XXIX, pp. 5-18, 107-120.

Elkins 1991 James Elkins, in The Art Bulletin, 73, no.1, 1991, pp. 53-62.

Durrieu 1910 Paul Durrieu, ‘Les “Très belles Heures de Notre-Dame” du duc Jean de Berry’, in Revue archéologique, 4e série, 16, 1910, pp. 30-51, 246-279. Durrieu 1911a Paul Durrieu, ‘Les aventures de deux splendides livres d’heures ayant appartenu au duc Jean de Berry’, in Revue de l’art ancien et moderne, 30, 1911, pp. 5-16, 91-103. Durrieu 1911b Paul Durrieu, ‘Notice d’un des plus importants livres de prières du roi Charles V: les Heures de Savoie ou “Très belles grandes heures” du roi’, in Bibliothèque de l’École des chartes, 72, 1911, pp. 550-555. Durrieu 1922 Paul Durrieu, Les ‘Très belles Heures de Notre-Dame’ du duc Jean de Berry, Paris 1922. Duverger 1945 Joseph Duverger, Het grafschrift van Hubert van Eyck en het quatrain van het Gentsche Lam Gods-retabel, met aanhangsel door E. Bontinck, Natuurwetenschappelijk onderzoek van de opschriften en de lijst van het Lam Gods-retabel (Verhandelingen van de Koninklijke Vlaamsche Academie voor Wetenschappen, Letteren en Schone Kunsten van België, Klasse der Schone Kunsten, 7), no. 4, 1945. Duverger 1955 Jozef Duverger, ‘Brugse schilders ten tijde van Jan van Eyck’, in Bulletin van de Koninklijke Musea voor Schone Kunsten, 4, 1955, pp. 83-120. Eastlake (1847) 1960 Sir Charles Lock Eastlake, Methods and Materials of Painting of the Great Schools and Masters, first published 1847 under the title Materials for a History of Oil Painting, republished London, 1960. Eckstein et al. 1975 Dieter Eckstein, Johannes A. Brongers and Josef Bauch, ‘Tree-ring Research in the Netherlands’, in Tree-ring Bulletin, 35, 1975, pp. 1-13. Eckstein et al. 1986 Dieter Eckstein et al., ‘New Evidence for the Dendrochronological Dating of Netherlandish Paintings’, in Nature, 320, 1986, pp. 465-466. Eckstein, Wrobel 2007 Dieter Eckstein and Sigrid Wrobel, ‘Dendrochronological Proof of Origin of historic Timber – Retrospect and Perspectives’, in Kristof Haneca et al. (eds), Proceedings of the Symposium on Tree Rings in Archaeology, Climatology and Ecology (TRACE), Jülich, 20-22 April 2006, (Schriften des Forschungszentrums Jülich, Reihe Umwelt/Environment, 74), Tervuren, 2007, pp. 8-20.

Elliger 1970 Karl Elliger, Biblischer Kommentar, Altes Testament, Jesaya II (40, 1-26), Neukirchen, 1970. Elsig 2010 Frédéric Elsig, ‘Compte rendu de Pierre Colman, Jan van Eyck et Jean sans Pitié (Académie royale de Belgique. Mémoire de la Classe des Arts. Collection in-8°, 3e série, t. XXVII), Brussels, 2009, pp. 73-78, 130-131’, in Revue de l’Art, 169, 2010, pp. 76-77. Emile-Mâle 1994 Gilberte Emile-Mâle, Inventaires et restauration au Louvre de tableaux conquis en Belgique, septembre 1794-février 1795, Mémoire de la Classe des Beaux-Arts (Académie royale des sciences, des lettres et des BeauxArts de Belgique. Classe des Beaux-Arts), 3e sér., t. 8., Brussels, 1994. Endres 2004 Johannes Endres, ‘Diderot, Hogarth, and the Aesthetics of Depilation’, in Eighteenth-Century Studies, 38/1, 2004, pp. 17-38. Enklaar 1950 Diederik Enklaar, De dodendans. Een cultuur-historische studie, Amsterdam, 1950. Erich 2006 S.J.F. Erich, NMR imaging of curing processes in alkyd coatings, PhD diss., Eindhoven University of Technology, 2006. Esther 1983 Jean-Pierre Esther, ‘Monumentenbeschrijving en bouwgeschiedenis van de Jeruzalemkapel’ in Adornes en Jeruzalem. International leven in het 15de- en 16de-eeuwse Brugge, exh. cat., Bruges (Jerusalem Chapel), 1983, pp. 51-81. Evans 2004 Helen Evans (ed.), Byzantium: Faith and Power (1261 - 1557), exh. cat., New York (The Metropolitan Museum of Art), 2004. Faggin 1976 Giorgio T. Faggin, L’Opera complete dei van Eyck, Milan, 1968. Faries 1999 Molly Faries, ‘The Underdrawing in Jan van Eyck’s Dresden Triptych’, in Hélène Verougstraete et al. (eds), Le dessin sous-jacent dans la peinture, Colloque XII, 1997, Louvain-La-Neuve, 1999, pp. 221-230. Faries 2001 Molly Faries, ‘Reshaping the Field: The Contribution of Technical Studies’, in Maryan W. Ainsworth (ed.), Early Netherlandish Painting at

570 the Crossroads. A Critical Look at Current Methodologies, Proceedings of a Conference in New York in 1998, New York, 2001, pp. 70-105. Faries 2003 Molly Faries, ‘Technical Studies of Early Netherlandish Painting: A Critical Overview of Recent Developments’, in Molly Faries and Ron Spronk (eds), Recent Developments in the Technical Examination of Early Netherlandish Painting: Methodology, Limitations & Perspectives, Proceedings of a conference in Harvard in 1996, Cambridge/Turnhout, 2003, pp. 1-38. Fensham 1959 F. Charles Fensham, ‘Thunder-Stones in Ugaritic’, in Journal of Near Eastern Studies, 18, 4, Oct. 1959, pp. 273-274. Ferrari, Salmi, Mellini 1972 Giorgio Ferrari, Mario Salmi and Gian Lorenzo Mellini, The Grimani Breviary, Reproduced from the Illuminated Manuscript belonging to the Biblioteca Marciana, Venice, London, 1972. Ferre i Puerto 1999 Josep Ferre i Puerto, ‘Trajectòria vital de Joan Reixac, pintor valencià del quatre-cents: La seua relació amb Andreu Garcia’, in Joaquín Yarza Luaces and Francesc Fité i Llevot (eds), L’Artista-artesà medieval a la Corona d’Aragó, Lleida, 1999, pp. 419-426. Fierens-Gevaert 1905 Hippolyte Fierens-Gevaert, La Renaissance septentrionale et les premiers Maîtres des Flandres, Brussels, 1905. Filedt Kok 2006 Jan Piet Filedt Kok, De heilige Lucas tekent en schildert de Madonna, Amsterdam, 2006. Filedt Kok et al. 2011 Jan Piet Filedt Kok, Esther van Duijn, Abbie Vandivere, Arie Wallert and Margreet Wolters, ‘De Leide schilders aan het werk’, in Christiaan Vogelaar, Jan Piet Filedt Kok, Huigen Leeflang and Ilja M. Veldman (eds), Lucas van Leyden en de Renaissance, exh. cat., Leiden (Museum De Lakenhal), 2011, pp. 79-101. Fingerlin 1971 Ilse Fingerlin, Gürtel des hohen und späten Mittelalters (Kunstwissenschaftliche Studien, XLVI), Berlin, 1971. Finot 1900 Jules Finot, Etudes historiques sur les relations commerciales entre la Flandre et la République de Gênes au Moyen Age, Paris, 1900. Fletcher 1976 John M. Fletcher, ‘A Group of English Royal Portraits painted soon after 1513, a Dendrochronological Study’, in Studies in Conservation, 21, 1976, pp. 171-178. Fletcher 1982 John M. Fletcher, ‘Panel Examination and Dendrochronology’, in The J. Getty Museum Journal, 10, 1982, pp. 39-44. Foister, Jones, Cool 2000 Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000. Folie 1996-1998 Jacqueline Folie, ‘Origine et premiers développements du Centre de Recherches sur les Primitifs flamands’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 27, 1996-1998, pp. 220-229.

Foot 2006 Miriam Foot, Bookbinders at Work. Their Roles and Methods, London, 2006. Fowler 1967 Harold North Fowler (ed.), Plato in twelve volumes: Eutyphro, Apology, Crtito, Phaedo, Phaedrus, London/Cambridge, 1967. Fraiture 1998 Pascale Fraiture, Etude dendrochronologique de tableaux des écoles septentrionales au XVIe siècle, méthode et application à travers quelques exemples conservés à Liège, Master Degree, University of Liège, 1998. Fraiture 2007 Pascale Fraiture, Les supports de peintures en bois dans les anciens PaysBas méridionaux de 1450 à 1650: analyses dendrochronologiques et archéologiques, PhD diss., University of Liège, 2007. Fraiture 2008 Pascale Fraiture, ‘Examen dendrochronologique du support’, in Dominique Vanwijnsberghe (ed.), Autour de la Madeleine Renders. Un aspect des collections, de la restauration et de la contrefaçon en Belgique dans la première moitié du XXe siècle (Scientia Artis, 4), Brussels, 2008, pp. 74-75. Fraiture 2009a Pascale Fraiture, ‘Dendrochronological Analysis of Pre-Eyckian Paintings’, in Cyriel Stroo (ed.), Pre-Eyckian Panel Painting in the Low Countries, 1, Catalogue (Contributions to Fifteenth-Century Painting in the Southern Netherlands and the Principality of Liège, 9), Brussels, 2009, pp. 47-69. Fraiture 2009b Pascale Fraiture, ‘Contribution of Dendrochronology to Understanding of Wood Procurement Sources for Panel Paintings in the Former Southern Netherlands from 1450 to 1650’, in Dendrochronologia, 27, 2009, pp. 95-111. Fraiture 2011 Pascale Fraiture, Report of Dendrochronological Analysis. Central Panels of the ‘Ghent Altarpiece’, Hubert and Jan van Eyck, Ghent, St Bavo Cathedral. Dossier IRPA/KIK 2008.09837, Brussels, 2011 [http:// closertovaneyck.kikirpa.be/#home/sub=documents]. Fraiture 2012 Pascale Fraiture, ‘Dendro-archaeological Examination of Paintings by Pieter Brueghel the Younger’, in Christina Currie and Dominique Allart, The Brueg[h]el phenomenon: Paintings by Pieter Bruegel the Elder and Pieter Brueghel the Younger with a Special Focus on Technique and Copying Practice (Scientia Artis, 8), Brussels, 2012, pp. 10021017. Fraiture, Crémer 2013 Pascale Fraiture and Sarah Crémer, Rapport d’analyse dendrochronologique. Structure de type « silo ». « Lohincou », rue de l’Avenir à Villers-leBouillet, Institut royal du Patrimoine artistique, (unpublished report), 2013. Frangenberg, Williams 2006 Thomas Frangenberg and Robert Williams (eds), The Beholder. The Experience of Art in Early Modern Europe, Aldershot, 2006. Fransen 2009 Bart Fransen, ‘Jan van Eyck, “el gran pintor del illustre duque de Borgoña”. El viaje a la Peninsula y “la Fuente de la Vida”’, in La senda española de los artistas flamencos, Madrid, 2009, pp. 105-125.

571 Fransen 2012a Bart Fransen, ‘Jan van Eyck and the Portraiture of his Predecessors’, in Stephan Kemperdick and Friso Lammertse (eds), The Road to Van Eyck, exh. cat., Rotterdam (Museum Boijmans Van Beuningen), 2012, pp. 76-79. Fransen 2012b Bart Fransen, ‘The Making of Portraits before Jan van Eyck: The Case of Wenceslas of Luxemburg’, in Marc De Mey, Maximiliaan P.J. Martens and Cyriel Stroo (eds), Vision and Material: Interaction between Art and Science in Jan van Eyck’s Time, Brussels, 2012, pp. 115127. Frémelet 2011 Armelle Frémelet, ‘Compte-rendu de Pierre Colman, Jan van Eyck et Jean sans Pitié (Académie royale de Belgique. Mémoire de la Classe des Arts. Collection in-8°, 3e série, t. XXVII), Brussels, 2009, pp. 73-78, 130-131’, in Histara, 25.8.2011 [http://histara.sorbonne.fr/cr. php?cr=1372]. Frezzato, Seccaroni 2010 Fabio Frezzato and Claudio Seccaroni, Segreti d’arti diverse nel regno di Napoli. Il manoscritto It. III.10 della Biblioteca Marciana di Venezia, Saonara (PD), Il Prato, 2010. Friedländer 1915 Max J. Friedländer, ‘Eyck, Jan (van)’, in Allgemeines Lexikon der bildenden Künstler, 11, 1915, pp. 129-133. Friedländer 1967 Max J. Friedländer, Early Netherlandish Painting, 1, The van EycksPetrus Christus, Leiden/Brussels, 1967. Friedländer 1924 Max J. Friedländer, Die altniederländische Malerei, 1, Die Van Eyck Petrus Christus, Berlin, 1924. Friedländer 1967-1976 Max J. Friedländer, Early Netherlandish Painting, 14 vols, Leiden, 19671976. Friedländer 1969 Max J. Friedländer, Early Netherlandish Painting, 4, Hugo van der Goes, Leiden/Brussels, 1969. Friedländer 1973 Max J. Friedländer, Early Netherlandish Painting, 9b, Joost van Cleve, Jan Provost, Joachim Patenier, Leiden/Brussels, 1973. Frinta 1963 Mojmir Frinta, ‘The use of Wax for Appliqué Relief Brocade on Wooden Statuary’, in Studies in Conservation, 8, 4, 186, pp. 136-149. Fris 1900 Victor Fris, ‘De onlusten te Gent in 1432-1435’, in Bulletijn der Maatschappij van Geschied- en Oudheidkunde te Gent/Bulletin de la Société d’Histoire et d’Archéologie de Gand, 8, Ghent, 1900, pp. 163-173. Fris 1907 Victor Fris, ‘Contributions à la Biographie Gantoise’, in Bulletijn van den Geschied- en Oudheidkundigen Kring van Gent/Bulletin du Cercle historique et archéologique de Gand, Ghent, 1907, pp. 62-92. Fromentin 1948 Eugène Fromentin, The Masters of Past Time. Dutch and Flemish Painting from Van Eyck to Rembrandt, Andrew Boyle (trans.), H. Gerson (ed.), London, 1948.

Fuks 1957 Lajb Fuks, The Oldest Known Literary Documents of Yiddish Literature (c. 1382), 2 vols, Leiden, 1957. Galassi 2006 Maria Clelia Galassi, ‘Pittura fiamminga per i Genovesi (secoli XV e XVI)’, in Piero Boccardo and Clario Di Fabio (eds), Genova e l’Europa atlantica. Opere, artisti, committenti e collezionisti. Inghilterra, Fiandre, Portogallo, Milan, 2006. Galassi 2007 Maria Clelia Galassi, ‘Episodi di “fiamminghismo” nella pittura ligure del Quattrocento. Verifiche sulla tecnica e sull’underdrawing’, in Caterina Limentani Virdis and Maddalena Bellavitis, Nord/Sud. Presenze e ricezioni fiamminghe in Liguria, Veneto e Sardegna. Prospettive di studio e indagini tecniche, Padua, 2007, pp. 33-46. García Marsilla 1996-1997 Juan Vicente García Marsilla, ‘El poder visible. Demanda y funciones del arte en la corte de Alfonso el Magnánimo’, in Ars longa: cuadernos de arte, 7-8, 1996-1997, pp. 33-47. Gaspar, Lyna 1984 Camille Gaspar and Frédéric Lyna F., Les Principaux Manuscrits à Peintures de la Bibliothèque Royale de Belgique, Deuxième partie, Brussels, 1984. Geelen 2012 Ingrid Geelen, ‘Modelling Splendour. Applied Brocade in the Ghent Altarpiece’, in Marc De Mey, Maximiliaan P.J. Martens and Cyriel Stroo (eds), Vision and Material. Interaction between Art and Science in Jan van Eyck’s Time, Brussels, 2012, pp. 129-139. Geelen, Steyaert 2009 Ingrid Geelen and Delphine Steyaert, ‘‘Ende wyldyt anders yet verheuen maken...’. Relief Decorations in the Art of around 1400’, in Cyriel Stroo (ed.), Pre-Eyckian Panel Painting in the Low Countries, 2, Essays (Contributions to Fifteenth-Century Painting in the Southern Netherlands and the Principality of Liège, 9), 2009, pp. 153-182. Geelen, Steyaert 2011 Ingrid Geelen and Delphine Steyaert, Imitation and Illusion. Applied Brocade in the Art of the Low Countries in the Fifteenth and Sixteenth Centuries (Scientia Artis, 6), Brussels, 2011. Geirnaert 1983 Noël Geirnaert, ‘De Adornes en de Jeruzalemkapel. Internationale contacten in het laatmiddeleeuwse Brugge’, in Adornes en Jeruzalem. International leven in het 15de- en 16de-eeuwse Brugge, exh. cat., Bruges (Jerusalem Chapel), 1983, pp. 11-49. Geirnaert 1987 Noël Geirnaert, Het Archief Van De Familie Adornes en de Jeruzalemstichting te Brugge.1. Inventaris, Bruges, 1987. Geirnaert 1989a Noël Geirnaert, Het Archief Van De Familie Adornes en de Jeruzalemstichting te Brugge. 2, Regesten van de Oorkonden en Brieven tot en met 1500, Bruges, 1989. Geirnaert 1989b Noël Geirnaert, ‘De Bibliotheek van de Jeruzalemkapel te Brugge tot ca. 1465’, in Biekorf, 89, 1989, pp. 313-321. Giernaert 1998 Noël Geirnaert, ‘Commentary: Some Experiences of an Archivist in Bruges’, in Early Netherlandish Painting at the Crossroads. A Critical Look

572 at Current Methodologies. Papers Presented at the Symposium on Nov. 7, 1998, at The Metropolitan Museum of Art, New York, NY, New York, 1998, pp. 40-45. Geirnaert 1999 Noël Geirnaert, ‘Anselm Adornes and his Daughters. Owners of Two Paintings of Saint Francis by Jan van Eyck?’, in Proceedings of the Jan van Eyck Symposium, The National Gallery, London, 13-14 March 1998, Turnhout, 1999, pp. 165-170.

Glatigny et al. 2010 Jean-Albert Glatigny et al., Report in progress: Observations on the supports and frames of the Ghent Altarpiece, 2010 (http://closertovaneyck. kikirpa.be/#home/sub=documents). Gludovatz 2005 Karin Gludovatz, ‘Der Name am Rahmen, der Maler im Bild. Künstlerselbstverständnis und Produktionskommentar in den Signaturen Jan van Eycks’, in Wiener Jahrbuch für Kunstgeschichte, 54, 2005, pp. 115175.

Giernaert 2000 Noël Geirnaert, ‘Anselm Adornes and his Daughters: Owners of Two Paintings of Saint Francis by Jan van Eyck?’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 165-170.

Goddard 1985 Stephen Goddard, ‘Brocade Patterns in the Shop of the Master of Frankfurt: An Accessory to stylistic analysis’, in The Art Bulletin, 67, 1985, 3, pp. 401-417.

Geirnaert,Vandewalle 1983 Noël Geirnaert, André Vandewalle (eds), Adornes en Jeruzalem. International leven in het 15de- en 16de-eeuwse Brugge, exh. cat., Bruges (Jerusalem Chapel), 1983.

Goedleven 1986 E. Goedleven, ‘De prijs van veiligheid: 8 jaar palaver, 5 maand werk’, in M&L Monuments, Landscapes & Archaeology in Flanders, 5.4, 1986, pp. 21-28.

Geschichte einiger Geistlichen Orden 1783 N.N., Geschichte einiger Geistlichen Orden, besonders derer, die in den kayserlich-königlichen erblanden bisher aufgehoben sind, Vienna, 1783.

Goethals 1849 Felix Goethals, Dictionnaire généalogique et héraldique des familles nobles du Royaume de Belgique, vol. 1, Brussels, 1849.

Ghent 1923 L’art ancien chez les collectionneurs gantois, exh. cat., Ghent (Cercle artistique et littéraire de Gand), 1923.

Goldschmidt 1930 Daisy Goldschmidt, Exposition internationale coloniale, maritime et d’art flamand: Section d’art flamand ancien, vol. I, Antwerp, 1930.

Giamatti 1966 A. Bartlett Giamatti, The Earthly Paradise and the Renaissance Epic, Princeton, 1966. Gifford 1995 Melanie Gifford, ‘Jan van Eyck’s Annunciation: Development and Alterations’, in Hélène Verougstraete and Roger Van Schoute (eds), Le dessin sous-jacent dans la peinture. Le dessin sous-jacent dans le processus de création, Colloque X, 5-7 septembre 1993, Louvain-La-Neuve, 1995, pp. 85-93. Gifford 1999 E. Melanie Gifford, ‘Van Eyck’s Washington Annunciation: Technical Evidence for Iconographic Development’, in The Art Bulletin, 81, 1999, 1, pp. 108-116. Gifford 2000 Melanie E. Gifford, ‘Assessing the Evolution of Van Eyck’s Iconography through Technical Study of the Washington Annunciation’, in Susan Foister, Susan Jones and Delphine Cool (eds), Investigating Van Eyck, London, 2000, pp. 59-66. Gifford et al. 2013 Melanie E. Gifford et al., ‘Jan van Eyck’s Washington ‘Annunciation.’ Painting Materials and Techniques’, in Facture, 1, 2013, pp. 128-153. Gilliodts-Van Severen 1871-1878 Louis Gilliodts-Van Severen, Inventaire des archives de la ville de Bruges, Inventaire des chartes, 1ère section, 6 vols, Bruges, 1871-1878. Gilliodts-Van Severen 1892 Louis Gilliodts-Van Severen, Coutumes des Pays et comté de Flandre. Quartier de Bruges (Coutumes des petites villes et seigneuries enclavées, 4), Brussels, 1892. Gilliodts-Van Severen 1904 Louis Gilliodts-Van Severen, Cartulaire de l’Ancienne Estaple de Bruges, V, 1, Bruges, 1904.

