Materials of Manufacture: The choice of materials in the working of bone and antler in northern and central Europe during the first millennium AD 9781841715599, 9781407326016

The original impetus for this volume came from a small and informal conference held at the British Museum in 1997. The c

195 90 11MB

English Pages [92] Year 2003

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Materials of Manufacture: The choice of materials in the working of bone and antler in northern and central Europe during the first millennium AD
 9781841715599, 9781407326016

Table of contents :
Front Cover
Title Page
Copyright
Contents
The Contributors
List of Figures
List of Tables
Acknowledgements
Introduction
ELK ANTLER AS A MATERIAL OF MANUFACTURE: Finds from Late Republican/Early Imperial ‘old Virunum’ on the Magdalensberg in Carinthia, southern Austria
THE IDENTIFICATION OF FALLOW DEER (DAMA DAMA) REMAINS FROM ROMAN MONKTON, THE ISLE OF THANET, KENT
RECYCLING WORKED BONE IN PANNONIA: Data on the curation of workshop debitage and worn/damaged objects in the Roman bone manufacturing industry
UN ATELIER DE PRODUCTION DE PEIGNES EN BOIS DE CERF DE LA FIN DE L’ANTIQUITÉ TARDIVE À SAINT-CLAIR-SUR-EPTE (EURE)
A LESSER MATERIAL: The Working of Roe Deer Antler in England during the Anglo-Saxon Period
BIAS FROM BONE WORKING AT MIDDLE SAXON HAMWIC, SOUTHAMPTON
LATE SAXON WORKED ANTLER WASTE FROM HOLY ROOD, SOUTHAMPTON (SOU106)

Citation preview

BAR S1193 2003

Materials of Manufacture The choice of materials in the working of bone and antler in northern and central Europe during the first millennium AD Edited by

RIDDLER (Ed): MATERIALS OF MANUFACTURE

B A R

Ian Riddler

BAR International Series 1193 2003

Materials of Manufacture The choice of materials in the working of bone and antler in northern and central Europe during the first millennium AD Edited by

Ian Riddler

BAR International Series 1193 2003

Published in 2016 by BAR Publishing, Oxford BAR International Series 1193 Materials of Manufacture © The editor and contributors severally and the Publisher 2003 Typesetting and layout: Darko Jerko The authors' moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781841715599 paperback ISBN 9781407326016 e-format DOI https://doi.org/10.30861/9781841715599 A catalogue record for this book is available from the British Library BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 1974 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd / Hadrian Books Ltd, the Series principal publisher, in 2003. This present volume is published by BAR Publishing, 2016.

BAR

PUBLISHING BAR titles are available from:

E MAIL P HONE F AX

BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

Contents The Contributors ..................................................................................................... ii List of Figures ......................................................................................................... iii List of Tables ........................................................................................................... v Acknowledgements ................................................................................................ vi Introduction ........................................................................................................... vii Kordula GOSTENČNIK ........................................................................................... 1 Elk Antler as a Material of Manufacture. Finds from Late Republican/Early Imperial ‘Old Virunum’ on the Magdalensberg in Carinthia, southern Austria Robin BENDREY ................................................................................................... 15 The Identification of Fallow Deer (Dama dama) Remains from Roman Monkton, the Isle of Thanet, Kent Maria T. BÍRÓ ........................................................................................................ 19 Recycling Worked Bone in Pannonia: Data on the Curation of Workshop Debitage and Worn/Damaged Objects in the Roman Bone Manufacturing Industry Annick THUET ...................................................................................................... 25 Un atelier de peignes en bois de cerf de la fin de l’Antiquité à Saint-Clair-sur-Epte (Eure) Ian RIDDLER ........................................................................................................ 41 A Lesser Material: the Working of Roe Deer Antler in Anglo-Saxon England Jennifer BOURDILLON ......................................................................................... 49 Bias from Boneworking at Middle Saxon Hamwic, Southampton, England Ian RIDDLER and Nicola TRZASKA-NARTOWSKI .............................................. 65 Late Saxon Worked Antler Waste from Holy Rood, Southampton (SOU106)

i

The Contributors Robin Bendrey 16 Newburgh Street, Winchester SO23 8UY England Maria T. Bíró Institute of Archaeological Sciences, Lorand Etvos University, H-1088 Budapest; Budapest V, Aranykez u7, 1052 Hungary Jennifer Bourdillon 44, Pettinger Gardens, Southampton, Hampshire SO17 2WL England Annick Thuet Base INRAP Nord-Picardie, 518, rue Saint Fuscien, 80000 Amiens, France; 6, rue du Hamelet, 80430 Lafresnoye, France Kordula Gostenčnik Rabenstein 12, A-9473 Lavamünd, Austria Ian Riddler and Nicola Trzaska-Nartowski Tatra, Diddies Road, Stratton, North Cornwall, EX23 9DW England