Gombrich 1976 Ernst H. Gombrich, The Heritage of Apelles: Studies in the Art of the Renaissance, Ithaca, 1976. Gómez Frechina 2001 José Gómez Frechina, ‘Algunas pautas flamencas en la pintura valenciana del siglo XV’, in La Clave Flamenca en los Primitivos Valencianos, exh. cat., Valencia (Museo de Bellas Artes), 2001, pp. 63-103. Gómez Frechina 2007 José Gómez Frechina, ‘La estética Flamenca en la pintura Valenciana: testimonios, influencias y protagonistas’, in Eduard Mira and An Delva (eds), A la búsqueda del Toisón de Oro. La Europa de los Príncipes. La Europa de las ciudades, vol. 1, exh. cat., Almudín (Museo de la ciudad), 2007, pp. 381-405. Goodgal 1981 Dana R. Goodgal, The Iconography of the Ghent Altarpiece, PhD diss., University of Pennsylvania, 1981 (Ann Arbor, 1984). Goodgal 1985 Dana R. Goodgal, ‘Joos Vijd: donateur de l’Agneau Mystique’, in Roger Van Schoute and Dominique Hollanders-Favart (eds), Le dessin sousjacent dans la peinture, Colloque V, 29-30 septembre- 1er octobre 1983. Dessin sous-jacent et autres techniques graphiques, Louvain-la-Neuve, 1985, pp. 25-52. Grandmaison 1912 Louis de Grandmaison, A Propos d’un Livre d’Heures écrit pour Michel Colombe par Pierre Fabri, de Bourges, en 1487, Réunion des Societés des beaux-arts des départements, 1912. Grimm 1888 Jacob Grimm, Teutonic Mythology, J. S. Stallybrass (trans.), London, 1888. Grosshans 1996 Reinhald Grosshans, Gemäldegalerie Berlin. Gesamtverzeichnis, Berlin, 1996.

573 Gysseling 1998 M. Gysseling (ed.), Der naturen bloeme. Jacob van Maerlant, The Hague/ Antwerp, 1998.

Hermens 1998 Erma Hermens (ed.), Looking Through Paintings. The Study of Painting Techniques and Materials in Support of Art Historical Research (Leids Kunsthistorisch Jaarboek, 11), Baarn/London, 1998.

Hall 1968 Edwin Hall, ‘Cardinal Albergati, St Jerome and the Detroit van Eyck’, in Art Quarterly, 31, 1968, pp. 2-34.

Herzner 1995 Volker Herzner, Jan van Eyck und der Genter Altar, Worms, 1995.

Halpine 1993 Susana M. Halpine, Analytical report, NGA Scientific Research Department, 28 July 1993.

Herzner 2011a Volker Herzner, Kunstgeschichte, 2011 [http://www.kunstgeschichteejournal.net/140/1/Lebensbrunnen.pdf].

Hand, Wolff 1986 John Oliver Hand and Martha Wolff, Early Netherlandish Painting (The Collections of the National Gallery of Art Systematic Catalogue), Washington, 1986.

Herzner 2011b Volker Herzner, ‘Der Madrider Lebensbrunnen aus der Werkstatt Jan van Eycks und die zielsicheren Irrwege der Forschung’, in Kunstgeschichte: Open Peer Reviewed Journal, 2011 [http://www.kunstgeschichteejournal. net/247/ I /Herzner ~ Madrider ~ Lebensbrunnen. pdf].

Hand, Metzger, Spronk 2006-2007 John Oliver Hand, Catherine A. Metzger and Ron Spronk, ‘The Diptych Format in Netherlandish Painting’, in Prayers and Portraits. Unfolding the Netherlandish Diptych, exh. cat., Washington (National Gallery of Art)/Antwerp (Koninklijk Museum voor Schone Kunsten Antwerpen), 2006-2007, pp. 1-13. Harbison 1991 Craig Harbison, Jan van Eyck. The Play of Realism, London, 1991. Harbison 1995 Craig Harbison, The Mirror of the Artist, Perspectives, Ann Arbor, 1995. Harbison 2012 Craig Harbison, Jan van Eyck. The Play of Realism (updated and expanded edition), London, 2012 (1991). Hargreaves-Mawdsley 1963 William Hargreaves-Mawdsley, History of legal dress in Europe until the end of the eighteenth century, Oxford, 1963. Hartweig 2005 Babette Hartwieg, ‘Rubens “Madonna mit stehendem Kind”, in Hannover – ein restauriergeschichteliches Exempel’, in Beiträge zur Erhaltung von Kunst- und Kulturgut, 2, 2005, pp. 38-46. Haydocke 1598 Richard Haydocke, Tracte containing the Artes of curious Paintinge Cruinge and Buildinge written first in Italian by Jo. Paul Lomatius painter of Milan and englished by R.H. student in Physik, London, 1598 (trans. of 1584 edn of Lomazzo). Heath 2008 Peter Heath, ‘Justice & Mercy. The Patron of Jan van Eyck’s Dresden Triptych’, in Apollo, March 2008, pp. 106-113. Heers 1971 Jacques Heers, Gênes au XVe siècle, Paris, 1971. Henderiks 2009 Valentine Henderiks, ‘Compte rendu de Pierre Colman Jan van Eyck et Jean sans Pitié (Académie royale de Belgique. Mémoire de la Classe des Arts. Collection in-8°, 3e série, t. XXVII), Brussels, 2009, pp. 73-78, 130-131’, in Revue belge de philologie et d’histoire, 87, 2009, pp. 848-852. Hendrix 2007-2012 Lee Hendrix, ‘Hoefnagel, Joris [Georg]’, in Grove Art online, Oxford University Press, 2007-2012 [http://www.oxfordartonline.com (accessed December 2012)].

Herzner 2011c Volker Herzner, ‘A Response to Hugo van der Velden, “The quatrain of The Ghent Altarpiece”‘, in Simiolus, 35, 2011, pp. 127-130. Herzner 2013-2014 Volker Herzner, ‘The Quatrain of The Ghent Altarpiece, Again’, in Simiolus, 37, 2013-2014, pp. 95-99. Hicks 2012 Carola Hicks, Girl in a Green Gown, Vintage Books, 2012. Higgitt, Spring, Saunders 2003 Catherine Higgitt, Marika Spring and David Saunders, ‘Pigmentmedium Interactions in Oil Paint Films containing Red Lead or Lead-tin Yellow’, in National Gallery Technical Bulletin, 24, 2003, pp. 75-95. Higgitt, White 2005 Catherine Higgitt and Raymond White, ‘Analyses of Paint Media: New Studies of Italian Paintings of the Fifteenth and Sixteenth Centuries’, in National Gallery Technical Bulletin, 26, 2005, pp. 89-104. Hillam, Morgan, Tyers 1987 Jennifer Hillam, R.A. Morgan and Ian Tyers, ‘Sapwood Estimates and the Dating of Short Ring Sequences’, in R.G.W. Ward (ed.), Applications of Tree-ring Studies, Current Research in Dendrochronology and Related Subjects (British Archaeological Reports, International Series, 333), 1987, pp. 165-185. Hinz 1974 Berthold Hinz, ‘Studien zur Geschichte des Ehepaarbildnisses’, in Marburger Jahrbuch für Kunstwissenschaft, 19, 1974, pp. 139-218. Hirschfeld 1970 Peter Hirschfeld, ‘Hat Meister Francke für Jean sans Peur von Burgund gearbeitet?’, in Zeitschrift für Kunstgeschichte, 33, 1970, pp. 54-60. Hodge, Spring, Marchant 1998 Sam Hodge, Marika Spring and Ray Marchant, ‘The construction and painting of a large Castilian retable: a study of techniques and workshop practices’, in Ashok Roy and Perry Smith (eds), Painting Techniques: History, Materials and Studio Practice (Contributions to the Dublin Congress 7-11 September 1998), London, 1998. Hodge, Spring, Marchant 2000 Sam Hodge, Marika Spring and Ray Marchant, with Zahira Veliz, ‘The Santa Maria retable from Mayorga, attributed to the Master of Palanquinos, c. 1490s’, in Hamilton Kerr Institute Bulletin, 3, Cambridge, 2000, pp. 7-40.

574 Hoffmann 1985 Richard C. Hoffmann, ‘Fishing for Sport in Medieval Europe: New Evidence’, in Speculum, 60, 1985, pp. 877-902. Hoffsummer 1995 Patrick Hoffsummer, Les charpentes de toitures en Wallonie. Typologie et dendrochronologie (XIe – XIXe siècle), Namur, 1995. Hollander 1978 Anne Hollander, Seeing Through Clothes, New York, 1978. Hoste 1993 D. Anselm Hoste, De Handschriften van ter Doest, Steenbrugge, 1993. Houbrechts, Vandervellen 2011 David Houbrechts and Pascale Vandervellen, ‘Dendrochronology applied to Belgian pianofortes’, in Pascale Fraiture (ed.), Tree Rings, Art, Archaeology. Proceedings of the conference, Brussels, Royal Institute for Cultural Heritage, 10-12 February 2010 (Scientia Artis, 7), Brussels, 2011, pp. 285-297. Houtart n.d. Maurice Houtart, Jacques Daret: peintre tournaisien du XVe siècle, Tournai, n.d. Howard 2003 Helen Howard, Pigments of English Medieval Wall Painting, London, 2003. Howarth 1985 David Howarth, Lord Arundel and his Circle, New Haven/London, 1985. Hulin de Loo 1911 Georges Hulin de Loo, Heures de Milan. Troisième partie des Très Belles Heures de Notre-Dame enluminées par les peintres de Jean de France, duc de Berry, et par ceux de Guillaume de Bavière, comte du Hainaut et de Hollande, Brussels/Paris, 1911. Hulin de Loo 1931 Georges Hulin de Loo, ‘Les Voyages des frères van Eyck avant 1425’, in Académie Royale de Belgique. Bulletin de la Classe des Beaux-Arts, 13, 1931, pp. 123-134. Hunter 1993 John Hunter, ‘Who is Jan van Eyck’s “Cardinal Nicolo Albergati”?’, in The Art Bulletin, 75, 1993, pp. 207-218. Hussmann 1977 Heinrich Hussmann, Über die Schrift, Wiesbaden, 1977. Hymans 1883 Henry Hymans, ‘Un tableau retrouvé de Jean van Eyck’, in Bulletin des Commissions Royales d’art et d’Archéologie, 22, 1883, pp. 108-116.

Itinéraire de Bruges (XVe siècle) 1908 Itinéraire de Bruges (XVe siècle), in Gilles le Bouvier, dit [le héraut] Berry, Le Livre de la description des pays, E. T. Hamy (ed.), Paris, 1908, Appendice IV, pp. 157-216. Jaerboeck van Buyck 1898 ‘Jaerboeck van Buyck’, in Annales de la Société de l’Emulation de Bruges, 6/1, 1898, pp. 297-378. Jamois 2005-2006 Thibault Jamois, Röser, peintre et restaurateur, 1737-1804, mémoire de muséologie de l’École du Louvre, sous la direction de Nathalie Volle, École du Louvre, Paris, 2005-2006. Jansen 1988 Dieter Jansen, Similitudo. Untersuchungen zu den Bildnissen Jan van Eycks, Cologne/Vienna, 1988. Jansen 1989 Dieter Jansen, ‘Jan van Eycks Selbstbildnis – der “Mann mit dem roten Turban” und der sogenannte “Timotheus” der Londoner National Gallery’, in Pantheon, 47, 1989, pp. 36-48. Jansma et al. 2012 Esther Jansma et al., ‘The DCCD: A digital data infrastructure for tree-ring research’, in Dendrochronologia, 30, 2012, pp. 249-251. Janssens de Bisthoven 1944 Aquilin Janssens de Bisthoven, ‘Het beeldhouwwerk van het Brugsche stadhuis’, in Gentsche Bijdragen tot de Kunstgeschiedenis,10, 1944, pp. 7-81. Janssens de Bisthoven 1981 Aquilin Janssens de Bisthoven, Stedelijk Museum voor Schone Kunsten (Groeningemuseum) Brugge (Corpus van de vijftiende-eeuwse Schilderkunst in de Zuidelijke Nederlanden,1), 3rd edn, (in French, trans. from the 2nd edition in Dutch, revised and augmented with the collaboration of M. Baes-Dondeyne and D. De Vos), Brussels, 1981. Janssens de Bisthoven 1983 Aquilin Janssens de Bisthoven, Musée Communal des Beaux-Arts (Musée Groeninge) Bruges (Corpus de la peinture des anciens Pays-Bas méridionaux au quinzième siècle, 1), 3rd edn, (in French, trans. from the 2nd edition in Dutch, revised and augmented with the collaboration of M. Baes-Dondeyne and D. De Vos), Brussels, 1983. Jerome 1982 Jerome, The Principal Works of Saint Jerome, Philip Schaff (ed.), New York, 1892. Jolly 1987 Penny Howell Jolly, ‘More on the Van Eyck question: Philip the Good of Burgundy, Isabelle of Portugal, and the Ghent Altarpiece’, in OudHolland, 101, 1987, pp. 237-253.

Hymans 1888 Henri Hymans, ‘Le Saint François d’Assise de Jean van Eyck’, in Gazette des Beaux-Arts, 2, 37, 1888, pp.78-83.

Jolly 1998 Penny H. Jolly, ‘Jan van Eyck’s Italian Pilgrimage. A Miraculous Florentine Annunciation and the “Ghent Altarpiece”’, in Zeitschrift für Kunstgeschichte, 61, 3, 1998, pp. 369-394.

IJpelaar 2012 Linda IJpelaar, Vispartij of valkenjacht? Een onderzoek naar de herkomst en voorstelling van het Vispartijtje, MA thesis, University of Amsterdam, Amsterdam, 2012.

Jones 2000 Sue Jones, ‘The use of patterns by Jan van Eyck’s assistants and followers’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 197-207.

Isaac 1984 Marie-Thérèse Isaac, Les Livres Manuscrits de l’Abbaye des Dunes d’après le Catalogue du XVIIe siècle, Aubel, 1984.

Jones 2014 Susan Frances Jones, ‘Jan van Eyck and Spain’, in Boletín del Museo del Prado, 32/50, 2014, pp. 30-49.

575 Joubert 1992 Fabienne Joubert, ‘Le Saint Christophe de Semur-en-Auxois’, in Bulletin monumental, 150/II, 1992, pp. 165-177.

Kemperdick, Lammertse 2012 Stephan Kemperdick and Friso Lammertse, The Road to Van Eyck, exh. cat., Rotterdam (Museum Boijmans Van Beuningen), 2012.

Kaczynski 1988 Bernice M. Kaczynski, Greek in the Carolingian Age. The St.Gall Manuscripts, Cambridge, MA, 1988.

Kempis 1949 Thomas À Kempis, Imitation of Christ, Book I, Milwaukee, 1949.

Kaemmerer 1898 Ludwig Kaemmerer, Hubert und Jan van Eyck, Bielefeld/Leipzig, 1898. Kahsnitz 1971 Rainer Kahsnitz, ‘The Gospel Book of Abbess Svanhild of Essen in the John Rylands Library: II’, in The Bulletin of the John Rylands Library, 53, 1971, pp. 360-396. Kay, Young, Posner 1964 M.I. Kay, R.A. Young and A.S. Posner, ‘Crystal Structure of hydroxyapatite’, in Nature, 204, 1964, pp. 1050-1052. Keck 1984 Sheldon Keck, ‘Some Picture Cleaning Controversies: Past and Present’, in Journal of the American Institute for Conservation, 1984, vol. 23, 2, pp. 73-87. Keith et al. 2011 Larry Keith, Ashok Roy, Rachel Morrison and Peter Schade, ‘Leonardo da Vinci’s Virgin of the Rocks: Treatment, Technique and Display’, in National Gallery Technical Bulletin, 32, 2011, pp. 32-56. Keller 1973 Renate Keller, ‘Leinöl als Malmittel’, in Maltechnik, vol. 2, 1973, pp. 74-105. Kemp 1991 Martin Kemp, ‘Linear Perspective from Brunelleschi to Leonardo’, in The Science of Art, 1990, and in Kunstchronik, 44, 5, 1991, pp. 285290. Kemperdick 1997 Stephan Kemperdick, Der Meister von Flémalle: die Werkstatt Robert Campins und Rogier van der Weyden (Ars Nova: Studies in Late Medieval and Renaissance Northern Painting and Illumination, 2), Turnhout, 1997. Kemperdick 2010a Stephan Kemperdick with Beatrix Graf and Regina Cermann, Deutsche und Böhmische Gemälde 1230-1430, Kritischer Bestandskatalog, Staatliche Museen zu Berlin, Gemäldegalerie, Petersberg, 2010. Kemperdick 2010b Stephan Kemperdick, ‘The First Generation’, in Till-Holger Borchert (ed.), Van Eyck to Dürer. Early Netherlandish Painting and Central Europe 1430-1530, exh. cat., Bruges (Groeningemuseum), 2010, pp. 5567. Kemperdick 2014 Stephan Kemperdick, ‘Die Geschichte des Genter Altars’, in Stephan Kemperdick and Johannes Rößler, Der Genter Altar der Brüder Van Eyck, exh. cat., Berlin (Staatliche Museen zu Berlin), 2014, pp. 9-69. Kemperdick, Klein 1997 Stephan Kemperdick and Peter Klein, ‘The Evidence of Dendrochronological and Art Historical Dating in Paintings of the Rogier van der Weyden Group’, in Roger Van Schoute and Hélène Verougstraete (eds), Le dessin sous-jacent et la technologie dans la peinture. Colloque XI: Dessin sous-jacent et technologie de la peinture. Perspectives, 14-16 septembre 1995, Louvain-la-Neuve, 1997, pp. 143-151.

Kern 1912 Joseph G. Kern, ‘Perspektive und Bildarchitektur bei Jan van Eyck’, in Repertorium für Kunstwissenschaft, 35, 1912, pp. 27-64. Kessler 1977 Herbert L. Kessler, The Illustrated Bibles from Tours (Studies in manuscript illumination,71), Princeton/Guilford, 1977. Kessler 2000 Herbert L. Kessler, Spiritual Seeing: Picturing God’s Invisibility in Medieval Art, Philadelphia, 2000. Kessler 2002 Herbert L. Kessler, Old Saint Peter’s and Church Decoration in Medieval Italy, Spoleto, 2002. Kessler, Wolf 1998 Herbert L. Kessler and Gerhard Wolf (eds), The Holy Face and the Paradox of Representation: papers from a colloquium held at the Bibliotheca Hertziana, Rome and the Villa Spellman, Florence, 1996, Bologna, 1998. Kessler, Zacharias 2000 Herbert L. Kessler and Johanna Zacharias, Rome 1300: On the path of the pilgrim, New Haven/London, 2000. Ketelsen et al. 2005 Thomas Ketelsen, Ina Reiche, Olaf Simon and Silke Merchel, ‘New Information on Jan van Eyck’s Portrait Drawing in Dresden’, in The Burlington Magazine, 147, no. 1224, London, 2005, pp. 170-175. Kieckhefer 1989 Richard Kieckhefer, Magic in the Middle Ages, Cambridge, 1989. Kilmer 1982 Martin Kilmer, ‘Genital Phobia and Depilation’, in Journal of Hellenic Studies, 52, 1982, pp. 104-112. Kirby 2012 Jo Kirby, ‘Aspects of Oil: Painting in Northern Europe and Jan van Eyck’, in Marc De Mey, Maximiliaan P.J. Martens and Cyriel Stroo (eds), Vision and Material. Interaction between Art and Science in Jan van Eyck’s Time, Brussels, 2012, pp. 255-278. Kirby, Nash, Cannon 2010 Jo Kirby, Susie Nash and Joanna Cannon (eds), Trade in Artists’ Materials. Markets and Commerce in Europe to 1700, London, 2010. Kirby, Saunders, Spring 2006 Jo Kirby, David Saunders and Marika Spring, ‘Proscribed pigments in Northern European Renaissance paintings and the case of Paris Red’, in David Saunders, Joyce H. Townsend and Sally Woodcock (eds), The Object in Context: Crossing Conservation Boundaries, Contributions to the Munich IIC Congress, 28 August–1 September 2006 , London, 2006, pp. 236-243. Kirby, Spring, Higgitt 2005 Jo Kirby, Marika Spring and Catherine Higgitt, ‘The Technology of Red Lake Pigment Manufacture: Study of the Dyestuff Substrate’, in National Gallery Technical Bulletin, 26, 2005, pp. 71-87. Kirchner 1955 Joachim Kirchner, Scriptura Latina libraria, Oldenburg, 1955, p. 57.

576 Klaas 2011 Jörg Klaas, Die “Ultramarinkrankheit”. Studien zu Veränderungen in ultramarinhaltigen Farbschichten an Gemälden, Munich, 2011. Klein 1983 Peter Klein, ‘Dating of Art-historical Objects’, in Dieter Eckstein, Sigrid Wrobel and R.W. Aniol (eds), Dendrochronology and Archaeology in Europe (Mitteilungen der Bundesforshungsanstalt für Forst und Holzwirtschaft, 141), Hamburg, 1983, pp. 209-222. Klein 1997 Peter Klein, ‘Dendrochronological Analyses of the Two Panels of “Saint Francis Receiving the Stigmata”’, in Jane Watkins (ed.), Jan van Eyck: Two Paintings of ‘Saint Francis Receiving the Stigmata’, Philadelphia, 1997, pp. 47-50. Klein 2010 Peter Klein, Report on the dendrochronological analysis of the panels “Adam and Eve” (Jan van Eyck), Universität Hamburg, 2010 [http:// closertovaneyck.kikirpa.be/#home/sub=documents]. Kmet 2003-2007 Karel Kmet, ‘Was Jan van Eyck in Praag ?’, in Agora Cultural Travels, 2003-2007. Kneepkens 2012 Indra Kneepkens, Understanding historical recipes for the modification of linseed oil: An experimental study into the properties of modified linseed oils for use as binding media in early northern European panel painting, Masters thesis, Department of Art History, University of Amsterdam, 2012. Knight 2012 Amy Knight Powell, Depositions: Scenes from the Late Medieval Church and the Modern Museum, New York, 2012. Kockaert 1973-1974 Leopold Kockaert, ‘Note sur les émulsions des Primitifs flamands’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 14, 1973-1974, pp. 133139.

König 1992 Eberhard König, Die Très Belles Heures de Notre-Dame des Herzogs von Berry, ms nouv. acq. lat. 3093 der Bibliothèque nationale, Paris, Lucerne, 1992. König 1998 Eberhard König, Die Très Belles Heures von Jean de France Duc de Berry, Lucerne, 1998. König 2005 Eberhard König, ‘Zur Farbigkeit der verbrannten Gefangennahme im Turiner Gebetbuch’, in Mara Hofmann and Caroline Zöhl (eds), Quand la Peinture était dans les livres, Mélanges en l’honneur de François Avril, Turnout/Paris, 2005, pp. 110-127. König et al. 1994 Eberhard König et al., Die Blätter im Louvre und das verlorene Turiner Gebetbuch, RF 2022-2025, Département des arts graphiques Musée du Louvre, Paris, und Handschrift K.IV.29, Biblioteca Nazionale Universitaria, Turin, Lucerne 1994. Kopp-Schmidt 2010 Gabriele Kopp-Schmidt, ‘Neue Blicke – Gefährliche Frauen’, in Anne Marie Bonnet and Gabriele Kopp-Schmidt, Die Malerei der deutschen Renaissance 1484-1555, Munich, 2010, pp. 41-48. Körner 1993 Joseph Leo Körner, Albrecht Dürer and the Moment of Self-Portraiture in German Renaissance Art, Chicago, 1993. Korteweg 2005 Anne Korteweg, ‘The form and content of Jean de Berry’s Books of Hours’, in Rob Dückers and Pieter Roelofs (eds), The Limbourg Brothers, Nijmegen Masters at the French Court 1400-1416, exh. cat., Nijmegen (Museum het Valkhof), 2005, pp. 134-147. Kren, McKendrick 2003 Thomas Kren and Scot McKendrick, Illuminating the Renaissance: The Triumph of Flemish Manuscript Painting in Europe, exh. cat., Los Angeles/ London, 2003, pp. 427-433.

Kockaert, Verrier 1978-1979 Leopold Kockaert and Monique Verrier, ‘Application des colorations à l’identification des liants de van Eyck’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 14, 17, 1978-1979, pp. 122-127.

Kroustallis, Bruquetas, Gómez 2011 Stefanos Kroustallis, Rocio Bruquetas and Marisa Gómez, ‘In Quest of Vermilion: Production, Commerce and Use of the Pigment in Eighteenth-Century Spain’, in Marika Spring (ed.), Studying Old Master Paintings. Technology and Practice, London, 2011, pp. 277283.

Kockaert, Gausset, Dubi-Rucquoy 1989 Leopold Kockaert, P. Gausset and M. Dubi-Rucquoy, ‘Detection of ovalbumin in paint media by immunofluorescence’, in Studies in Conservation, 34(4), pp. 183-188.

Kupfer 2008 Marcia Kupfer, ‘“…lectres …plus vrayes”: Hebrew Script and Jewish Witness in the Mandeville Manuscript of Charles V’, in Speculum, 83, 2008, pp. 58-111.

Koerner 1993 Joseph Leo Koerner, The Moment of Self-Portraiture in German Renaissance Art, Chicago/London, 1993.

Kurz 1956 Otto Kurz, ‘A Fishing Party at the Court of William VI Count of Holland, Zeeland and Hainaut. Notes on a Drawing in the Louvre’, in Oud Holland, 71, 1956, pp. 117-131.

Köhler 1906-1907 Wilhelm Köhler, ‘Aktenstücke zur Geschichte der Wiener Kunstkammer in der herzoglichen bibliothek zu Wolffenbüttel’, in Jahrbuch der kunsthistorischen Sammlungen der allerhöchsten Kaiserhauses, 26, 2, 1906-1907, pp. I-XX. König 1982 Eberhard König, Französische Buchmalerei um 1450, der Jouvenel-Maler, der Maler des Genfer Boccaccio und die Anfänge Jean Fouquets, Berlin, 1982.

Kuys 2006 Jan Kuys, ‘Weltliche funktionen spätmittelalterlicher Pfarrkirchen in den nördlichen Niederlanden’, in Paul Trio and Marjan De Smet, The use and abuse of sacred places in late medieval towns (Mediaevalia Lovaniensia, series I/Studia XVIII), 2006, pp. 27-45. Lacarra Ducay 2007 María del Carmen Lacarra Ducay, La pintura gótica durante el siglo XV en tierras de Aragón y en otros territorios peninsulares, Zaragoza, 2007.

577 Laemers 2008 Suzanne Laemers, ‘“A matter of character”. Max J. Friedländer et ses relations avec Emile Renders et Jef Van der Veken’, in Dominique Vanwijnsberghe (ed.), Autour de la Madeleine Renders. Un aspect de l’histoire des collections, de la restauration et de la contrefaçon en Belgique dans la première moitié du XXe siècle (Scientia Artis, 4), Brussels, 2008, pp. 147-176. Laenen 1934 Jean Laenen, Geschiedenis van Mechelen tot op het einde der Middeleeuwen, Mechelen, 1934. Laffitte 1989 Marie-Pierre Laffitte, ‘Le vocabulaire médiéval de la reliure d’après les anciens inventaires’, in Olga Weijers (ed.), Vocabulaire du livre et de l’écriture au Moyen Âge. Actes de la table ronde, Paris 24-26 septembre 1987, Turnhout, 1989, pp. 61-78. Laleman 1984 Marie-Christine Laleman, ‘Gentbrugge, Bassijnstraat 51’, in Stadsarcheologie, 8, 1984, pp. 61-62. Lambert 2006 Georges-Noël Lambert, Dendrochronologie, histoire et archéologie, modélisation du temps. Le logiciel ‘Dendron II’ et le projet ‘Historic Oaks’, Thèse d’Habilitation à Diriger les Recherches, Université de Franche-Comté, 2006. Lamont 2005 Koen Lamont, Het wereldbeeld van een zestiende-eeuwse Gentenaar Marcus van Vaernewijck: een ideeën- en mentaliteitshistorische studie op basis van zijn kroniek, Ghent, 2005. Lane 1970 Barbara G. Lane, ‘Petrus Christus: A Reconstructed Triptych with an Italian motif’, in The Art Bulletin, 52, 1970, pp. 390-393.

Leclerq 1861 Emile Leclerq, ‘Van Eyck et Stuerbout’, in Gazette des Beaux-Arts, 2, 1861, pp. 283-290. Lederman, Klatzky 2004 Susan J. Lederman and Roberta L. Klatzky, ‘Multisensory Texture Perception’, in G. Calvert, C. Spence and B. Stein (eds), Handbook of Multisensory Processes, Cambridge, 2004, pp. 107-122. Leeuwenberg 1965 Jaap Leeuwenberg, ‘Die Ausstrahlung Utrechter Tonplastik’, in Studien zur Geschichte der europäischen Plastik; Festschrift Theodor Müller, Munich, 1965, pp. 151-166. Lehmann 2004 Ann-Sophie Lehmann, Mit Haut und Haar. Jan van Eyck’s Adam und Eva Tafeln des Genter Altars, PhD diss., University of Utrecht, 2004. Lehmann 2008 Ann-Sophie Lehmann, ‘Fleshing out the Body. The Colours of the Naked in Dutch Art Theory and Workshop Practice 1400-1600’, in H. Roodenburg (ed.), Body and Embodiment in Netherlandish Art, 2008, pp. 108-131. Lehmann 2009 Ann-Sophie Lehmann, ‘The Missing Sex. Absence and Presence of a Female Body Part in the Visual Arts’, in Barbara Baert and Hilde van Gelder (eds), Fluid Flesh: The Body, Religion and the Visual Arts, Leuven, 2009. Lehmann 2012 Ann-Sophie Lehmann, ‘Voetstappen op weg naar Van Eyck’, in Kunstschrift, 5/56, 2012, pp. 8-13. Lehmann 2014 Ann-Sophie Lehmann, In the Flesh. Jan van Eyck’s Adam and EvePanels and the Making of the Northern Nude, 2014.

Lane 1984 Barbara G. Lane, The Altar and the Altarpiece: Sacramental Themes in Early Netherlandish Painting, New York, 1984.

Leitch 2008 Stephanie Leitch, ‘The Wild Man. Charlemagne and the German Body’, in Art History, 31, 2008, pp. 283-302.

Lane 1988 Barbara G. Lane, ‘Sacred versus profane in early Netherlandish painting’, in Simiolus, 18, 3, 1988, pp. 107-115.

Lejeune 1956 Jean Lejeune, Les Van Eyck, peintres de Liège et de sa cathédrale, Liège, 1956.

Lauer 1992 R. Lauer, ‘Köln in Flandern. Ein Domansicht auf dem Genter Altar’, in Kölner Domblatt, 57, 1992, pp. 309-314.

Lejeune 1972 Jean Lejeune, Jean et Marguerite van Eyck et le roman des Arnolfini, Liège, 1972.

Laurie 1910 Arthur P. Laurie, The Materials of the Painter’s Craft in Europe and Egypt, London, 1910.

Lemaire, Henry, Rouzet 1991 Claudine Lemaire, Michèle Henry and Anne Rouzet, Isabelle van Portugal, hertogin van Bourgondië, 1397-1471, exh. cat., Brussels (Koninklijke Bibliotheek Albert I), 1991.

Laurie 1926 Arthur P. Laurie, The Painter’s Methods and Materials, London, 1926. Laurie 1933 Arthur P. Laurie, ‘The Use of an Emulsion as a Painting Medium’, in Technical Studies in the Field of the Fine Arts, 2(1), 1933, pp. 11-14. Laurie 1934 Arthur P. Laurie, ‘Notes on the Medium of Flemish painters’, in Technical Studies in the Field of the Fine Arts, 2(3), 1934, pp. 124-128. Le Glay 1839 André J. G. Le Glay, Correspondance de l’Empereur Maximilien et de Marguerite d’Autriche, sa fille, Gouvernante des Pays-Bas, de 1507 à 1519. Publiée d’après les manuscrits originaux, 2 vols, Paris, 1839.

Leonardi 2007 Renzo Leonardi, ‘Jan van Eyck, les Alpes et la lune’, in Revue de l’Art, 155, 2007, 1, pp. 51-57. Leonardi 2010 Renzo Leonardi, in Revue de l’Art, 168, 2010-2012, p. 77 (note 14). Leroquais 1934 Victor Leroquais, Les bréviaires manuscrits des bibliothèques publiques de France, 6 vols, Paris, 1934. Lessing 1774 Gotthold E. Lessing, Vom Alter der Oelmalerey, aus dem Theophilus Presbyter, Braunschweig, 1774.

578 Letts 1949 Malcolm Letts, Sir John Mandeville. The Man and his Book, Batchworth Press, London, 1949. Leuven 1993 Maurits Smeyers, Vlaamse Minaturen voor Van Eyck (ca. 1380 – ca. 1420). Catalogus, (Corpus of Illuminated Manuscripts, 6, Low Countries series, 4), exh. cat., Leuven (Cutureel Centrum Romaanse Poort), 1993. Leuven 2009 Lorne Campbell and Jan van der Stock, Rogier van der Weyden, Master of Passions, exh. cat., Leuven (M – Museum), 2009. Lichterstein 1971 Aaron Lichterstein, ‘Portugal’, in Encyclopedia Judaica, 13, Jerusalem, 1971, cols 919-922. Liège 1968 Liège et Bourgogne, exh. cat., Liège (Musée de l’art wallon), 1968. Lievois 2000 Daniel Lievois, ‘Op bezoek bij de hertogen van Bourgondië’, in M.C. Laleman (ed.), Het Prinselijk Hof ten Walle in Gent, Ghent, 2000, pp. 77-200. Lomax 1993 Suzanne Q. Lomax, Analytical report, NGA Scientific Research Department, September 21, 1993. Lomax 2000 Suzanne Q. Lomax, Analytical report, NGA Scientific Research Department, October 30, 2000. Lomax 2012 Suzanne Q. Lomax, Analytical report, NGA Scientific Research Department, January 13, 2012. London 2006 Ernst Vegelin van Claerenberge (ed.), David Tenier and the Theatre of Painting, exh. cat., London (Courtauld Institute of Art Gallery), 2006. London 2008 Lorne Campbell, Miguel Falomir, Jennifer Fletcher and Luke Syson (eds), Renaissance Faces. Van Eyck to Titian, exh. cat., London (The National Gallery), 2008. Los Angeles/London 2003 Thomas Kren and Scot McKendrick (eds), Illuminating the Renaissance, the Triumph of Flemish Manuscript Painting in Europe, exh. cat., Los Angeles (J. Paul Getty Museum)/London (Royal Academy), 2003.

Luttervelt 1957 R. van Luttervelt, ‘Bijdragen tot de Iconographie van de Graven van Holland naar aanleiding van de beelden uit de Amsterdamse vierschaar II’, in Oud Holland, 72, 1957, pp. 139-149. Luxembourg/Budapest 2006 Sigismund Rex et Imperator, exh. cat., Luxembourg (Musée National d’Histoire et d’Art) /Budapest (Museum of Fine Arts), 2006. Lyna 1955 Frédéric Lyna, ‘Les Van Eyck et les Heures de Turin et de Milan’, in Musées royaux des Beaux-Arts de Belgique. Bulletin. Koninklijke Musea voor Schone Kunsten van België, IV, Miscellanea Erwin Panofsky, 1955, pp. 7-20. McCartney 1923 Eugene S. McCartney, ‘On Fossil Thunderstones and Fingerstones’, in The Classical Journal, 18, 7, April 1923, pp. 425-426. McKinnon 1978 James W. McKinnon, ‘Representations of the Mass in Medieval and Renaissance Art’, in Journal of the Mass in Medieval and Renaissance Art, 31, 1, Spring 1978, pp. 21-52. McLachlan 1978 Elizabeth Parker McLachlan, ‘The Bury Missal in Laon and Its Crucifixion Miniature’, in Gesta, 17, 1, 1978, pp 27-35. Macleod Higgins 1997 Iain Macleod Higgins, Writing East: the ‘travels’ of Sir John Mandeville (The Middle Ages Series), Philadelphia, PA, 1997. Madou 1986 Mireille Madou, Le costume civil (Typologie des sources du Moyen Âge occidental, 47), Turnhout, 1986. Madrid/Valencia 2001 Mauro Natale (ed), El Renacimiento Mediterránneo. Viajes de artistas e itinerarios de obras entre Italia, Francia y España en el siglo XV, exh. cat., Madrid (Museo Thyssen-Bornemisza)/Valencia (Museo de Belles Artes de Valencia), 2001. Mahler 1905 Arthur Mahler, ‘L’original de la Vénus des Médicis’, in Comptes rendus des séances de l’Académie des Inscriptions et Belles-Lettres, 1905, pp. 623628. Mander 1994-1999 Hessel Miedema (ed.), Karel van Mander: The Lives of the Illustrious Netherlandish and German Painters: From the First Edition of the ‘Schilderboeck’ (1603-1604), 5 vols, Doornspijk, 1994-1999.

Lu et al. 1999 Rong Lu, Jan Koenderink and Astrid Kappers, ‘Specularities on Surfaces with Tangential Hairs and Grooves’, in Proceedings of the Seventh IEEE International Conference on Computer Vision, Los Alamitos, 1999, pp. 2-7.

Manning 1999 Christopher D. Manning and Hinrich Schütze, Foundations of Statistical Natural Language Processing, Cambridge, 1999.

Luber 1998 Katherine C. Luber, ‘Jan van Eyck, the Adornes Family, and Two Paintings of “Saint Francis in Portraiture”’, in Philadelphia Museum of Art Bulletin, 91, 386/387, Spring 1998, pp. 24-37.

Mantovani 2000 Thierry Mantovani, ‘Pierre Fabri et la poétique des Puys dans le second livre du Grand et vrai art de pleine rhétorique’, in Nouvelle Revue du XVIe Siècle, 18, no. 1, Arts poétiques de la Renaissance, 2000, pp. 41-54.

Lugt 1968 Frits Lugt, Musée du Louvre. Inventaire général des dessins des Écoles du Nord, Paris, 1968.

Marconi 1965 Bohdan Marconi, ‘The Transfer of Panel Paintings on Linen by Siderov (Hermitage Museum, St. Petersburg) in the Nineteenth Century’, in Applications of Science in Examination of Works of Art: Proceedings of the Seminar: September 7-16, 1965, Research Laboratory, Museum of Fine Art, Boston, 1965, pp. 246-54.

Lukiesh 1922 Matthew Luckiesh, in Visual Illusions, 1922.

579 Mareel 2010 Samuel Mareel, Voor vorst en stad. Rederijkersliteratuur en vorstenfeest in Vlaanderen en Brabant (1432-1561), Amsterdam, 2010.

Martin 1967 Paul Martin, Armes et armures de Charlemagne à Louis XIV, Paris, 1967.

Marijnissen 1969 Roger Marijnissen, ‘Notes sur l’enluminure eyckienne: “La naissance de Saint Jean Baptiste” des Heures de Milan-Turin’, in Scriptorium, 23, 1969, pp. 225-226.

Martin 2000 John Martin, ‘Inventing Sincerity, Refashioning Prudence: The Discovery of the Individual in Renaissance Europe’, in American Historical Review, 102 (5), December 1997, pp. 1309-1317, 1320-1342; reprinted in Keith Whitlock (ed.),The Renaissance in Europe. A Reader, New Haven/London, 2000, pp. 11-31.

Marijnissen, Grosemans 1982-1983 Roger H. Marijnissen and Jozef Grosemans, ‘Twee specifieke paneelproblemen. De Johannes de Doper van het Lam Gods and Rubens’ Kruisoprichting’, in Bulletin de l’Institut royal du Patrimoine artistique/ Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 29, 19821983, pp. 120-132.

Masschelein-Kleiner 2000 Liliane Masschelein-Kleiner, ‘Les cinquante ans de l’IRPA’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 27, 1996-1998, pp. 14-47.

Marinis 1994 Fabrizio de Marinis, Velvet. History, Techniques, Fashions, Milan, 1994.

Masschelein-Kleiner 2005 Lilliane Masschelein-Kleiner, ‘La technique des “Primitifs Flamands”‘, in Studies in Conservation, 50, 2005, pp. 67-68.

Marks 2000 Laura U. Marks, The Skin of Film: Intercultural Cinema, Embodiment, and the Senses, Durham, 2000.

Mather 1906 Frank J. Mather, Jr., ‘Recent Additions to the Collection of Mr. John G. Johnson, Philadelphia’, in The Burlington Magazine, 9, 1906, pp. 358359.

Marrow 1979 James H. Marrow, ‘Passion Iconography in Northern European Art of the Late Middle Ages and Early Renaissance. A Study of the Transformation of Sacred Metaphor into Descriptive Narrative’, in Ars Neerlandica. Studies in the History of Art in the Low Countries, 1, Kortrijk, 1979.

Matthäi 1835 Friedrich Matthäi, Verzeichnis der Königlich Sächsischen Gemälde-Galerie zu Dresden. Erste Hauptabtheilung. Die Gemälde der äusseren Galerie; Zweite Hauptabtheilung. Die Gemälde der inneren Galerie, der Abteilung H. der äusseren Galerie und des Pastellcabinets, 2 vols, Dresden, 1835.

Marrow 1986 James Marrow, ‘Symbol and Meaning in Northern European Art of the Late Middle Ages and the Early Renaissance’, in Simiolus, 16, 2/3, 1986, pp. 150-169. Marrow 1991 Janes H. Marrow, ‘Dutch Manuscripts in Context: Encounters with the Art of France, the Southern Netherlands and Germany’, in Koert van der Horst and Johann-Christian Klamt (eds), Masters and Miniatures. Proceedings of the Congress on Medieval Manuscript Illumination in the Northern Netherlands (Utrecht, 10-13 December 1989), Doornspijk, 1991, pp. 53-88. Marrow 2002 James H. Marrow, ‘Une page inconnue des Heures de Turin’, in Revue de l’Art, 135, 2002, pp. 67-76. Martens 1929 Bela Martens, Meister Francke, 2 vols, Hamburg, 1929. Martens 1999 Maximiliaan P. J. Martens, ‘The Position of the Artist in the Fifteenth Century: Salaries and Social Mobility’, in Wim Blockmans and Antheun Janse (eds), Sharing Status, Turnhout, 1999, pp. 387-414. Martens 2011 Didier Martens, ‘Compte rendu de Pierre Colman, Jan van Eyck et Jean sans Pitié (Académie royale de Belgique. Mémoire de la Classe des Arts. Collection in-8°, 3e série, t. XXVII), Brussels, 2009, pp. 73-78, 130-131’, in Bulletin d’information de la Société royale d’archéologie de Bruxelles, 64, 2011, pp. 2-8. Martens 2012 Didier Martens, ‘Compte rendu de Pierre Colman, Jan van Eyck et Jean sans Pitié (Académie royale de Belgique. Mémoire de la Classe des Arts. Collection in-8°, 3e série, t. XXVII), Brussels, 2009, p. 73-78 et 130131’, in Revue belge d’archéologie et d’histoire de l’art, 1, 2012, pp. 171-172.

Matthäi 1846 Friedrich Matthäi [revised under the editorship of Julius Schnorr von Carolsfeld], Catalog der Königlichen Gemälde-Galerie zu Dresden, Dresden, 1846. Mayer 1926 August L. Mayer, ‘A Jan van Eyck Problem’, in The Burlington Magazine, 48, 1926, p. 200. Meckelnborg 2014 Christina Meckelnborg, ‘Die Inschrift des Genter Altars. Eine Philologische Betrachtung’, in Stephan Kemperdick and Johannes Rößler (eds), Der Genter Altar der Brüder Van Eyck, exh. cat., Berlin (Staatliche Museen zu Berlin), 2014, pp. 113-121. Meine 1983 Karl-Heinz Meine, ‘Zur Weltkarte des Andreas Walsperger, Konstanz 1448’, in Wolfgang Scharfe (ed.), Kartenhistorisches Colloquium Bayreuth 1982, Berlin, 1983, pp. 17-30. Meiss 1967 Millard Meiss, French Painting in the Time of Jean de Berry. The Late Fourteenth Century and the Patronage of the Duke, New York/London, 1967. Meiss 1968 Millard Meiss, French Painting in the Time of Jean de Berry. The Boucicaut Master, London, 1968. Meiss 1974 Millard Meiss, French Painting in the Time of Jean de Berry. The Limbourgs and their contemporaries, New York, 1974. Melion 1991 Walter S. Melion, Shaping the Netherlandish Canon. Karel van Mander’s Schilder-Boeck, Chicago, 1991. Mellinkoff 1993 Ruth Mellinkoff, Outcasts: Signs of Otherness in Northern European Art of the Late Middle Ages, Berkeley/Los Angeles/Oxford, 1993.

580 Mély 1921 Fernand de Mély, ‘Le retable de l’Agneau des Van Eyck et les pierres gravées talismaniques’, in Revue archéologique, 1921, pp. 1-16. Mendelsohn 1982 Leatrice Mendelsohn, Paragoni: Benedetto Varchi’s ‘Due lezzioni’ and Cinquecento Art Theory, Ann Arbor, 1982. Mérimée 1830 Jean-François-Léonor Mérimée, De la peinture à l’huile, ou des procédés matériels émployés dans ce genre de peinture, depuis Hubert et Jean van Eyck jusqu’à nos jours, Paris, 1830. Merrifield 1849 Mary Philadelphia Merrifield, Original Treatises, Dating from the XIIth to the XVIIIth Centuries on the Arts of Painting, in Oil, Miniature, Mosaic and on Glass; of Gilding, Dyeing, and the Preparation of Colours and Artificial Gems, 2 vols, London, 1849. Messling 2012 Guido Messling, ‘Medewerker Jan van Eyck, Kruisiging van Christus, ca. 1440’, in Stephan Kemperdick and Friso Lammertse (eds), De Weg naar Van Eyck, exh. cat., Rotterdam (Museum Boijmans Van Beuningen), 2012, pp. 302-304. Michaelis, Martin Vogtherr 2011 Rainer Michaelis and Christoph Martin Vogtherr, ‘Die erste Anordnung der Gemäldegalerie im Alten Museum 1830’, in Kunst als Kulturgut, II, Munich, 2011, pp. 227-247. Michelant 1871 Henri Michelant, ‘Inventaire des vaisselles, joyaux, tapisseries, peintures, manuscrits etc. de Marguerite d’Autriche, régente et gouvernante des Pays-Bas, dressé en son palais de Malines, le 9 juillet 1523’, in Académie Royale des Sciences, des Lettres et des Beaux-Arts de Belgique. Compte-rendu des séances de la Commission royale d’histoire, ou Recueil de ses Bulletins, 3iéme sér., 12, 171, pp. 5-78, 83-136. Miller 1986 Naomi Miller, ‘Paradise Regained: Medieval Garden Fountains’, in Elisabeth B. MacDougall (ed.), Medieval Gardens, Washington, 1986, pp. 135-154. Mills, White 1994 John S. Mills and Raymond White, The Organic Chemistry of Museum Objects, 2nd edn, Burlington, 1994. Mills, White 1985 John Mills and Raymond White, ‘Analyses of Paint Media’, in National Gallery Technical Bulletin, 9, 1985, pp.70-71. Mira 2007 Eduard Mira, ‘Siti Perillos. El Toisón de Oro y la gran fisura de Europa’, in Eduard Mira and An Delva (eds), A la búsqueda del Toisón de Oro. La Europa de los Príncipes. La Europa de las ciudades, vol. 1, exh. cat., Almudín (Museo de la ciudad), 2007, pp. 257-285. Mocholí Roselló 2009 Asunción Mocholí Roselló, Estudi dels documents del pintors i altres artífexs valencians segles XIII, XIV i XV (1238-1450), PhD diss., Universitat Politècnica de Valencia, 2009. Monnas 1986 Lisa Monnas, ‘Developments in figured velvet weaving in Italy during the fourteenth century’, in Bulletin de Liaison du Centre International d’Étude des Textiles Anciens, 63/ 64, 1986, pp. 63-100.

Monnas 2000 Lisa Monnas, ‘Silk textiles in the Paintings of Jan van Eyck’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 147-162. Monnas 2008 Lisa Monnas, Merchants, Princes and Painters. Silk Fabrics in Italian and Northern Paintings 1300-1550, New Haven/London, 2008. Monstrelet 1861 Enguerrand de Monstrelet, La Chronique d’Enguerrand de Monstrelet, Louis Claude Douët-d’Arcq (ed.), 5, Paris, 1861. Montero Tortajada 2013a Maria de la Encarnación Montero Tortajada, La transmisión del conocimiento en los oficios artísticos. Valencia, 1370-1450, PhD diss., Universitat de Valencia, Valencia, 2013. Montero Tortajada 2013b Maria de la Encarnación Montero Tortajada, ‘The Oligarch and the Brushes: a Biographical Sketch of Andreu Garcia, priest/El oligarca y los pinceles: breve semblanza del presbítero Andreu Garcia’, in Espacio, tiempo y forma, 1, 2013, pp. 25-41, 401-418. Montero Tortajada 2015 Maria de la Encarnación Montero Tortajada, La transmisión del conocimiento en los oficios artísticos. Valencia, 1370-1450, Valencia, 2015. Montgomery 1996 Scott L. Montgomery, ‘The Eye and the Rock: Art, Observation and the Naturalistic Drawing of Earth Strata’, in Earth Sciences History, 15, 1996, pp. 3-24. Mortier 1894 Etienne Mortier, ‘Fouilles effectuées dans la crypte de la cathédrale St-Bavon’, in Bulletijn van de Geschied- en Oudheidkundige kring van Gent, 1, 1894, pp. 239-246. Moser 1999 Stephanie Moser, Ancestral Images. The Iconography of Human Origins, Ithaca, 1998. Müller 1906 Walter A. Müller, Nacktheit und Enblößung in der altorientalischen und älteren griechischen Kunst, Borna/Leipzig, 1906. Müller-Hofstede 1998 Justus Müller-Hofstede, ‘Der Künstler im Humilitas gestus. Altniederländische Selbstporträts und ihre Signifikanz in Bildkontext. Jan van Eyck - Dieric Bouts - Hans Memling - Joos van Cleve’, in Gunther Schweikhart (ed.), Autobiographie und Selbstportrait in der Renaissance, Cologne, 1998, pp. 38-68. Nagel, Wood 2005 Alexander Nagel and Christopher S. Wood, ‘Interventions: Toward a New Model of Renaissance Anachronism’, in Art Bulletin, LXXXVII, 3, September 2005, pp. 403-415. Nash 2008 Susie Nash, Northern Renaissance Art, (Oxford History of Art), Oxford, 2008. Nash 2010 Susie Nash, ‘“Pour couleurs et autres prise de Lui…”: The Supply, Acquisition, Cost and Employment of Painters’ Materials at the Burgundian Court’, in Jo Kirby, Susie Nash and Joanna Cannon (eds), Trade in Artists’ Materials. Markets and Commerce in Europe to 1700, London, 2010, pp. 97-185.

581 Neeffs 1876 Emmanuel Neeffs, Histoire de la peinture et de la sculpture à Malines, 1, Ghent, 1876. Neidhardt, Schölzel 2000 Uta Neidhardt and Christopher Schölzel, ‘Jan van Eyck’s Dresden Triptych’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 25-39. Neidhardt, Schölzel 2005 Uta Neidhardt and Christoph Schölzel, ‘Jan van Eycks Dresdener Marientriptychon. Entstehung und Funktion’, in Thomas Ketelsen and Uta Neidhardt (eds), Das Geheimnis des Jan van Eyck: die frühen niederländischen Zeichnungen und Gemälde in Dresden, exh. cat., Dresden (Staatliche Kunstsammlungen), 2005, pp. 14-21. New York 1980 The Wild Man: Medieval Myth and Symbolism, exh. cat., New York (The Metropolitan Museum of Art), 1980. New York 1994 Maryan W. Ainsworth (ed.), Petrus Christus Renaissance Master of Bruges, exh. cat., New York (The Metropolitan Museum of Art), 1994. New York 2005 Prague, The Crown of Bohemia 1347-1437, exh. cat., New York (The Metropolitan Museum of Art), 2005.

Oltrogge 2008 Doris Oltrogge, ‘Jan van Eyck, Lucca-Madonna, Auswertung der Vis-messungen vom 28.–30.11.07’, Cologne Institute of Conservation Sciences/Institut für Restaurierungs- und Konservierungswissenschaft Fakultät für Kulturwissenschaften der Fachhochschule Köln, 2008, unpublished report. Oosterwijk 2005 Sophie Oosterwijk, ‘Food for Worms – Food for Thought. The Appearance and Interpretation of the “Verminous” Cadaver in Britain and Europe’, in Church Monuments, 20, 2005, pp. 40-80, 133-140. Oudemans 1852 A.C. Oudemans (ed.), Die bediedenisse van der missen, Leiden, 1852. Cd-Rom Middelnederlands. Pächt 1956 Otto Pächt, ‘Panofsky’s “Early Netherlandish Painting”‘, in The Burlington Magazine, 98, 1956, no. 637, pp. 110-116; no. 641, pp. 267279. Pächt 1989 Otto Pächt, Van Eyck. Die Begründer der altniederländischen Malerei, Munich, 1989. Pächt 1994 Otto Pächt, Van Eyck and The Founders of Early Netherlandish Painting, London, 1994.

New York 2010-2011 Maryan W. Ainsworth (ed.), Man, Myth, and Sensual Pleasures. Jan Gossart’s Renaissance: The Complete Works, exh. cat., New York (The Metropolitan Museum of Art), 2010-2011.

Pächt 1999 Otto Pächt, Van Eyck and the Founders of Early Netherlandish Painting, London, 1999.

Nijmegen 2005 De Gebroeders van Limburg, exh. cat., Nijmegen (Museum Het Valkhof), 2005.

Panhans 1974 Günter Panhans, ‘Florentine Maler verarbeiten ein eyckisches Bild’, in Wiener Jahrbuch für Kunstgeschichte, 27, 1974, pp. 188-198.

Noad 2002 Tim Noad, Fifteen Worthies, ten Virtues and Saint George. An important Cycle of Pre-Eyckian Wall Paintings of the Fourteenth Century in Bruges, unpublished Master’s thesis, Courtauld Institute, London, 2002. Nordström 1967 Carl-Otto Nordström, The Duke of Alba’s Castilian Bible. A Study of the Rabbinical Features of the Miniatures (Acta Universitatis Upsaliensis. Figura Nova Series, 5), Uppsala, 1967. Notebaert 1939 Emile Notebaert, ‘Contribution à l’identification du paysage architectural dans les œuvres des Primitifs flamands’, in Revue de Saint-Luc, 1939, pp. 2-13. Nys 2004 Ludovic Nys, ‘L’art à la cour de Flandre à l’époque du mariage de Philippe le Hardi et Marguerite de Male’, in L’art à la cour de Bourgogne, exh. cat., Paris, 2004, pp. 52-64. Nys, Lievois 2002 Ludovic Nys and Daniel Lievois, ‘Not Tymotheus again! The Portrait of Goedevaert de Wilde, receiver of Flanders and Artois’, in Bert Cardon, Jan van der Stockt and Dominique Vanwijnsberghe (eds), Als ich xan. Liber Amicorum in memory of Maurits Smeyers, vol. II, Leuven, 2002, pp. 1037-1057. Olszewski 1997 E.J. Olszewski, Armenini’s treatise on painting, PhD diss., University of Minnesota, 1997.

Panofsky 1927 Erwin Panofsky, ‘Imago Pietatis: Ein Beitrag zur Typengeschichte des “Schmerzensmannes” und der Media Mediatrix’, in Festschrift fur Max J. Friedlander zum 60. Geburtstage, Leipzig, 1927, pp. 261-308. Panofsky 1935 Erwin Panofsky, ‘The Friedsam Annunciation and the Problem of the Ghent Altarpiece’, in The Art Bulletin, 17, 1935, pp. 433-473. Panofsky 1938 Erwin Panofsky, ‘Once more the “Friedsam Annunciation and the Problem of the Ghent Altarpiece”‘, in The Art Bulletin, 20, 1938, pp. 419-442. Panofsky 1953 Erwin Panofsky, Early Netherlandish Painting. Its Origins and Character, 2 vols, Cambridge, 1953. Panofsky 1971 Erwin Panofsky, Early Netherlandish Painting: its origins and character, 2 vols, New York, 1971. Panofsky 2006 Erwin Panofsky, Korrespondenz 1950-1956, Wiesbaden, 2006 (Erwin Panofsky, Korrespondenz 1910 bis 1968. Eine kommentierte Auswahl in fünf Bänden, Dieter Wuttke (ed.), 3). Panofsky 2008 Erwin Panofsky, Korrespondenz 1957-1961, Wiesbaden, 2008 (Erwin Panofsky, Korrespondenz 1910 bis 1968. Eine kommentierte Auswahl in fünf Bänden, Dieter Wuttke (ed.), 4).

582 Panofsky 2011 Erwin Panofsky, Korrespondenz 1962-1968, Wiesbaden, 2011 (Erwin Panofsky, Korrespondenz 1910 bis 1968. Eine kommentierte Auswahl in fünf Bänden, Dieter Wuttke (ed.), 5).

Paviot 2006 Jacques Paviot and Dina Goren, ‘Les inscriptions grecques et hébraïques dans les tableaux eyckiens’, in Revue belge d’archéologie et d’histoire de l’art, 75, 2006, pp. 53-73.

Paris 2004 Paris 1400. Les arts sous Charles VI, exh. cat., Paris (Musée du Louvre), 2004.

Paviot 2010 Jacques Paviot, ‘La Crucifixion et le Jugement dernier attribués à Jan van Eyck (Diptyque de New York)’, in Zeitschrift für Kustgeschichte, 2, 2010, pp. 145-68.

Parma 1999 Elena Parma, Italy and the Low Countries. Artistic Relations, The Fifteenth Century, Proceedings of the Symposium held at Museum Catharijneconvent, Utrecht, 14 March 1994, Florence, 1999, pp. 79-96. Parma 2002 Elena Parma, ‘Genoa – Gateway to the South’, in Till-Holger Borchert (ed.), The Age of Van Eyck 1430-1530, The Mediterranean World and Early Netherlandish Painting, exh. cat., Bruges (Groeningenmuseum), 2002, pp. 94-107. Parshall 1993 Peter Parshall, ‘Imago contrafacta: Images and Facts in the Northern Renaissance’, in Art History, 16, 1993, pp. 554-579. Passavant 1841 Johann David Passavant, ‘Beiträge zur Kenntniss der altniederländischen Malerschulen des 15. Und 16. Jahrhunderts’, in Kunstblatt, 3, 12 January 1841, p. 9. Paulus Burgensis (1435) Paulus Burgensis, ‘Scrutinium scripturarum Pauli Burgensis contra perfidiam Iudaeorum’, in Biblia sacra cum glossa ordinaria, 6, Antwerp, 1634. Pauwels 1985 Henri Pauwels, ‘Van Eycks Madonna van der Paele opnieuw bekeken’, in V. Vermeersch (ed.), Jaarboek 1983-84. Stad Brugge Stedelijke Musea, 2, Bruges, 1985, pp. 223-240. Paviot 1990a Jacques Paviot, ‘Oliviero Maruffo et la cour de Bourgogne’, in La Storia dei Genovesi, 10, Atti del Convegno […] 23-26 Maggio 1989, [Genova, 1990], pp. 369-393. Paviot 1990b Jacques Paviot, ‘La vie de Jan van Eyck selon les documents écrits’, in Revue des Archéologues et Historiens d’Art de Louvain, XXIII, 1990, pp. 83-93. Paviot 1991 Jacques Paviot, ‘La Mappemonde attribuée à Jan van Eyck par Fàcio: une pièce à retirer du catalogue de son œuvre’, in Revue des Archéologues et Historiens de l’Art de Louvain, 24, 1991, pp. 57-62.

Paviot 2013 Jacques Paviot, ‘Review of: Albert Châtelet, Hubert et Jan van Eyck, créateurs de l’Agneau Mystique, Dijon, 2011’, in Francia-Recensio, 2013-1, Mittelalter - Moyen Âge (500-1500), [first published: http:// www.recensio.net/rezensionen/zeitschriften/francia-recensio/2013-1/ mittelalter-moyen-age-500-1500/hubert-et-jan-van-eyck-createursde-l2019agneau-mystique/@@generate-pdf-recension?language=de (accessed 25 September 2015)]. Paviot, Goren 2006 Jacques Paviot, with Dina Goren, ‘Les inscriptions grecques et hébraïques dans les tableaux eyckiens’, in Revue belge d’archéologie et d’histoire de l’art, 75, 2006, pp. 53-73. Pemán y Pemartín 1969 Cesar Pemán y Pemartín, Juan van Eyck y España, Cádiz, 1969. Pentcheva 2006 Bissera Pentcheva, ‘The Performative Icon’, in The Art Bulletin, 88, 4, December 2006, pp. 631-655. Périer D’Ieteren 1982-1983 Catheline Périer D’Ieteren, ‘Dessin au poncif et dessin perforé. Leur utilisation dans les anciens Pays-Bas au XVe siècle’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 19, 1982/83, pp. 74-94. Périer-D’Ieteren 1985 Catheline Périer-D’Ieteren, Colyn de Coter et la technique picturale des peintres flamands du xve siècle, Brussels, 1985. Périer-D’Ieteren 1990 Catheline Périer-D’Ieteren (ed.), Paul Philippot. Pénétrer l’art. Restaurer l’oeuvre. Une vision humaniste. Hommage en forme de florilège, Kortrijk, 1990. Périer-D’Ieteren 2005 Catheline Périer-D’Ieteren, Thierry Bouts: L’œuvre complet, Brussels, 2005. Périer-D’Ieteren 2006 Catheline Périer-D’Ieteren, Dieric Bouts: the Complete Works, Brussels, 2006.

Paviot 1995a Jacques Paviot, ‘The Sitter for Jan van Eyck’s “Léal Souvenir”’, in Journal of the Warburg and Courtauld Institutes, 58, 1995, pp. 210215.

Petti Balbi 1996 Giovanna Petti Balbi, Mercanti e Nationes nelle Fiandre: i Genovesi in età bassomedievale, Pisa, 1996.

Paviot 1995b Jacques Paviot (ed.), Portugal et Bourgogne au XVe siècle. Recueil de documents extraits des archives bourguignonnes (1384-1482), Lisbon/ Paris, 1995.

Petti Balbi 1997 Giovanna Petti Balbi, ‘Le strategie mercantili di una grande casata genovese: Francesco Spinola tra Bruges e Malaga (1420-1456)’, in Serta Antiqua et Mediaevalia, I, 1997, pp. 379-393.

Paviot 2003 Jacques Paviot, Les ducs de Bourgogne, la croisade et l’Orient (fin XIVeXVe siècle), Paris, 2003.

Petti Balbi 2005 Giovanna Petti Balbi, Nogoziare fuori patria. Nazioni e Genovesi in età medievale, Bologna, 2005.

583 Pfeiffer 1861 Franz Pfeiffer, Das Buch der Natur: Die erste Naturgeschichte in deutsche Sprachedurch Conrad von Megenberg, Stuttgart, 1861.

Ploss 1962 Emil E. Ploss, Ein Buch von alten Farben, Heidelberg/Berlin, 1962, pp. 113-114.

Philip 1971 Lotte B. Philip, The Ghent Altarpiece and the Art of Jan van Eyck, Princeton, 1971.

Pollard 1900 William Pollard (ed.), The Travels of Sir John Mandeville: the Version of the Cotton Manuscript in Modern Spelling, London, 1900.

Philippe 1960 Joseph Philippe, Van Eyck et la genèse mosane de la peinture des anciens Pays-Bas, Liège, 1960.

Post 1921 Paul Post, ‘Wen stellen die vier ersten Reiter auf dem Flügel der Gerechten Richter am Ganter Altar’, in Jahrbuch der preussischen Kunstsammlungen, 42, 1921, pp. 67-81.

Philippot 1960 Paul Philippot, ‘Réflexions sur le problème de la formation des restaurateurs de peintures et de sculptures’, in Studies in Conservation, 5, 1960, pp. 61-69. Philippot 1967 Paul Philippot, ‘Le problème des relations entre historiens d’art et restaurateurs’, in Bulletin du Congrès International d’Histoire de l’Art, II, 1967, pp. 8-9 (reprinted in Périer-D’Ieteren 1990, pp. 505-507). Philippot 2008 Paul Philippot, ‘De l’histoire de la restauration comme affaire de famille’, in Dominique Vanwijnsberghe (ed.), Autour de la Madeleine Renders. Un aspect de l’histoire des collections, de la restauration et de la contrefaçon en Belgique dans la première moitié du XXe siècle (Scientia Artis, 4), Brussels, 2008, pp. 189-196.

Post 1922 Paul Post, ‘Die Stifter des Lebensbrunnens der van Eyck’, in Jahrbuch der preussischen Kunstsammlungen, 43, 1922, pp. 120-125. Power 1910 D’Arcy Power, Treatises of Fistula in Ano, Haemorrhoids, and Clysters by John Arderne from an Early Fifteenth-Century Manuscript Translation, London/Oxford, 1910. Preimesberger 2011 Rudolf Preimesberger, Paragons and Paragone. Van Eyck. Raphael. Michelangelo. Caravaggio. Bernini, Los Angeles, 2011.

Pietrangella 2004 Daniela Pietrangella (ed.), Bartolomeo Facio, Rerum gestarum Alfonsi regis libri, Alessandria, 2001.

Prevenier 1974 Walter Prevenier, ‘Officials in Town and Countryside in the Low Countries. Social and Professional Developments from the Fourteenth to the Sixteenth-Century’, in Acta Historiae Neerlandicae. Studies on the History of the Netherlands, 7, 1974, pp. 1-17.

Pinchart 1860 Alexander Pinchart (ed.), Archives des arts, sciences et lettres: Documents inédits, vol. I, Ghent, 1860.

Prevenier, Blockmans 1983 Walter Prevenier and Wim Blockmans, De Bourgondische Nederlanden, Antwerp, 1983.

Piponnier, Mame 1995 Françoise Piponnier and Perrine Mame, Se vêtir au Moyen Age, Paris, 1995.

Purtle 2003 Carol J. Purtle, ‘Jan van Eyck’s Madonna in a Church: re-viewing Stylistic Assumptions’, in Hélène Verougstraete and Roger Van Schoute (eds), Jérôme Bosch et son entourage et autres études. Le dessin sousjacent et la technologie dans la peinture, colloque XIV, Bruges, 13-15 septembre 2001, Leuven/Paris/Dudley, MA, 2003, pp. 250-259.

Plahter 1997 Unn Plahter, ‘Norwegian Medieval Oil Paintings with Special Attention to the Heddal Frontal’, in Das Aschaffenburger Tafelbild: Studien zur Tafelmalerei des 13. Jahrhunderts, Arbeitshefte des Bayerischen Landesamtes für Denkmalpflege, 89, Munich, 1997, pp. 335-347. Pleij 1999 Herman Pleij, ‘Gent en de stadscultuur in de Nederlanden’, in Hugo Soly and Johan Van de Wiele (eds), Carolus. Keizer Karel V 1500-1558, exh. cat., Ghent (Kunsthal De Sint-Pietersabdij), 1999, pp. 123-131.

Raft 1982 Adam Raft, ‘Die Quellentexte zur Verwendung besonderer Bindemittel fur blaue Farben’, in Maltechnik/Restauro, 88, 1982, pp. 112-119. Randall 1997 Lilian Randall, Medieval and Renaissance Manuscripts in the Walters Art Gallery, Volume III, Belgium 1250-1530, 2 vols, Baltimore, 1997.

Plesters 1956 Joyce Plesters, ‘Cross-Sections and Chemical Analysis of Paint Samples’, in Studies in Conservation, 2(3), 1956, pp. 110-157.

Ranferling 1867 M. Ranferling, Geschichte der Juden in Portugal, Leipzig, 1867.

Plesters 1980 Joyce Plesters, ‘Possible Causes of Blanching Involving Changes in Pigments or Interaction of Pigment and Medium’, in National Gallery Technical Bulletin, 4, 1980, pp. 61-63.

Raspe 1781 Rudolph Erich Raspe, A critical essay on oil-painting: proving that the art of painting in oil was known before the pretended discovery of John and Hubert van Eyck, London, 1781.

Plesters 1993 Joyce Plesters, ‘Ultramarine Blue, Natural and Artificial’, in Ashok Roy (ed.), Artists’ Pigments. A Handbook of Their History and Characteristics, vol. 2, Washington, 1993, p. 45 and p. 58.

Raymundus Martinus (c. 1280) Raymundus Martinus (c. 1280), Pugio Fidei adversus Mauros et Judaeos, Leipzig, 1687; reprint Westmead, 1967.

Pliny the Elder 1855 Pliny the Elder, The Natural History, Book II, J. Bostock and H.T. Riley (eds), London, 1855.

Redonensis 1893 Marbodus Redonensis episcopus, Liber Lapidum, Patrologiae cursus completus: Series Latina, Jacques Paul Migne (ed.), Paris, 1893, [PL 171] 1737-1770.

584 Reiche et al. 2004 Ina Reiche et al., ‘Spatially Resolved Synchrotron-Induced X-ray Fluorescence Analyses of Metalpoint Drawings and their Mysterious Inscriptions’, in Spectrochimica Acta, Part B, 59, 2004, pp. 1657-1662. Reiche et al. 2005 Ina Reiche, Silke Merchel, Thomas Ketelsen and Olaf Simon, ‘Alc ixh xan. Zum zeichnerischen Kalkül Jan van Eycks’, in Thomas Ketelsen and Uta Neidhardt (eds), Das Geheimnis des Jan van Eyck. Die frühen niederländischen Zeichnungen und Gemälde in Dresden, exh. cat., Dresden (Staatliche Kunstammlungen), 2005. Reiset 1866 Frédéric Reiset, Notice des dessins, cartons, pastels, miniatures et émaux exposés dans les salles du 1er étage au Musée imperial du Louvre. Première partie: écoles d’Italie, écoles allemande, flamande et hollandaise; précédé d’une introduction historique et du résumé de l’inventaire général des dessins, Paris, 1866, reprinted 1878 and 1887. Remo 1923 Remo (Norbertijn van Tongerloo), ‘Middeleeuwsche Handelsbetrekkingen tusschen Vlaanderen en Genua (Ligurie)’, in Biekorf, XXIX, 1923, pp. 145-166, 178-186, 227-234. Reynaud 1989 Nicole Reynaud, ‘Barthélemy d’Eyck avant 1450’, in Revue de l’Art, 84, 1989, pp. 22-43. Reynolds 2000 Catherine Reynolds, ‘The King of Painters’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 1-16. Reynolds 2012 Catherine Reynolds, ‘Self Portrait and Signature in the Brussels “Justice Scenes”: Rogier van der Weyden’s Fame and the Status of Painting’, in Lorne Campbell, Jan van der Stock, Catherine Reynolds and Lieve Watteeuw (eds), Rogier van der Weyden in Context. Papers presented at the Seventeenth Symposium for the Study of Underdrawing and Technology in Painting held in Leuven, 22-24 October 2009, Leuven, 2012, pp. 79-91. Ricciardi et al. 2012 Paola Ricciardi et al., ‘Near Infrared Reflectance Imaging Spectroscopy to Map Paint Binders In Situ on Illuminated Manuscripts’, in Angewandte Chemie, 124, 2012, pp. 5705-5708. Ricketts 1906 Charles Ricketts, ‘Pictures in the Collection of Mr. John G. Johnson, of Philadelphia’, in The Burlington Magazine, 9, 1906, p. 426. Ridderbos 1998 Bernhard Ridderbos, ‘The Man of Sorrows. Pictorial Images and Metaphorical Statements’, in Alasdair MacDonald, Bernhard Ridderbos and Rita Schlusemann (eds), The Broken Body, Passion Devotion in LateMedieval Culture, Groningen, 1998, pp. 143-182. Ridderbos 2005 Bernhard Ridderbos, Anne Van Buren and Henk Van Veen (eds), Early Netherlandish Paintings: Rediscovery, Reception, and Research, Amsterdam, 2005. Ridderbos 2008 Bernhard Ridderbos, ‘Creating Frameworks. The Social Function of the Ghent Altarpiece’, in Herman W. Hoen and Mary G. Kemperink (eds), Vision in Text and Image. The Cultural Turn in the Study of Arts (Groningen Studies in Cultural Change, 30), Leuven/Paris/Dudley, 2008, pp. 33-52.

Ridderbos 2014 Bernhard Ridderbos, Schilderkunst in de Bourgondische Nederlanden, Leuven, 2014. Ringbom 1965 Sixten Ringbom, Icon to Narrative. The Rise of the Dramatic Close-Up in Fifteenth-Century Devotional Painting, Abo, 1965. Rishel 1997 Joseph J. Rishel, ‘The Philadelphia and Turin Paintings: The Literature and Controversy over Attribution’, in Jane Watkins (ed.), Jan van Eyck: Two Paintings of Saint Francis Receiving the Stigmata, Philadelphia, 1997, pp. 3-12. Roberts-Jones 1992 François Roberts-Jones, De chef-d’oeuvre à la copie. Des frères Van Eyck à Michiel Coxcie, Brussels, 1992. Roggen 1938 Domien Roggen, ‘Gentsche beeldhouwkunst der XIVde en XVde eeuw’, in Gentsche Bijdragen tot de Kunstgeschiedenis, 5, 1938, pp. 51-73. Romano 2004 Serena Romano, ‘Giusto di Ravensburg e i pittori svizzero-tedeschi a Santa Maria di Castello’, in Piero Boccardo and Clario Di Fabio (eds), Genova e l’Europa continentale. Opere, artisti, committenti, collezionisti. Austria, Germania, Svizzera, Milan, 2004, pp. 32-47. Rooses 1902 Max Rooses, ‘De tekeningen der vlaamsche meesters’, in Onze kunst, 1, 1902, pp. 3-8. Rösener 2000 Werner Rösener, ‘Codex Falkensteinensis. Zur Erinnerungskultur eines Adelsgeschlechts im Hochmittelalter’, in Werner Rösener, Adelige und bürgerliche Erinnerungskultur des Spätmittelalters und der Frühen Neuzeit (Formen der erinnerung, 8), Göttingen, 2000, pp. 35-55. Rotelle 2002 John E. Rotelle (ed.), The Works of Saint Augustine. On Genesis: A Refutation of the Manichees Unfinished Literal Commentary on Genesis, New York, 2002. Rotterdam 2002 Benoît Boëlens van Waesberghe (ed.), European Master Drawings Unveiled. Van der Goes, Michelangelo, Van Goyen, Fragonard and other Masters from Belgian collections, exh. cat., Rotterdam (Kunsthal), 2002. Rotterdam 2012 Stephan Kemperdick and Friso Lammertse, The Road to Van Eyck, exh. cat., Rotterdam (Museum Boijmans Van Beuningen), 2012. Rotterdam 2012 Stephan Kemperdick and Friso Lammertse, De weg naar Van Eyck, exh. cat., Rotterdam (Museum Boijmans Van Beuningen), 2012. Rousseau 1951 The Care of paintings (intr. Théodore Rousseau), Paris, 1951 Roy 2000 Ashok Roy, ‘Van Eyck’s Technique: The Myth and the Reality, I’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 97-100. Rudwick 1985 Martin J. S. Rudwick, The Meaning of Fossils. Episodes in the History of Paleontology, Chicago/London, 1985.

585 Rüfenacht 2014 Andreas Rüfenacht, Johann Gottlob von Quandt (1787-1859). Kunstschriftsteller, Sammler und Förderer der Künste in Dresden, Dissertationsprojekt, Universität Zürich, 2010-2015. Ruiz Quesada 2007 Francesc Ruiz Quesada, ‘Lluís Dalmau y la influencia del realismo flamenco en Cataluña’, in María del Carmen Lacarra Ducay (ed.), La pintura gótica durante el siglo XV en tierras de Aragón y en otros territorios peninsulares, Zaragoza, 2007, pp. 243-297. Runciman 1931 Steven Runciman, ‘Some Remarks on the Image of Edessa’, in Cambridge Historical Journal, 3, 3, 1931, pp. 238-252. Rupprich 1956 Hans Rupprich (ed), Albrecht Dürer: Schriftlicher Nachlass, vol. 1, Berlin, 1956. Ruusbroec 1989 Jan van Ruusbroec, Boecsken der Verclaringhe (Corpus Christianorum, Continuatio Mediaevalis, CI), H. Rolfson (trans.), G. de Baere (ed.), Turnhout, 1989. Saitzyk 1987 Steven Saitzyk, Art hardware: The definitive guide to artists’ materials, 1987 [excerpt at: http://www.trueart.info/drying_oils.htm]. Salvadó et al. 2008 Nati Salvadó, Salvador Butí, Francesc Ruiz Quesada, Hermann Emerich and Trinitat Pradell, ‘Mare de Déu dels Consellers, de Lluís Dalmau. Una nova tècnica per a una obra singular’, in Butlletí Museu del Nacional d’Art de Catalunya, 9, 2008, pp. 43-61. Salzman 1979 Louis Francis Salzman, Building in England down to 1540, reprint of the 1967 edition published by Oxford University Press, 1979. Sander 1992 Jochen Sander, Hugo van der Goes, Mainz, 1992. Sander 1993 Jochen Sander, ‘Niederländische Gemälde im Städel 1400-1550’, in Kataloge der Gemälde im Städelschen Kunstinstitut Frankfurt am Main II, Mainz 1993. Sargent, Hogg 1985 Michael Sargent and James Hogg (eds), The Chartae of the Carthusian General Chapter, Paris, Bibliothèque Nationale MS Latin 10888, Part I, 1457-65 (Ff. 1-157v) (Analecta Cartusiana, 100:5), Salzburg, 1985. Sauerberg et al. 2009 Marie Louise Sauerberg, Ashok Roy, Marika Spring, Spike Bucklow and Mary Kempski, ‘Materials and Techniques’, in Paul Binski, A. Massing et al. (eds), The Westminster Retable: History, Technique and Conservation, London/Turnhout, 2009, pp. 244-246. Sauerberg et al. 2012 Marie Louise Sauerberg, Helen Howard, Jo Kirby, Rachel Morrison and David Peggie, ‘The final touches: Evidence from the Study of Varnishes on Medieval Polychromy in England’, in Die Sprache des Materials – Kölner Maltechnik des Spätmittelalters im Kontext, Zeitschrift für Kunsttechnologie und Konserveierung, 26 (1), 2012, pp. 241-258. Scafi 2006 Alessandro Scafi, Mapping Paradise: a History of Heaven on Earth, London, 2006.

Schäfke 1991 Werner Schäfke (ed.), Die Kölner Kartause um 1500. Aufsatzband, Cologne, 1991. Scheewe 1933 Ludwig Scheewe, Hubert und Jan van Eyck. Ihre literarische Würdigung bis ins 18. Jahrhundert, Utrecht, 1933. Scheller 1968 Robert W. Scheller, “‘Als ich can’”, in Oud Holland, 73, 1968, pp. 13539. Scheller 1995 Robert W. Scheller, Exemplum. Model-Book Drawings and the Practice of Artistic Transmission in the Middle Ages (ca. 900-ca. 1470), Amsterdam, 1995. Schiffman 2000 Harvey Richard Schiffman, Sensation and Perception, 2000. Schiller 1972 Gertrude Schiller, Iconography of Christian Art, 2, The Passion of Jesus Christ, J. Seligman (trans.), Greenwich, 1972. Schmidt 2005 Peter Schmidt, Het Lam Gods, Leuven/Baarn, 2005. Shorr 1935 Philip Shorr, ‘The Genesis of Prehistorical Research’, in Isis, 23, 2, September 1935, pp. 425-433. Schottmüller 1936 Frida Schottmüller, Unvollendetes und unveröffentlichtes Typoskript zum Thema Schinkel-Rahmen von 1936, Staatliche Museen zu Berlin, Zentralarchiv, I/GG 39. Schouteet 1989-2004 Albert Schouteet, De Vlaamse Primitieven te Brugge: Bronnen voor de schilderkunst te Brugge tot de dood van Gerard David (Fontes Historiae Artis Neerlandicae, Bronnen in de kunstgeschiedenis van de Nederlanden, 2), 2 vols, Brussels, 1989-2004. Schrader 2010 Stephanie Schrader, ‘Gossaert’s Mythological Nudes and the Shaping of Philip of Burgundy’s Erotic Identity’, in Maryan W. Ainsworth (ed.), Man, Myth, and Sensual Pleasures: Jan Gossart’s Renaissance: The Complete Works, exh. cat., New York (The Metropolitan Museum of Art), 2010, pp. 57-67. Schwarzfuchs 1971 Simon R. Schwarzfuchs, ‘Belgium’, in Encyclopedia Judaica, 5, Jerusalem, 1971, cols 417-418. Scott 1986 Margaret Scott, A Visual History of Costume, The Fourteenth and Fifteenth Centuries, London, 1986. Scott 2000 Margaret Scott, ‘Dress in Van Eyck’s Painting’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 133-148. Scribner 1998 Bob Scribner, ‘Ways of Seeing in the Age of Dürer’, in Dagmar Eichberger and Charles Zika (eds), Dürer and his Culture, Cambridge, 1998, pp. 93-117.

586 Segal 1984 Sam Segal, ‘Die Pflanzen im Genter Altar’, in De Arte et Libis. Festschrift Erasmus 1934-84, Amsterdam, 1984. Seidel 2008 Linda Seidel, ‘Adam and Eve: Shameless First Couple of the Ghent Altarpiece’, in Different Visions: A Journal of New Perspectives on Medieval Art, 1, 2008, pp. 2-18. Serrano 1942 Luciano Serrano, Los conversos D. Pablo de Santa Maria y D. Alfonso de Cartagena (Consejo superior de investigaciones científicas. Instituto Arias Montano – Escuela de estudios hebraicos, Serie B, 1), Madrid, 1942. Serrure, Blommaert 1839-1840 Constant Philippe Serrure and Philip Marie Blommaert (eds), Kronyk van Vlaenderen van 580 tot 1467, 2 vols, Ghent, 1839-1840.

Smeyers 1989 Maurits Smeyers, ‘Answering some questions about the Turin-Milan Hours’, in Le dessin sous-jacent dans la peinture. Colloque VII, Géographie et chronologie du dessin sous-jacent, 1987, Louvain-la-Neuve, 1989, pp. 55-70. Smeyers et al. 1993 Maurits Smeyers et al., Naer natueren ghelike: Vlaamse miniaturen voor Van Eyck (ca. 1350 - ca. 1420), Leuven, 1993. Smeyers 1995 Maurits Smeyers, ‘An Eyckian Vera Ikon in a Book of Hours’, in Serta devota in memoriam Guillermi Lourdaux. Pars posterior: Cultura Mediaevalis. Mediaevalia Lovaniensia Series 1: Studie, 21, Werner Verbeke, Marcel Haverals, Rafaël De Keyser and Jean Goossens (eds), Leuven, 1995, pp. 195-224.

Shearman 1992 John Shearman, Only Connect Art and the Spectator in the Italian Renaissance, Princeton, 1992.

Smeyers 1996 Maurits Smeyers, ‘Jan van Eyck, Archaeologist? Reflections on Eyckian Epigraphy’, in M. Lodewijckx (ed.), Archaeological and Historical Aspects of West-European Societies. Album Amicorum André Van Doorselaer (Acta Archaeologica Lovaniensia, Monographiae 8), Leuven, 1996.

Sheldon, Costaras 2006 Libby Sheldon and Nicola Costaras, ‘Johannes Vermeer’s “Young Woman Seated at a Virginal”‘, in The Burlington Magazine, CXLVIII, 2006, pp. 89-97.

Smeyers 1998a Maurits Smeyers, Vlaamse miniaturen van de 8ste tot het midden van de 16de eeuw, Leuven, 1998.

Sheldon et al. 2012 Libby Sheldon, Tarnya Cooper, Sophie Plender and Charlotte Bolland, ‘Towards a History of the Techniques of the Tudor Artist’s Workshop: Recent Research Using a Combination of Analytical Surface Microscopy and Reconstructions’, Poster Presented at Conference The Renaissance Workshop, London (The British Museum), 10-11 May 2012. Siefert 2010 Helge Siefert, Rahmenkunst. Auf Spurensuche in der Alten Pinakothek, Bayrische Staatsgemäldesammlung München, Munich, 2010. Sievers 1950 Johannes Sievers, Die Möbel, Karl Friedrich Schinkel Lebenswerk, vol. 6, Berlin, 1950. Silva Maroto 2006 Pilar Silva Maroto, ‘Le dessin sous-jacent de deux peintures eyckiennes du Musée du Prado’, in La peinture ancienne et ses procédés. Copies, répliques, pastiches. Le dessin sous-jacent et la technologie dans la peinture, colloque XV, Bruges, 11-13 septembre 2003, Louvain, 2006, pp. 4250. Simonson Fuchs 1982 Anne Simonson Fuchs, ‘The Virgin of the Councillors by Luis Dalmau (1443-1445). The Contract and its Eyckian Execution’, in Gazette des Beaux-Arts, XCIX, 1982, pp. 45-54. Skemer 2006 Don C. Skemer, Binding Words. Textual Amulets in the Middle Ages (Magic in History), Pennsylvania, 2006. Smalley 1952 Beryl Smalley, The Study of the Bible in the Middle Ages, Oxford, 1952. Smeyers 1988 Maurist Smeyers, ‘A Mid-Fifteenth Century Book of Hours from Bruges in the Walters Art Gallery (Ms.721) and its Relation to the TurinMilan Hours’, in Journal of the Walters Art Gallery, 46, 1988, pp. 5576.

Smeyers 1998b Maurits Smeyers, ‘The Living Bread. Dirk Bouts and the Last Supper’, in Dirk Bouts, exh. cat., Leuven (Sint-Pieterskerk and Predikherenkerk), 1998, pp. 35-58. Smeyers, Cardon 1995 Maurits Smeyers and Bert Cardon (eds), ‘Flanders in a European Perspective. Manuscript Illumination around 1400’, in Flanders and Abroad. Proceedings of the International Colloquium, Leuven 7-10 September 1993, Leuven, 1995. Smith 2003 Jamie L. Smith, ‘The Divine Self-Portrait in Northern Devotion: Petrus Christus’s ‘Christ as the Man of Sorrows’’, in Philadelphia Symposium of Art History, Philadelphia, PA (Philadelphia Museum of Art), 2003. Smith 2008 Jamie L. Smith, So moeti den schilt draghen; Dien God veruwede met roder greine: Jan van Eyck’s Critical Principles of Oil Painting and Their Middle Dutch Antecedents, PhD diss., The Johns Hopkins University, Baltimore, MD, 2008. Smith 2009 Jamie L. Smith, ‘Als Ich Can: Poetic Origins of Jan van Eyck’s Device Elucidate Artistic Identity in The Madonna with Canon Van der Paele’, College Art Association, Annual Conference, Los Angeles, CA, 2009. Smith 2012 Jamie L. Smith, ‘Als ich can: How Jan van Eyck Extended the Vernacular from Dutch Poetry to Oil Painting’ in The Transformation of Vernacular Expression in Early Modern Arts, Intersections Yearbook for Early Modern Studies, Joost Keizer and Todd Richardson (eds), Leiden, 2012, pp. 273-312. Smith 2014 Jamie L. Smith, ‘Typology and Spiritual Knighthood in Jan van Eyck’s Madonna with Canon Van der Paele’, in Imago Exegetica: Visual Images as Exegetical Instruments, 1400 - 1700, Intersections Yearbook for Early Modern Studies, Walter Melion and Michel Weemans (eds) Leiden, 2014, pp. 43-71.

587 Snyder, 1976 James Snyder, ‘Jan van Eyck and Adam’s Apple’, in The Art Bulletin, 58, 1976, pp. 511-515. Snyder 1997 James Snyder, Observations on the Iconography of Jan van Eyck’s ‘Saint Francis Receiving the Stigmata,’ in Jane Watkins (ed.), Jan van Eyck: Two Paintings of ‘Saint Francis Receiving the Stigmata, Philadelphia, 1997, pp. 75-87. Sobchack 2000 Vivian Sobchack, ‘What My Fingers Knew: The Cinesthetic Subject, or Vision in the Flesh’, in Senses of Cinema, 5, 2000, pp. 1-29. Sohar et al. 2012 Kristina Sohar, Adomas Vitas and Alar Laanelaid, ‘Sapwood Estimates of Pedunculate Oak (Quercus robur L.) in Eastern Baltic’, in Dendrochronologia, 30-1, 2012, pp. 49-56. Sommé 1998 Monique Sommé, ‘Le cérémonial de la naissance et de la mort de l’enfant princier à la cour de Bourgogne au XVe siècle’, in Jean-Marie Cauchies (ed.), A la cour de Bourgogne. Le duc, son entourage, son train (Burgundica, 1), Turnhout, 1998, pp. 33-48. Sosson 1977 Jean-Pierre Sosson, Les travaux publics de la ville de Bruges, XIVeXVe siècles: les matériaux, les hommes (Collection Histoire Pro Civitate, 48), Brussels, 1977.

Spring 2012a Marika Spring, ‘The materials of Rogier van der Weyden and his contemporaries in context’, in Lorne Campbell, Jan van der Stock, Catherine Reynolds and Lieve Watteeuw (eds), Rogier van der Weyden in Context. Papers presented at the Seventeenth Symposium for the Study of Underdrawing and Technology in Painting held in Leuven, 22-24 October 2009, Leuven, 2012, pp. 93-105. Spring 2012b Marika Spring, ‘Colourless Powdered Glass as an Additive in Fifteenthand Sixteenth-Century European Paintings’, in National Gallery Technical Bulletin, 33, 2012, pp. 4-26. Spring, Higgitt 2006 Marika Spring and Catherine Higgitt, ‘Analyses reconsidered: The Importance of the Pigment Content of Paint in the Interpretation of the Results of Examination of Binding Media’, in Jilleen Nadolny (ed.), Medieval Painting in Northern Europe: Techniques, Analysis, Art History. Studies in commemoration of the 70th birthday of Unn Plahter, London, 2006, pp. 223-229. Spring et al. 2008 Marika Spring, Camilla Ricci, David Peggie and Sergei Kazarian, ‘ATR-FTIR Imaging for the Analysis of Organic Materials in Paint Cross Sections: Case Studies on Paint Samples from the National Gallery, London’, in Analytical and Bioanalytical Chemistry, 392(1-2), 2008, pp. 37-45.

Spantigati 1997 Carlenrica Spantigati, ‘The Turin van Eyck “Saint Francis Receiving the Stigmata”’, in Jane Watkins (ed.), Jan van Eyck: Two Paintings of Saint Francis Receiving the Stigmata, Philadelphia, 1997, pp. 13-27.

Spring et al. 2012 Marika Spring, Rachel Billinge, David Peggie and Rachel Morrison, ‘The technique and materials of two paintings from fifteenth-century Cologne in the National Gallery London’, in Zeitschrift für Kunsttechnologie und Konservierung, Die Sprache des Materials, 26(1), 2012, pp. 88-99.

Spitzbarth 2013 Anne-Brigitte Spitzbarth, Ambassades et ambassadeurs de Philippe le Bon, troisième duc Valois de Bourgogne (1419-1467), Turnhout, 2013.

Stechow 1995 Wolfgang Stechow (ed.), Northern Renaissance Art, 1400 - 1600: Sources and Documents, Englewood Cliffs, 1995.

Spring 2000 Marika Spring, ‘The Technique and Materials of the Blue Draperies’, in Sam Hodge, Marika Spring and Ray Marchant, with Zahira Veliz, ‘The Santa Maria Retable from Mayorga, attributed to the Master of Palanquinos, c. 1490s’, in Hamilton Kerr Institute Bulletin, 3, Cambridge, 2000, pp. 24-28.

Stege et al. 2012 Heike Stege, Claudia Tilenschi and Jana Sanyova, ‘Neues zu den Pigmenten der Altkölner Malerei’, in Zeitschrift für Kunsttechnologie und Konservierung, Die Sprache des Materials, 26(1), 2012, pp. 71-79.

Spring 2004 Marika Spring, ‘Perugino’s Painting Materials: Analysis and Context within Sixteenth-Century Easel Painting’, in Bruno G. Brunetti, Claudio Seccaroni and Antonio Sgamellotti (eds), The Painting Technique of Pietro Vannucci, Called Il Perugino, Proceedings of the LabS Tech Workshop, Quaderni di Kermes, Florence, 2004, pp. 21-28. Spring 2007a Marika Spring, ‘Pigments in Sixteenth-Century Painting of the Germans School’, in Pantxika Béguerie-De Paepe and Michel Menu (eds), La technique picturale de Grünewald et des ses contemporains/The pictorial technique of Grüneward and his peers/Die Maltechnik von Grünewald und seinen Zeigenossen, Paris/Colmar, 2007, pp. 136-144. Spring 2007b Marika Spring, ‘Raphael’s materials: Some New Discoveries and their Context within Early Sixteenth-Century Painting’, in Ashok Roy and Marika Spring (eds), Raphael’s Painting Technique: Working Practices before Rome, Proceedings of the Eu-ARTECH workshop organised by The National Gallery and Eu-ARTECH, London, 11 November 2004, Quaderni di Kermes, Florence 2007, pp. 77-86.

Steppe 1983 Jan Karel Steppe, De Mappemonde geschilderd door Jan van Eyck voor Filips de Goede (Mededelingen van de Koninklijke Academie voor Wetenschappen, Letteren en Schone Kunsten van België, Academiae Analecta, 44), no. 2, 1983, pp. 87-131. Steppe, Smeyers, Lauwerys 1973 Jan Karel Steppe, Maurits Smeyers and Joseph Lauwerys, Wereld van vroomheid en satire. Laat-gotische koorbanken in Vlaanderen, Kasterlee, 1973. Sterling 1976 Charles Sterling, ‘Jan van Eyck avant 1432’, in Revue de l’Art, 33, 1976, pp. 7-82. Steyaert 2014 Delphine Steyaert, ‘Medieval Sculptures Re-polychromed and Regilded in the Nineteenth Century by Adrien Hubert Bressers’, in Making and Transforming Art: Changes in Artists’ Materials and Practice, Proceedings of the Fifth Interim Meeting of the ICOM-CC Working Group Art Technological Source Research, Brussels, Royal Institute for Cultural Heritage (KIK-IRPA), 22-23 November 2012, London, 2014.

588 Strasbourg 2008 Philippe Lorentz (ed.), Strasbourg 1400, exh. cat., Strasbourg (Musée de l’Œuvre de Notre-Dame), 2008. Streeton 2011a Noëlle L.W. Streeton, ‘Jan van Eyck’s Dresden Triptych: New Evidence for the Giustiniani of Genoa in the Borromei ledger for Bruges, 1438’, in Journal of Historians of Netherlandish Art, 3, 1, 2011. Streeton 2011b Noëlle L.W. Streeton, Painting in Bruges in the Early Fifteenth Century. Investigations of the Technique of Jan van Eyck and Imported Artists’ Materials, Leading to an Evaluation of Methods for Technical Art History, unpublished PhD diss., University College London, 2011 [http://www. jhna.org/index.php/past-issues/vol-3-1/136-janvaneycksdresden]. Streeton 2013 Noëlle L.W. Streeton, Perspectives on the Painting Technique of Jan van Eyck: Beyond the Ghent Altarpiece, London, 2013. Strehlke 2001 Carl B. Strehlke, ‘Valencia. Early Flemish and Valencian Art. Review’, in The Burlington Magazine, 143, 1182, Sept. 2001, pp. 590-593. Strehlke, Israëls 2015 Carl Brandon Strehlke and Machtelt Israëls, Catalogue of the Paintings in the Berenson Collection, Florence: Villa I Tatti, the Harvard University Center for Italian Renaissance Studies, 2015. Stroo 2009 Cyriel Stroo (ed.), Pre-Eyckian Panel Painting in the Low Countries (Contributions to Fifteenth-Century Painting in the Southern Netherlands and the Principality of Liège, 9), 2 vols, Brussels, 2009. Stroo, Smeyers 1994 Cyrel Stroo and Maurits Smeyers, ‘Hubert et Jean van Eyck’, in Les Primitifs flamands et leur temps, Brigitte de Patoul and Roger Van Schoute (eds), Louvain-la-Neuve, 1994, pp. 291-292. Stroo, Syfer-d’Olne 1996 Cyriel Stroo and Pascale Syfer-d’Olne, The Flemish Primitives I. The Master of Flémalle and Rogier van der Weyden Groups.Catalogue of Early Netherlandish Painting in the Royal Museums of Fine Arts of Belgium, Brussels, 1996. Swenden 1960 Karel Swenden, ‘Denys le chartreux’, in Dictionnaire d’histoire et de géographie ecclésiastiques, XIV (fasc 78), Paris, 1960, cols 256-260. T.S. 1614 T.S. (Juliana Berners?), A Iewell for Gentrie, London, 1614 [The English experience, its record in early printed books in facsimile no. 890, Amsterdam 1977]. Tanimoto, Verri 2009 Satoko Tanimoto and Giovanni Verri, ‘A Note on the Examination of Silverpoint Drawings by Near-Infrared Reflectography’, in Studies in Conservation, 54, 2009, pp. 106-116. Taubert 1975 Johannes Taubert, ‘Beobachtungen zum schöpferischen Arbeitsprozesz bei einigen altniederländischen Malern’, in Nederlands Kunsthistorisch Jaarboek, 26, 1975, pp. 41-72. Taubert 1978 Johannes Taubert, Farbige Skulpturen. Bedeutung, Fassung, Restaurierung, Munich, 1978.

The Hague/Antwerp 1998 Cd-Rom Middelnederlands, The Hague/Antwerp, 1998. The Jewish Encyclopedia 1901-1906 The Jewish Encyclopedia, 12  vols, New York, 1901-1906 [www. jewishencyclopedia.com]. Thomas Cantimpratensis 1973 Thomas Cantimpratensis liber de natura rerum. Editio princeps secundum codices manuscriptos, 1 (text), Berlin/New York, 1973. Thomas, Pavitt 1922 William Thomas and Kate Pavitt, The Book of Talismans, Amulets and Zodiacal Gems, third revised edn, London, 1922. Throop 1998 Priscilla Throop (ed.), Hildegard von Bingen’s Physica: the Complete English Translation of her Classic Work on Health and Healing, Rochester, 1998. Thurston 1897 Herbert J. Thurston, ‘Genuflexion at Mass’, in The Month, LXXXIX, October 1897, p. 400. Thurston 1916 Herbert J. Thurston, ‘Genuflections and Aves: a Study in Rosary Origins’ in The Month, CXXVII, 1916, pp. 441, 546. Tietzel 1994 Brigitte Tietzel, ‘Sein und Schein in Jan van Eycks gemalten Stoffen’, in Ingrid Guntermann and Brigitte Tietzel (eds), Festschrifte für Brigitte Klesse, Berlin, 1994, pp. 217-231. Toscano 2007 Gennaro Toscano, ‘Alfonso el Magnánimo y el Norte. Obras flamencas en las colecciones reales’, in Eduard Mira and An Delva (eds), A la búsqueda del Toisón de Oro. La Europa de los Príncipes. La Europa de las ciudades, vol. 1, exh. cat., Almudín (Museo de la ciudad), 2007, pp. 347-363. Tossavainen 1994 Jouko Tossavainen, Dutch Forest Products’ Trade in the Baltic from the Late Middle Ages to the Peace of Munster in 1648, PhD diss., University of Jyväskylä (Finland), 1994. Tóth-Ubbens 2012 Magdi Tóth-Ubbens, ‘Een vorstelijke jachtpartij in Frankrijk. De kunst van het jagen met valken’, in Journal of Eurasian Studies, OctoberDecember 2012 [www.federatio.org/joes.html]. Tronzo 1983 William Tronzo, ‘The Hildesheim Doors: An Iconographic Source and Its Implications’, in Zeitschrift für Kunstgeschichte, 46, 1983, pp. 357-366. Tumosa, Mecklenburg 2005 Charles S. Tumosa and Marion F. Mecklenburg, ‘The Influence of Lead Ions on the Drying of Oils’, in Reviews in Conservation, 6, 2005, pp. 39-47. Turel 2011 Noa Turel, ‘Living Pictures: Rereading “au vif”, 1350-1550’, in Gesta, 50/2, 2011, pp.163-182. Turnhout, 1992 Album palaeographicum XVII Provinciarum, Turnhout, 1992. Tyers, Tyers 2007 Cathy Tyers and Ian Tyers, Tree-Ring Analysis of the Nave Ceiling of Peterborough Cathedral, City of Peterborough, Research Department Report 004/2007, 2007.

589 Underhill 1910 Evelyn Underhill, ‘The Fountain of Life: An Iconographical Study’, in The Burlington Magazine for Connoisseurs,17, 1910, pp. 99-109. Unterkircher 1984 Franz Unterkircher, Das Rothschild-Gebetbuch, die schönsten Miniaturen, Graz, 1984. Upton 1990 Joel Upton, Petrus Christus: his Place in Fifteenth-Century Flemish Painting: A Study of the Changing Relationship of Painter, Painting, and Patron, University Park/London, 1990. Utrecht/New York 1989-1990 Henri L.M. Defoer, Anne S. Korteweg and Wilhelmina C.M. Wüstefeld, The Golden Age of Dutch Manuscript Painting, exh. cat., Utrecht (Rijksmuseum het Catharijneconvent)/New York (Pierpont Morgan Library), 1989-1990. Uytven 1998 Raymond van Uytven, De zinnelijke middeleeuwen, Leuven, 1998. Vagnini et al. 2009 M. Vagnini, C. Miliani, L. Cartechini, P. Rocchi, B. G. Brunetti and A. Sgamellotti, ‘FT-NIR Spectroscopy for Non-Invasive Identification of Natural Polymers and Resins in Easel Paintings’, in Analytical and Bioanalytical Chemistry, 395, 2009, pp. 2107-2118. Vaivre 1985 Jean-Bernard de Vaivre, ‘Chasse à l’oiseau et cour d’amour’ in Journal des Savants, 1985, pp. 313-339. Valencia 2001 Fernando Benito Doménech and José Gómez Frechina (eds), La Clave Flamenca en los Primitivos Valencianos, exh. cat., Valencia (Museo de Bellas Artes), 2001. Van Anrooij 1997 Wim van Anrooij, Helden van weleer. De Negen Besten in de Nederlanden (1300-1700), Amsterdam, 1997. Van Asperen de Boer 1974 J. R. J. van Asperen de Boer, ‘An Examination of Particle Size Distributions of Azurite and Natural Ultramarine in Some Early Netherlandish Paintings’, in Studies in Conservation, 19, 1974, pp. 233-243. Van Asperen de Boer 1979 J. R. J. van Asperen de Boer, ‘A Scientific Re-examination of the Ghent Altarpiece’, in Oud Holland, 93, 1979, pp. 141-214. Van Asperen de Boer 1990 J. R. J. van Asperen de Boer, ‘Technical Aspects of Some Eyckian Portrait Paintings’, in Jan De Jong (ed.), Werk. Opstellen voor Hans Locher, Groningen, 1990, pp. 8-12. Van Asperen de Boer 1992 J. R. J. van Asperen de Boer, ‘Over de techniek van Jan van Eyck’s: De Heilige Barbara’, in Jaarboek van het Koninklijk Museum voor Schone Kunsten Antwerpen, 1992, pp. 9-13. Van Asperen de Boer 1995 J. R. J. van Asperen de Boer, ‘Infrared Reflectograms of two Paintings by Jan van Eyck in Bruges’, in Hélène Verougstraete and Roger Van Schoute (eds), Le dessin sous-jacent dans la peinture, Colloque X. Le dessin sous-jacent dans le processus de création, Louvain-la-Neuve, 1995, pp. 81-84.

Van Asperen de Boer 1997 J. R. J. van Asperen de Boer, ‘Some Technical Observations on the Turin and Philadelphia Versions of “Saint Francis Receiving the Stigmata”’, in Jane Watkins (ed.), Jan van Eyck: Two Paintings of Saint Francis Receiving the Stigmata, Philadelphia, 1997, pp. 51-63. Van Asperen de Boer 2004 J. R. J. van Asperen de Boer, ‘A Note on the Original Disposition of the Ghent Altarpiece and the Beaune Polyptych’, in Oud Holland, 117, 2004, pp. 107-118. Van Asperen de Boer, Giltaij 1987 J. R. J. van Asperen de Boer and Jeroen Giltaij, ‘Een nader onderzoek van “De Drie Maria’s aan het Graf” – een schilderij uit de “groep van Eyck” in Rotterdam’, in Oud Holland, 101(4), 1987, pp. 254-274. Van Asperen de Boer, Faries 1990 J. R. J. van Asperen de Boer and Molly Faries, ‘La Vierge au Chancelier Rolin de Van Eyck: examen au moyen de la réflectographie à l’infrarouge’, in Revue du Louvre et des Musées de France, 40, 1990, 1, pp. 3749. Van Asperen de Boer, Ridderbos, Zeldenrust 1991 J. R. J. van Asperen de Boer, Bernhard Ridderbos and Manja Zeldenrust, ‘Portrait of a Man with a Ring by Jan van Eyck’, in Bulletin van het Rijksmuseum, 39, 1991, pp. 8-35. Van Autenboer 1972 Eugeen Van Autenboer, ‘Bijdrage tot de geschiedenis der Kolveniersgilden in het Oude Hertogdom Brabant’, in Taxandria. Jaarboek van de Koninklijke geschied- en oudheidkundige kring van de Antwerpse Kempen, no. XLIV, 1972. Van Belle 2002 Ronald Van Belle, ‘De vrucht in Eva’s hand op het Lam Gods te Gent. Een «  Pomum Adami  » of een andere vrucht  ?’, in Biekorf, 2002, pp. 50-59. Van Belle 2006 Ronald Van Belle, Vlakke grafmonumenten en memorietaferelen met persoonsafbeeldingen in West-Vlaanderen, een inventaris, funeraire symboliek en overzicht van het kostuum, Bruges, 2006. Van Belle 2012 Ronald Van Belle, ‘De vrucht in Eva’s hand op het Lam Gods van Van Eyck. De etrog of wild lemoen (Citrus hystrix)’, in Biekorf, 2012, pp. 91-93. Van Buren 1996 Anne H. Van Buren, ‘Jan van Eyck’, in Dictionary of Art, 10, 1996, pp. 705-714. Van Buren 2011 Anne H. Van Buren with Roger S. Wieck, Illuminating Fashion, Dress in the Art of Medieval France and the Netherlands 1325-1515, New York/ London, 2011. Van Buren, Marrow, Pettenati 1996a Anne H. Van Buren, James H. Marrow and Silvana Pettenati, Heures de Turin-Milan, Inv. No 47, Museo Civico d’Arte Antica, Torino, Kommentar, Lucerne, 1996. Van Buren, Marrow, Pettenati 1996b Anne H. Van Buren, James H. Marrow and Silvana Pettenati, Heures de Turin-Milan, Inv. No 47, Museo Civico d’Arte Antica, Torino, Lucerne, 1996 (facsimile).

590 Van de Velde 1855 Hippolyte Van de Velde, ‘Anciennes moeurs et coûtumes. Histoire de la procession de Furnes’, in Annales de la Société d’Emulation, 2nd series, 1855, pp. 143-228.

Van Duijn 2013b Esther van Duijn, All that Glitters is not Gold. The Depiction of GoldBrocaded Velvets in Fifteenth- and Early Sixteenth-Century Netherlandish Paintings, unpublished PhD diss., Amsterdam, 19 June 2013.

Van den Berg et al. 2000 Klaas Jan van den Berg et al., ‘On Copper Green Glazes in Paintings’, in Jacques Goupy and Jean-Pierre Mohen, Art et Chimie, la Couleur: Actes du Congrès, Paris, 2000, pp. 18-21.

Van Duijn, Roeders 2012 Esther van Duijn and Jessica Roeders, ‘Gold-brocaded Velvets in Paintings by Cornelis Engebrechtsz’, in Journal for Historians of Netherlandish Art, 4, 2012, 1, pp. 1-13 (pagination on the PDF file) [www. jhna.org].

Van den Kerkhove, Baldewijns 1986 Antoine Van den Kerkhove and Jeannine Baldewijns, Museum voor stenen voorwerpen (Ruines van de SintBaafsabdij): gids voor de bezoeker, Ghent, 1986. Van der Meer 1938 Frits Van der Meer, Maiestas Domini. Théophanies de l’Apocalypse dans l’art chrétien. Études sur les origines d’une iconographie spéciale du Christ (Studi di antichità cristiana, 13), Vatican City, 1938. Van der Meer 1978 Frits Van der Meer, Apocalypse. Visioenen van het Boek der Openbaring in de Kunst, Antwerp, 1978. Van der Velden 2010 Hugo Van der Velden, ‘The Genesis of the Ghent Altarpiece’, paper presented at the conference Vision & Material. Interaction between Art and Science in Jan van Eyck’s Time, Vlaams Academish Centrum, Brussels, 26 November 2010. Van der Velden 2011a Hugo Van der Velden, ‘The Quatrain of ‘The Ghent Altarpiece’’, in Simiolus, 35, 2011, pp. 5-39. Van der Velden 2011b Hugo Van der Velden, ‘A Reply to Volker Herzner and a Note on the Putative Author of the Ghent Quatrain’, in Simiolus, 35, 2011, pp. 131141. Van der Weerd et al. 2002 Jaap van der Weerd, Jaap J. Boon, Muriel Geldof, Ron M.A. Heeren and Petria Noble, ‘Chemical Changes in Old Master Paintings: Dissolution, Metal Soap Formation and Remineralization Processes in Lead Pigmented Paint Layers of 17th Century Paintings’, in Zeitschrift für Kunsttechnologie und Konservierung, 16, 2002, pp. 36-51. Van Delf 1937 Dirc van Delf, Tafel van den Kersten Ghelove. Winterstuk. vol. 2, L. M. Daniels (ed.) Antwerp, 1937. Cd-Rom Middelnederlands. Van Delf 1938 Dirc van Delf, Tafel van den Kersten Ghelove. Somerstuc, vol. 3, L. M. Daniels (ed.), Antwerp, 1938. Cd-Rom Middelnederlands. Van Dixmude 1835 Olivier van Dixmude, Merkwaerdige gebeurtenissen, vooral in Vlaenderen en Brabant, en ook in de aengrenzende landstreken, van 1377 tot 1443, Jean-Jacques Lambin (ed), Ypres, 1835. Van Doren 1859 Pierre Joseph Van Doren, Inventaire des archives de la ville de Malines, 1, Mechelen, 1859. Van Duijn 2013a Esther van Duijn, ‘Gold-Brocaded Velvets in Paintings by Geertgen tot Sint Jans. A Study of Geertgen’s Painting Technique to Imitate Gold-Brocaded Velvets and a Comparison with the “Raising of Lazarus” by Albert van Ouwater’, in ArtMatters, 5, 2013, pp. 27-47 [http://www.artmattersjournal.org].

Van Eikema Hommes 2004 Margriet van Eikema Hommes, Changing Pictures. Discolouration in 15th-17th-Century Oil Paintings, London, 2004. Van Gelder 1951 Jan Gerrit Van Gelder, ‘Het zogenaamde portret van Dieric Bouts op het “Werc van den Heilichen Sacrament”‘, in Oud Holland, 66, 1951, pp. 51-52. Van Gelder 1972 Jan Gerrit Van Gelder, ‘Der Teufel stiehlt das Tintenfaß’, in Artur Rosenauer and Gerold Weber (eds), Kunsthistorische Forschungen, Salzburg, 1972, pp. 173-183. Van Grevenstein 2011 Anne van Grevenstein, Lasting Support. An Interdisciplinary Research Project to Assess the Structural Condition of the Ghent Altarpiece, 2011. Van Grevenstein, Spronk 2011 Anne van Grevenstein and Ron Spronk, Lasting Support: An Interdisciplinary Research Project to Assess the Structural Condition of the Ghent Altarpiece. Final project report, 9 October 2011 [http://closertovaneyck. kikirpa.be/#home/sub=documents (accessed March 2012)]. Van Grevenstein-Kruse et al. 2011 Anne van Grevenstein-Kruse et al., Conservatie en materieel onderzoek van het Retabel van het Lam Gods, Gent, Sint-Baafskathedraal. Deel I: Verslag van het onderzoek naar de materiële conditie en van de urgente conservatiebehandeling (april tot november 2010) dir. Prof. Anne van Grevenstein-Kruse (University of Amsterdam), 16 March 2011 [http:// closertovaneyck.kikirpa.be/#home/sub=documents]. Van Herreweghen 1964 G. van Herreweghen, ‘Het oudste cijnsboek van Pamel (1391)’, in Eigen Schoon en de Brabander. Maandschrift van het geschied- en oudheidkundig Genootschap van Vlaams-Brabant, 47, 1964, pp. 169-203. Van Hese 1931 Jean Van Hese, ‘L’occupation française à Bruges en 1792 et 1793, Journal contemporain de Jean Van Hese’, in E. Hosten and E. Strubbe (eds), Bronnen en bijdragen tot de Vlaamse geschiedvorsing, 1, Bruges, 1931. Van Leeuwen 2004 Jacoba van Leeuwen, De Vlaamse wetsvernieuwing. Een onderzoek naar de jaarlijkse keuze en aanstelling van het stadsbestuur in Gent, Brugge en Ieper in de Middeleeuwen (Verhandelingen van de Koninklijke Vlaamse Academie van België voor Wetenschapppen en Kunsten, Nieuwe Reeks, 15), Brussels, 2004. Van Loon 2008 Annelies van Loon, ‘The Relationship Between Preservation and Technique in the Paintings of the Oranjezaal’, in Color Changes and Chemical Reactivity in Seventeenth-Century Oil Paintings, Amsterdam, 2008.

591 Van Maerlant 1863 Jacob van Maerlant, Spiegel historiael, M. de Vries and E. Verwijs (eds), met de fragmenten der later toegevoegde gedeelten, bewerkt door Philip Utenbroeke en Lodewijc van Velthem. 3 vols, Leiden, 1863. CD-Rom Middelnederlands. Van Maerlant 1954 Sinte Franciscus Leven van Jacob van Maerlant, P. Maximilianus (ed.), Part 2, Zwolle, 1954. Cd-Rom Middelnederlands. Van Maerlant 1981 Jacob van Maerlant, Der naturen bloeme, M. Gysseling (ed.), Corpus van Middelnederlandse teksten (tot en met het jaar 1300). II: Literaire handschriften. Deel 2, Der Naturen Bloeme, The Hague, 1981. CDRom Middelnederlands. Van Mander 1604 Karel van Mander, Het Schilder-Boeck, Haarlem 1604. Van Mander 1969 Karel van Mander, Het schilder-boeck, waerin voor eerst de leerslustighe Iueght den gront der edel vry schilderconst in verscheyden deelen work voorghedraeghen. Reprint of the first edn from Haarlem, P.v. Westusch, 1604, Utrecht, 1969. Van Mander 1995 Karel Van Mander, The Lives of the Illustrious Netherlandish and German Painters, vol. II, Hessel Miedema (ed.), Doornspijk, 1995. Van Melckebeke 1874 Guillielmus Judocus Joannes Van Melckebeke, Geschiedkundige aanteekeningen rakende de Sint-Christoffel of Kolveniers-gilde te Mechelen, Mechelen, 1874. Van Melckebeke 1913 Guillielmus Judocus Joannes Van Melckebeke, Geschiedkundige aanteekeningen rakende de kruis- of voetboog-gilde te Mechelen, Mechelen, 1913. Van Nuffel 1979 Herman van Nuffel, ‘Vaernewyck, Marcus van’, in Nationaal biografisch woordenboek, vol. 8, 1979, pp. 795-809. Van Os 1974 Henk W. Van Os, ‘Saint Francis of Assisi as a Second Christ in Early Italian Painting’, in Simiolus, 7, 3, 1974, pp. 115-132. Van Puyvelde 1940 Léo van Puyvelde, De reis van Jan van Eyck naar Portugal (Koninklijke Vlaamse Academie voor Taal-en Letterkunde. Verslagen en Mededeelingen), 1940, pp. 17-27.

Van Vaernewyck 1566-1568 (1872) Marcus van Vaernewyck, Van die beroerlicke Tijden in die Nederlanden en voornamelick in Ghendt 1566-1568, Ferdinand Vanderhaegen (ed.), 5 vol., Ghent, 1872-1881 (vol. 1, Ghent, 1872, pp. 143-146). Van Vaernewyck 1568 Marcus van Vaernewyck, Den Spieghel der Nederlandscher Oudtheyt, inhoudende die constructie ofte vergaderinghe van Belgis, Ghendt, 1568, fols 117r-119r. Van Werveke 1895-1896 Alfons Van Werveke, ‘Ontleding van de tekst van M. Van Vaernewijck betreffende het graf en de grafzerk van Hubrecht Van Eyck’, in Handelingen van den geschied- en oudheidkundige kring van Gent, 2, 1895-1896, pp. 1-8. Van Werveke 1898 Alfons Van Werveke, ‘De grafzerk van Huibrecht van Eyck’, in Inventaire archéologique, 1898, no. 73. Van Werveke 1947 Hans van Werveke, Gent. Schets van een sociale geschiedenis (Uitgave van het Willemsfonds, 183), Ghent, 1947. Vandamme 1974 E. Vandamme, ‘Een 16e-eeuws Zuidnederlands receptenboek’, in Jaarboek van het Koninklijk Museum voor Schone Kunsten Antwerpen, 1974, pp. 101-113. Vandekerchove et al. 2012 Véronique Vandekerchove et al., ‘The Leuven Trinity. A New Look at a Fifteenth-Century Painting in the M Collection’, in Lorne Campbell et al. (eds), Rogier van der Weyden in Context. Papers presented at the Seventeenth Symposium for the Study of Underdrawing and Technology in Painting held in Leuven, 22-24 October 2009, Paris/Leuven/Walpole, 2012, pp. 179-194. Vandenbroeck 1981 Paul Vandenbroeck, Rondom plechtige intredes en feestelijke stadsversieringen Antwerpen 1594-1599-1635. Documentaire tentoonstelling, Antwerp, 1985. Vandenbroeck 1985 Paul Vandenbroeck, Catalogus schilderijen 14e en 15e eeuw. Koninklijk Museum voor schone Kunsten Antwerpen, Antwerp, 1985. Vandenbroeck 1987 Paul Vandenbroeck, Beeld van de andere, vertoog over het zelf: van wilden en narren, boeren en bedelaars, Antwerp, 1987. Vandenbroeck 1989 Paul Vandenbroeck, ‘Jheronimus Bosch’ zogenaamde Tuin der Lusten. I’,in Jaarboek Koninklijk Museum voor Schone Kunsten Antwerpen, 1989, pp. 9-210.

Van Rompaey 1967 Jan Petrus Cornelis Van Rompaey, Het grafelijk baljuwsambt in Vlaanderen tijdens de Boergondische periode (Verhandelingen van de Koninklijke Vlaamse Academie voor Wetenschappen, Letteren en Schone Kunsten van België, Klasse der Letteren, 29), Brussels, 1967.

Vandenbroeck 1990 Paul Vandenbroeck, ‘Jheronimus Bosch’ zogenaamde Tuin der Lusten. II’, in Jaarboek Koninklijk Museum voor Schone Kunsten Antwerpen, 1990, pp. 9-192.

Van Steenberghe 1935 Edmond Van Steenberghe, ‘Jean Gerson à Bruges’, in Revue d’histoire ecclésiastique, 31, 1935, pp. 5-52.

Vander Linden 1940 Herman Vander Linden, Itinéraires de Philippe le Bon, duc de Bourgogne (1419-1467) et de Charles, comte de Charolais (1433-1467), Brussels, 1940.

Van Uytven 1991 Raymond Van Uytven, ‘De Mechelse burgers tegenover hun politieke meesters’, in Raymond Van Uytven (ed.), De geschiedenis van Mechelen. Van Heerlijkheid tot Stadsgewest, Tielt, 1991, pp. 75-80.

Vanderhaeghen 1872-1881 Ferdinand Vanderhaeghen (ed.), Marcus van Vaernewyck, Van die beroerlicke tijden in die Nederlanden en voornamelick in Ghendt 15661568, Ghent, 1872-1881.

592 Vandewalle 1994 Andre Vandewalle, ‘El consulado de Burgos en los Países Bajos’, in Actas del V Centenario del consulado de Burgos I, Burgos, 1994, pp. 283300. Vandivere 2011a Abbie Vandivere, ‘In Search of van Mander’s Primuersel: Intermediate Layers in Early Netherlandish Paintings’, in Janet Bridgeland (ed.), ICOM-CC 16th Triennial Conference, Lisbon, 19-23 September 2011, Almada, 2011, pp. 1-8. Vandivere 2011b Abbie Vandivere, ‘“Cloeck en veerdigh”: Energetic and Skillful Painting Techniques of the Sixteenth-Century Leiden School’, in Journal of the Historians of Netherlandish Art, 4, 1, Winter 2011 [www.jhna.org]. Vandivere 2013 Abbie Vandivere, From the Ground Up: Surface and Sub-Surface Effects in Fifteenth- and Sixteenth-Century Netherlandish Paintings, PhD diss., University of Amsterdam, 2013. Vanwijnsberghe 2008 Dominique Vanwijnsberghe (ed.), Autour de la Madeleine Renders. Un aspect de l’histoire des collections, de la restauration et de la contrefaçon en Belgique dans la première moitié du XXe siècle (Scientia Artis, 4), Brussels, 2008. Vanwijnsberghe 2011 Dominique Vanwijnsberghe, ‘De “Vlaamse” miniatuur. Route naar een verfijnde kaart van een transregionale productie’, in Bernard Bousmanne and Thierry Delcourt (eds), Vlaamse miniaturen 1404-1482, Brussels/Paris, 2011, pp. 19-37. Vasari 1568 Giorgio Vasari, Lives of the Painters, Sculptors and Architects, 1568, Gaston du C. de Vere (trans.), London, 1912-15, reprint, New York, 1996. Vasari 1568 (1968) G. Vasari, Le Vite de’ più eccellenti pittori, scultori ed architettori […], R. Bettarini (ed.), Verona, 1966-87. Vasconcelos 1897 Joaquim de Vasconcelos, ‘Voyage de Jehan Van Eyck – Viaje de Jan van Eyck’, in Revista de Guimarães, XIV, 1897, pp. 10-45. Vegas 1998 Liana C. Vegas, Jan van Eyck, and Hubert van Eyck, Milan, 1998. Veliz 2000 Zahira Veliz, ‘Appendix: The 1493 Ordenanzas de Córdoba for Regulating the Profession of Painting’, in Sam Hodge, Marika Spring and Ray Marchant, with Zahira Veliz, ‘The Santa Maria Retable from Mayorga, Attributed to the Master of Palanquinos, c. 1490s’, in Hamilton Kerr Institute Bulletin, 3, Cambridge, 2000, pp. 35-40. Veltman 1977 Kim H. Veltman, ‘Panofky’s Perspective: A Half Century Later’, in Atti del convegno internazionale di studi: la prospettiva rinascimentale, Milan, 1977, pp. 565-584.

Vermeersch 1981 Valentin Vermeersch, Brugge Duizend jaar Kunst, Antwerp, 1981. Veronee-Verhaegen 1973 Nicole Veronee-Verhaegen, L’Hôtel-Dieu de Beaune (Corpus de la peinture des anciens Pays-Bas méridionaux au quinzième siècle, 13) Brussels, 1973. Verougstraete, Van Schoute 1987 Hélène Verougstraete and Roger Van Schoute, ‘Les cadres de l’Agneau Mystique de Van Eyck’, in Revue de l’Art, 77, 1987, pp. 73-76. Verougstraete-Marcq, Van Schoute 1989 Hélène Verougstraete-Marcq and Roger Van Schoute, Cadres et supports dans la peinture flamande aux 15e et 16e siècles, Heure-le-Romain, 1989. Verougstraete, Van Schoute 1995 Hélène Verougstraete and Roger Van Schoute (eds), Le dessin sousjacent dans la peinture, Colloque X: Le dessin sous-jacent dans le processus de création (Université Catholique de Louvain, Institut supérieur d’Archéologie et d’Histoire de l’Art, Document de travail no. 28), Louvain-laNeuve, 1995. Verougstraete, Van Schoute 1999 Hélène Verougstraete and Roger Van Schoute (ed.), Le dessin sousjacent dans la peinture, Colloque XII: La peinture dans les Pays-Bas au 16e siècle. Pratique d’atelier. Infrarouges et autres méthodes d’investigation, Bruges, 11-13 septembre 1997, Leuven, 1999. Verougstraete, Van Schoute 2000 Hélène Verougstraete and Roger Van Schoute, ‘Frames and Supports of some Eyckian Paintings’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 107-118. Verwijs 1980 Eelco Verwijs (ed.), Jacob Van Maerlant’s Naturen Bloeme (Bibliotheek van Middelnederlandse letterkunde), Arnhem, 1878 – anast. reprint. Groningen, 1980. Verwijs, Verdam 1882-1952 Eelco Verwijs and Jakob Verdam, Middelnederlandsch Woordenboek, 11  vols, The Hague/Zedelgem, 1882-1952. Cd-Rom Middelnederlands. Viaene 1965 Antoon Viaene, ‘Het Grafpaneel van Kanunnik van der Paele voltooid in 1436 door Jan van Eyck (Groeningemuseum Brugge)’, in Biekorf, 66, 1965, pp. 257-64. Viaene 1975-1976 Antoon Viaene, ‘De legende van de drie levenden en de drie doden in beeld en in vlaamse processiespelen 1400-1820’, in Biekorf, 76, 19751976, pp. 101-105. Viaene 1977 Antoon Viaene, ‘Processie van de drie levenden en de drie doden Nieuwpoort 1413-1431’, in Biekorf, 77, 1977, p. 119.

Verelst 1984 Dirk Verelst, Geschiedenis van volk en land van Beveren, onder leiding van Gustaaf Asaert, 1. Politieke en Institutionele Geschiedenis, Beveren, 1984.

Villela-Petit 2011 Inès Villela-Petit, ‘Quatre feuillets des Très Belles Heures de Notre Dame de Jean de Berry’, in François Avril, Nicole Reynaud and Dominique Cordelier (eds), Les enluminures du Louvre: Moyen Age et Renaissance, Paris, 2011, nos 77-80, pp. 140-162.

Vermeersch 1976 Valentin Vermeersch, Grafmonumenten te Brugge voor 1578, Bruges, 1976.

Vinckier 2002 Josef Vinckier, ‘Étude dendrochronologique de quelques panneaux de l’Agneau Mystique’, in Bulletin de l’Institut royal du Patrimoine

593 artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 28, 1999-2000, Brussels, 2002, pp. 237-240. Von Bode 1929 Wilhelm von Bode, ‘Bilderrahmen in alter und neuer Zeit’, in Antike Rahmen, exh. cat., Berlin (Künstlerhaus), 1929, pp. 7-45 (Wiederabdruck von v. Bode 1898-1899). Von Euw, Plotzek 1982 Anton von Euw and Joachim Plotzek, Die Handschriften der Sammlung Ludwig, 4 vols, Cologne, 1982. Von Oettingen 1890 Wolfgang von Oettingen (ed.), ‘Antonio Averlino Filarete’s “Tractat uber die Bukunst”‘, in Quellenschriften für Kunstgeschichte und Kunsttechnik des Mittelalters und der Neuzeit, neue folge, III, Vienna, 1890. Von Quandt 1816 Johann Gottlob von Quandt, ‘Ueber den gegenwärtigen Zustand der Gemäldegallerie in Dresden’, in Journal für Literatur, Kunst, Luxus und Mode, 31, December 1816, pp. 831-843.

Waagen, Rassman, de Bast 1825 Gustav Friedrich Waagen, Notice sur le chef-d’œuvre des frères Van Eyck, traduit de l’allemand par Friedrich Rassmann, Augmentée de notes inédites sur la vie et sur les ouvrages par Liévin de Bast, Ghent, 1825. Wallert 2013 Arie Wallert, ‘Function and Meaning of a Metalpoint Drawing by Jan van Eyck’, in ArtMatters, 5, 2013, pp. 2-9. Wallert et al. 2012 Arie Wallert, Katrien Keune, Jolanda van Iperen and Joana Pedroso, ‘Early Northern Netherlandish Painting and Gilding Techniques and Cologne: Some Parallels’, in Zeitschrift für Kunsttechnologie und Konservierung, 26(2), 2012, pp. 230-240. Walsch 2012 Peter G. Walsh with Christopher Husch (eds and trans.), One Hundred Latin Hymns from Ambrose to Aquinas (Dumbarton Oaks Medieval Library, 18), Cambridge, 2012. Ward 1975 John L. Ward, in The Art Bulletin, 57, no. 2, 1975, pp. 196-220.

Von Roenne 2007 Bettina von Roenne, Ein Architekt rahmt Bilder. Karl Friedrich Schinkel und die Berliner Gemäldegalerie, exh. cat., Munich/Berlin, 2007.

Ward 1983 John L. Ward, in The Art Bulletin, 65, no. 4, 1983, pp. 680-690.

Von Roenne 2011 Bettina von Roenne, Schinkel rahmt. Die Bilderrahmen Karl Friedrich Schinkels, diss. at Freie Universität Berlin, Berlin, 2011.

Wardwell 1977 Anne E. Wardwell, ‘The Stylistic Development of 14th- and 15thCentury Italian Silk Design’, in Aachner Kunstblätter, 47, 1976/77, pp. 177-226.

Von Roenne 2012 Bettina von Roenne, ‘Bilderrahmen neu entdeckt. Das in Vergessenheit geratene Rahmenwerk Schinkels’, in Karl Friedrich Schinkel. Geschichte und Poesie - Das Studienbuch, Herausgegeben von Hein-th Schulze Altcappenberg and Rolf Johannsen, Berlin/Munich, 2012. Von Sandrart 1994 Joachim von Sandrart, Teutsche Academie der edlen Bau-, Bild- und Malerei-Künste, Nuremberg 1675-80 (facsimile edn, Nördlingen 1994), bk. 3, pt. 1, ch. 7. Von Schlosser 1904 Julius von Schlosser, ‘Über einige Antiken Ghibertis’, in Jahrbuch der kunsthistorischen Sammlungen des allerhöchsten Kaiserhauses, 24, fasc. 4, 1904, pp. 125-159. Vynckier 1992 Josef Vynckier, ‘Dendrochronological Dating of the Wings’, in Peter Paul Rubens’s Elevation of the Cross. Study, Examination and Treatment (Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 24), 1992, pp. 61-63. Vynckier 1999/2000 Jozef Vynckier, ‘Étude dendrochronologique de quelques panneaux de l’Agneau Mystique de van Eyck/Dendrochronologisch onderzoek van enkele panelen uit van Eyck’s retabel van het Lam Gods’, in Bulletin de l’Institut royal du Patrimoine artistique/Bulletin van het Koninklijk Instituut voor het Kunstpatrimonium, 28, 1999/2000, pp. 237-240. Waagen 1824 Gustav Friedrich Waagen, ‘Ueber das von den Brüdern Hubert und Johann van Eyck ausgeführte Altargemälde’, in Kunst-Blatt, 5, 1824, no. 23, pp. 89-91, no. 24, pp. 93-96, no. 25. pp. 97-100, no. 26, pp. 103104, no. 27. Pp. 105-108. Waagen 1830 Gustav Friedrich Waagen, Verzeichnis der Gemälde-Sammlung des Königlichen Museums, Berlin, 1830.

Washington/Antwerp 2006 John Oliver Hand, Catherine A. Metzger and Ron Spronk, Prayers and Portraits. Unfolding the Netherlandish Diptych, exh. cat., Washington (National Gallery of Art)/Antwerp (Koninklijk Museum voor Schone Kunsten), 2006. Waterschoot 1966 W. Waterschoot, ‘Lucas de Heere en Marcus van Vaernewijck voor het Lam Gods’, in Jaarboek Koninklijke Soevereine Hoofdkamer van Retorica ‘De Fonteine’ te Gent, 16, 1966, pp. 109-118. Watkins 1997 Jane Watkins (ed.), Jan van Eyck: Two Paintings of Saint Francis Receiving the Stigmata, Philadelphia, 1997. Watrous 1957 J. Watrous, The Craft of Old Master Drawings, Madison, 1957. Ważny 2002 Tomasz Ważny, ‘Baltic Timber in Western Europe : An Exciting Dendrochronological Question’, in Dendrochronologia, 20/3, 2002, pp. 313320. Ważny 2005 Tomasz Ważny, ‘The Origin, Assortments and Transport of Baltic Timber: Historic-Dendrochronological Evidence’, in Carl Van de Velde et al. (eds), Constructing Wooden Images, Proceedings of the Symposium held in Brussels, 25-26 October 2002, Brussels, 2005, pp. 115-126. Ważny, Eckstein 1991 Tomasz Ważny and Dieter Eckstein, ‘The Dendrochronological Signal of Oak (Quercus spp.) in Poland’, in Dendrochronologia, 9, 1991, pp. 35-49. Weale 1864-1865 W. H. James Weale, ‘Inventaires du trésor de la collégiale de Saint Donatien à Bruges 1347-1539’, in Le Beffroi. Arts, Heraldique, Archéologie, 2, 1864-1865, pp. 9-30, 104-138.

594 Weale 1886a W. H. James Weale, ‘Jan van Eyck at the Academy Old Masters Exhibition’, in The Times (London), 3 February 1886, p. 7. Weale 1886b W. H. James Weale, ‘The Royal Academy (First Notice)’, in The Athenaeum, January 9, 1886, p. 73. Weale 1886c W. H. James Weale, ‘The Royal Academy (Fifth Notice)’, in The Athenaeum, February 20, 1886, p. 269. Weale 1903 W. H. James Weale, Hubert and John van Eyck, London, 1903. Weale 1904 W. H. James Weale, ‘Portraits by John van Eyck in the Vienna Gallery’, in The Burlington Magazine, 5, 1904, pp. 190-197. Weale 1908 W. H. James Weale, Hubert and John van Eyck, Their Life and Work, London/New York, 1908. Weale 1909 W. H. J. Weale, ‘A Fifteenth-Century Painting by a Follower of John van Eyck’, in The Burlington Magazine, XV, 1909, p. 49, f. Weale et al. 1863-1873 W. H. J. Weale, Le Beffroi. Arts Heraldique Archeologie, 4 vols, Bruges, 1863-1873. Weale, Brockwell 1912 W. H. James Weale and Maurice W. Brockwell, The Van Eycks and their Art, London, 1912. Weaver 1950 James R.H. Weaver, George L. Stout and Paul Coremans, ‘The Weaver Report on the Cleaning of Pictures in the National Gallery’, in Museum International, 3, 1950, pp. 113-176. Wechsler 1987 Bettina Wechsler, ‘Cleaning Controversy. Zur Diskussion der Gemäldereinigung in England von 1946-1963’, in Zeitschrift für Kunsttechnologie und Konservierung, Jahrgang 1, Heft 2, 1987, pp. 88-132. Wehle, Salinger 1947 Harry B. Wehle and Margaret Salinger, A Catalogue of Early Flemish, Dutch, and German Paintings, New York, 1947. Weis 1974 E. Weis, ‘Johannes der Täufer’, in W. Braunfels (ed.), Lexikon der christlichen Ikonografie. Ikonografie der Heiligen, 7. Innozenz bis Melchisedech, Rome, 1974, pp. 164-190. Weiss 1955 Roberto Weiss, ‘Jan van Eyck’s “Albergati” Portrait’, in The Burlington Magazine, 97, 1955, pp. 145-147.

Werner 2007 Elke Anna Werner, ‘The Veil of Venus: A Metaphor of Seeing in Lucas Cranach the Elder’, in Bodo Brinkman, Cranach, exh. cat., London (Royal Academy of Arts), 2007, pp. 89-109. Werner-Tschochner 1974 F. Werner-Tschochner, ‘Jodok(us) (Iudocus, Josse, Jost)’, in Engelbert Kirschbaum, Wolfgang Braunsfels (eds), Lexikon der christlichen Ikonographie, 7. Ikonographie der Heiligen. Innozenz bis Melchisedech, Rome/ Vienna, 1974, cols 70-71. White 1995 Raymond White, ‘Analyses of Norwegian Medieval Paint Media. A Preliminary Report’, in Magne Malmanger, Lazlo Berczelly and Signe Fuglesang (eds), Norwegian Medieval Altar Frontals and Related Material. Papers from the Conference in Oslo 16th to 19th December 1989, ACTA 11, Rome, 1995, pp. 127-135. White 2000 Raymond White, ‘Van Eyck’s Technique: The Myth and the Reality II’, in Susan Foister, Sue Jones and Delphine Cool (eds), Investigating Jan van Eyck, Turnhout, 2000, pp. 101-105. White, Kirby 2006 Raymond White and Jo Kirby, ‘Some Observations on the Binder and Dyestuff Composition of Glaze Paints in Early European Panel Paintings’, in Medieval Painting in Northern Europe: Techniques, Analysis, Art History, 2006, London, pp. 215-222. White, Pilc 1993 Raymond White and Jennifer Pilc, ‘Analyses of Paint Media’, in National Gallery Technical Bulletin, 14, 1993, pp. 86-94. Wieck 1991 Roger S. Wieck, ‘The Savoy Hours and its Impact on Jean, duc de Berry’, in Beinecke Studies in Early Manuscripts, The Yale University Gazette, 66, 1991, pp.159-180. Wieck 2005 Roger S Wieck, ‘Bibliophilic Jealousy and the Manuscript Patronage of Jean, Duc de Berry’, in Rob Dückers and Pieter Roelofs (eds), The Limbourg Brothers, Nijmegen Masters at the French Court 1400-1416, exh. cat., Nijmegen (Museum het Valkhof), 2005, pp. 120-133. Wisse 1998 Jacob Wisse, ‘Distinguishing Between Bouts and Stuerbout as Official City Painters’, in Maurits Smeyers and Katharina Smeyers (ed.), Dirk Bouts (ca. 1410-1475) een Vlaams primitief in Leuven, exh. cat., Leuven, 1998, pp. 19-33. Wolfthal 2006 Diane Wolfthal, ‘Renaissance Fingerpainting’, in Hélène Verougstraete and Jacqueline Couvert (eds), Le dessin sous-jacent dans la peinture. La peinture ancienne et ses procédés: copies, répliques, pastiches, Colloque XV, Bruges, 11-13 septembre 2003, Bruges, 2006, pp. 91-97.

Weiss 1956 Roberto Weiss, ‘Jan van Eyck and the Italians’, in Italian Studies, 11, 1956, pp. 1-15.

Wösten 2010 Lidwien Wösten, Blotting: Onderzoek naar het gebruik van blotting in 15de en 16de eeuwse schilderijen van de Leidse School, Masters thesis, Department of Conservation and Restoration, University of Amsterdam, 2010.

Weitz 2012 Armelle Weitz, La provenance du bois d’œuvre en région BruxellesCapitale: première approche de la question à travers l’étude et la révision des données du laboratoire de dendrochronologie de l’ULg/CEA, Master 2 diss., University of Liège, 2012.

Woudhuysen-Keller 1995 Renate Woudhuysen-Keller, ‘Aspects of Painting Technique in the Use of Verdigris and Copper Resinate’, in A. Wallert, E. Hermensand M. Peek (eds), Historical Painting Techniques, Materials and Studio Practice, Los Angeles, 1995, pp. 65-69.

595 Wright 1912 A.R. Wright, ‘The Thunderweapon in Religion and Folklore. A Study in Comparative Archaeology’, in Folklore, 23, 1, March 1912, pp. 122-123.

Ziloty 1941 Alexandre Ziloty, La Découverte de Jean van Eyck et l’évolution du procédé de la peinture à l’huile du moyen age à nos jours, Paris, 1941.

Yiu 2006 Yvonne Yiu, ‘Hinging Past and Present: Diptych Variants of Jan van Eyck’s Virgin in the Church’, in John Oliver Hand and Ron Spronk (eds), Essays in Context. Unfolding the Netherlandish Diptych, Cambridge/ New Haven/London, 2006, pp. 110-123.

Zoppi et al. 2010 A. Zoppi, C. Lofrumento, N.F.C. Mendes and E.M. Castellucci, ‘Metal Oxalates in Paints: a Raman Investigation on the Relative Reactivities of Different Pigments to Oxalic Acid Solutions’, in Analytical and Bioanalytical Chemistry, 397, 2010, pp. 841-849.

Zarncke 1876 Friedrich Zarncke, Der Priester Johannes, Abhandlungen der philologischhistorischen Classe der Koeniglichen Sachsischen Gesellschaft der Wissenschaften, vols 7-8, Leipzig, 1876.

Zunde 1998-1999 Maris Zunde, ‘Timber Export from Old Riga and its Impact on Dendrochronological Dating in Europe’, in Dendrochronologia, 16-17, 19981999, pp. 119-130.

Zehnder 1990 Frank Günter Zehnder, Katalog der Altkölner Malerei (Kataloge des Wallraf-Richartz-Museums, XI), Cologne, 1990.

Zunde 2011 Maris Zunde, ‘New Dendrochronological, Historical and Archaeological Evidence of Long-Distance Floating of Timbers to Riga’, in Pascale Fraiture (ed.), Tree Rings, Art, Archaeology. Proceedings of the conference, Brussels, Royal Institute for Cultural Heritage, 10-12 February 2010 (Scientia Artis, 7), Brussels, 2011, pp. 125-135.

Zeman 2002 Georg Zeman, ‘Eine altniederländische Silberstiftzeichnung und ihre Bedeutung für Jan van Eycks Dresdner Marienaltar’, in Zeitschrift für Kunstgeschichte, vol. 65, 2002, pp. 91-104.

Illustration Credits

Every effort has been made to contact the copyright holders of the illustrations in this volume. Copyright holders whom we have been unable to reach or to whom inaccurate acknowledgement has been made are invited to contact the respective author.

Antwerp, KMSKA: Figs 34.3; 34.6c. Artothek: Fig. 26.1 AXES research group, University of Antwerp: Figs 30.6; 30.7; 30.8; 30.10

24.2; 24.3; 24.4; 24.5; 24.6; 24.7; 24.8; 24.9; 24.10; 24.11; 24.12; 24.13; 26.3a; 27.6; 27.8a-b; 28.5a-f; 28.6a-b; 28.7; 28.8a-c Dresden, Hochschule für Bildende Künste Dresden: Figs 22.4 (Bertram Lorenz); 22.6 (Bertram Lorenz) Dresden, Staatliche Kunstsammlungen Dresden: Figs 20.2a; 22.1 (Hans Peter Klut/Elke Estel); 22.2 (Konstanze Krüger/Christoph Schölzel); 22.3(Gerhard Rüger); 22.9 (Konstanze Krüger/Christoph Schölzel); 22.10 (Konstanze Krüger/Christoph Schölzel); 22.11 (Hans Peter Klut/Elke Estel); 24.16b; 24.16d; 32.1; 32.2 Dwarswaard, Henry: Figs 10.7a-b Edelmann, U.: Fig. 26.1

Baltimore, The Walters Art Gallery: Figs 32.4; 32.8

Fiorello, M.: Figs 32.3; 32.5; 32.6, 32.7

Barcelona, Museu Nacional d’Art de Catalunya: Figs 31.3; 31.4; 31.5; 31.6; 31.7a-b

Frankfurt, Städel Museum: Fig. 26.1 Ghent, Stad Gent – Dienst Stadsarcheologie: Fig. 36.4

Berlin, Staatliche Museen zu Berlin, Gemäldegalerie: Figs 12.2 (photo Schwarz); 12.3 (photo Schulz); 12.4 (photo Schulz); 12.5 (photo Schulz); 12.6 (photo Schulz); 12.8 (photo Schulz); 34.2; 34.6b Berlin, Staatliche Museen zu Berlin, Kupferstichkabinett: Fig. 12.7 (photo Schulz) Birmingham, Birmingham Art Gallery: Fig. 25.4 Boswank, Herbert, Staatliche Kunstsammlungen Dresden: Fig. 15.9 Bruges, Musea Brugge, Groeningemuseum: Figs 15.1; 27.1 Brussels, KIK-IRPA: Figs 1.8 (LB 3080); 1.11 (X058728L);1.12a (detail of X058837L); 1.13 (LB030040); 1.16a (Photo atelier, 2010); 1.16b (LB 3008); 1.19 (LB3054); 1.2 (LB03062); 1.20 (LB5615); 1.21 (KIK-IRPA Archives); 1.22 (LB3086); 1.23 (KIK-IRPA Archives); 1.24 (LB3056); 2.4; 4.1; 4.3 (KIK-IRPA Archives); 4.4 (Archives Coremans 1949-1965, no. 2117); 4.5 (KIK-IRPA Archives); 4.7 (photos: labo); 5.1 (dendro. lab.); 5.2a-c (dendro. lab.); 5.3 (dendro. lab.); 5.4a-d (dendro. lab.); 5.5a (Archive); 5.6a-c (dendro. lab.); 5.7a (dendro. lab.); 5.7b (dendro. lab.); 5.8a-b (dendro. lab.); 5.9a (dendro. lab.); 5.9b (dendro. lab.); 6.1; 6.2; 6.3; 6.4; 6.5; 6.6; 6.7; 6.8; 6.9; 6.10; 6.11; 8.10; 8.8; 8.9a; 8.9b; 15.14; 20.2c; 20.2e; 20.7a-b; 20.8; 20.9; 25.3; 25.7; 26.3b (Noëlle Streeton); 26.3c (Noëlle Streeton); 27.1; 27.2; 27.5; 27.6; 33.1; 33.2; 33.3; 33.4; 33.5; 33.7; 33.8; 35.1; 35.2; 35.3; 35.4; 35.5 (C11051); 35.6 (C1104); 35.7 (KN8862); 36.3; 36.6; 36.7; 36.8; 36.9 Brussels, KIK-IRPA Centre for the Study of the Flemish Primitives: Fig. 34.4 Closer to Van Eyck (http://closertovaneyck.kikirpa.be): Figs 1.12b; 1.14a-b; 1.15a-c; 1.17; 2.1; 2.2; 3.1; 3.2; 3.4; 3.6; 3.7; 3.8; 3.9; 7.1; 7.2; 7.3a-b; 7.4; 7.7; 8.1; 8.12a-c; 8.2a-b; 8.3a-b; 8.4a-b; 8.5a-b; 8.6a-b; 10.1; 10.2; 10.3; 10.4; 10.5; 10.6; 15.10a-b; 20.11; 20.12; 20.4a-b; 20.5;

Ghent, St Bavo’s Cathedral, Lukasweb.be – Art in Flanders vzw – photo KIK-IRPA: Plates A, B, Figs 1.6, 1.11, 1.12a, 6.1. Heidelberg, Universitätsbibliothek Heidelberg: Fig. 2.3 Karpinski, Jürgen, Dresden: Fig. 22.5 Kemperdick, Stephan: Figs 21.2; 21.5; 21.8 Liège, Musée du Grand Curtius: Figs 34.1; 34.5; 34.6a; 34.7a London, The National Gallery: Figs 8.7a-b; 8.9c; 13.1; 13.10; 13.11; 13.12; 13.13; 13.2; 13.3; 13.4; 13.5; 13.6; 13.7; 13.8; 13.9; 15.1; 15.2; 15.3; 15.4; 15.5; 15.6; 15.7; 15.8; 16.1; 16.2; 16.3; 16.4; 16.5; 16.6; 16.7; 16.8; 16.9; 18.1; 18.2; 18.3; 18.4; 18.5; 18.6; 18.7; 18.8; 20.1; 20.2b; 20.3; 24.14; 24.15 Lukasweb.be – Art in Flanders vzw: Figs 15.13 (Hugo Martens); 20.2d (Hugo Martens); 34.3; 34.6 Madrid, Museo Nacional del Prado: Figs 2.6a; 2.7a-b; 17.1; 17.2; 17.3; 20.10a-b Madrid, Museo Thyssen-Bornemisza: Figs 24.16a; 24.16c Munich, Alte Pinakothek: Fig. 7.6 New York, The Metropolitan Museum of Art: Figs 14.1; 14.2; 14.3; 14.4; 14.5a-b; 14.6a-b; 14.7a-c; 14.8a-b; 21.7; 24.1a Nuremberg, Germanisches Nationalmuseum Nürnberg: Fig. 22.8 Onroerend Erfgoed: Figs 29.1 (K. Vandevorst); 29.2 (K. Vandevorst); 29.3 (K. Vandevorst); 29.5 (K. Vandevorst); 29.6 (K. Vandevorst); 29.7 (K. Vandevorst); 29.8 (K. Vandevorst); 29.1 (K. Vandevorst); 29.4a-b (Brigitte De Schaepmeester)

598 Philadelphia, Philadelphia Museum of Art, John G. Johnson Collection: Figs 25.2; 25.6; 26.2

Turin, Galleria Sabauda: Fig. 25.1 Valencia, Museo de Bellas Artes: Fig. 31.1

Réunion des musées nationaux / Musée du Louvre Paris: Figs 21.6 (Hervé Lewandowski); 27.3a; 30.1

Van Asperen de Boer, J.R.J.: Fig. 27.3b

Smith, Jamie: Fig. 25.8

Van Duijn, Esther: Figs 27.5 (Closer to Van Eyck. http://closertovaneyck.kikirpa.be); 27.7; 27.8c-f; 27.9

Spronk, Ron: Figs 15.11 (IRR courtesy of Kunsthistorisches Museum Wien, Gemäldegalerie /Queen’s University); 15.12 (IRR courtesy of Kunsthistorisches Museum Wien, Gemäldegalerie /Queen’s University) Streeton, Noëlle: Figs 26.3a-d (printed with permission of Joseph Rishel and Philadelphia Museum of Art); 26.4a-d (printed with permission of Joseph Rishel and Philadelphia Museum of Art); 26.5 (printed with permission of Jochen Sander and the Städel Museum) The Hague, Netherlands Institute for Art History (RKD): Figs 1.9 (Friedländer Archive, no. 0000215246); 1.10 (Friedländer Archive, no. 0000215066); 27.3b; 30.5 The Hague, Koninklijke Bibliotheek: Fig. 25.5 The Impact of Oil: Figs 2.5b; 2.6b-c; 28.1a-b (Groeningemuseum); 28.2a (Groeningemuseum); 28.2b (Washington, National Gallery); 28.3 a-c (Groeningemuseum); 28.4 a-c (Groeningemuseum)

Van Grevenstein-Kruse, Anne: Figs 1.1; 1.4 Verburg, Adri: Fig. 34.7b Verroken, Erik: Fig. 4.6 Vienna, Albertina: Fig 32.6 Vienna, Kunsthistorisches Museum: Figs 21.3; 21.4 Wallert, Arie; Figs 37.1; 37.2; 37.3a-b; 37.4a-b; 37.5a-b; 37.6; 37.7 Washington, National Gallery of Art: Figs 19.1; 19.2a-b; 19.3; 19.4ad; 19.5a-d; 19.6; 19.7; 20.6; 27.4; 28.2b Wittenberg, Jörg: Fig. 22.7