ii

List of Figures Kordula Gostenčnik Figures 1.1 Magdalensberg. General outline of the excavations (after Piccottini) 1.2 Current distribution of elks in Northern Europe (after MacGregor) 1.3 An elk in comparison to a red deer stag (after Catalogue München) 1.4 Diptych-panel of Late Antiquity from Liverpool Museum, venatio with five elks (after Delbrueck) 1.5 Wickerwork vessels from the Kertch-region (after Rieth and Vaulina et al) Plates Pl. 1 Catalogue N°s 1-2; scale 1:1 Pl. 2 Catalogue N°s 3-4; scale 1:1 Pl. 3 Catalogue N°. 5; scale 1:1 Pl. 4 Catalogue N°. 1 Pl. 5 Catalogue N°. 2 Pl. 6 Catalogue N°. 3 Pl. 7 Catalogue N°. 4 Pl. 8 Catalogue N°. 5 Pl. 9 Addendum Catalogue N. 6 Maria T. Bíró 3.1 Cut off red deer antler tines: A: Tác - Gorsium; B: Dunaújváros/Dunapentele Intercisa and C: rope-tying toggle: Esztergom - Solva 3.2 Awl (Tác - Gorsium) 3.3 Drill handle (Szombathely - Savaria) 3.4 Amulets made from antler burrs: A: Dunaújváros/Dunapentele - Intercisa; B: Szombathely - Savaria 3.5 Fragment of an Indian comb (Tác - Gorsium) 3.6 Lid of ivory casket (A: Bégram); Ivory carving (B: Bégram) 3.7 Heavily worn bird-shaped pin (Tác - Gorsium) 3.8 Renewed bird-shaped pin (Tác - Gorsium) 3.9 Renewed bird-shaped pin (Dunaújváros/Dunapentele - Intercisa) Annick Thuet 4.1 Terminologie relative au bois de cerf, d’après Billamboz 1979, 96 4.2 Différentes traces de découpes sur les meules 4.3 Bois de massacre nº 6, avec les traces de découpe 4.4 Découpes 12 à 17 repositionnés sur un bois 4.5 Intersection merrain-andouiller nº 18 4.6 Découpes sur intersection merrain-andouiller 4.7 Découpes originales sur trois fragments iii

4.8 Intersection nº 48 4.9 Intersection nº 56 4.10 Intersection nº 70 4.11 Andouiller fendu longitudinalement 4.12 Tronçon d’andouiller nº 136 4.13 Tronçon d’andouiller débité longitudinalement 4.14 Plaque façonnée sur une face 4.15 Plaque façonnée sur une face 4.16 Plaque dont les deux faces sont travaillées 4.17 Plaque dont les deux faces sont travaillées 4.18 Les Objets semi-finis et complet : A Plaque de forme trapézoidale nº 446 B Plaque d’extrémité de peigne à double denture nº 447 C Plaquette d’extrémité de peigne à double denture nº 448 D Plaquette décorée d’extrémité de peigne à double denture nº 449 E Ebauche d’épingle à cheveux en os (?) nº 453 F Plaque de traverse brisée nº 451 G Deux traverses brisées au niveau des rivets, enserrant une plaquette de peigne à double denture nº 452 H Plaque formant traverse nº 450 4.19 Chaîne opératoire de débitage sur bois de cerf Ian Riddler 5.1 A: Roe Deer Antler Implement from Barton Court Farm, Oxfordshire (after de Hoog 1984). B: Roe Deer Antler Stamp from Hamwic 5.2 A: Roe Deer Antler Implement from Ipswich Buttermarket B: Roe Deer Antler Implement from Westbury, Buckinghamshire (after Ivens, Busby and Shepherd 1995) 5.3 Roe Deer Antler from Dover, Townwall Street (drawn by Beverley Leader) Jennifer Bourdillon 6.1 Cow Metatarsus Distal Breadth: Six Dials Studies 6.2 Cow Metatarsus Distal Breadth: Site 32, all contexts 6.3 Cow Metacarpus Distal Breadth: Six Dials Studies 6.4 Cow Metacarpus Distal Breadth: Site 32, all contexts 6.5 Cow Tibia Distal Breadth: Six Dials Studies 6.6 Cow Tibia Distal Breadth: Site 32, all contexts 6.7 Cow Tibia Distal Breadth: Six Dials late contexts Ian Riddler and Nicola Trzaska-Nartowski 7.1 Distribution of Waste from Feature 96, located on a Red Deer Antler 7.2 Measurement Points for Antler Burrs (after Müller-Using 1953) 7.3 Antler Burr Circumferences from Hamwic and Holy Rood, Southampton 7.4 First Diameter Measurement, Holy Rood and Hamwic 7.5 Second Diameter Measurement, Holy Rood and Hamwic 7.6 Third Diameter Measurement, Holy Rood and Hamwic 7.7 Fourth Diameter Measurement, Holy Rood and Hamwic 7.8 Antler Burr Circumferences for Hamwic, Holy Rood and West Quay, Southampton 7.9 A: Modified Burr from Feature 95W; B: Section of beam with incised scoring mark, from Feature 52 iv

7.10 Antler Tine Implements: A: Holy Rood; B: Ipswich, Buttermarket; C: Fragmentary Wedge from Holy Rood 7.11 Unfinished Antler Objects from Holy Rood: A Unfinished Tooth Segment blank (75); B Antler strip with decoration at one end (77); C Unfinished comb connecting plate (76); D undecorated broad antler strip (48)

List of Tables Robin Bendrey 1 Scoring of Lister’s (1996) identification characters for the fallow deer remains, and measurements 2 Distribution and Date of allow deer finds from Roman Monkton Jennifer Bourdillon 1 Hamwic, Six Dials. SOU31, Feature 2048. Identified Fragments from Normal Recovery 2 Hamwic, Six Dials. SOU31, Feature 2048. Weights of Identified Mammal Fragments 3 Hamwic, Six Dials. SOU31, Feature 2048. Unidentified Material 4 Hamwic, Six Dials. Unidentified Material, excluding Worked Offcuts 5 Hamwic, Six Dials. Mean Fragment Weights of Cattle, Sheep, Pig and Horse 6 Hamwic, Six Dials. SOU31, Feature 2048. Mean Fragment Weights of Worked Bone Fragments (g/fragment) 7 SOU14. Chapel Road. Worked Bone and Antler Offcuts, by Fragment Count 8 Hamwic, Six Dials. SOU31, Feature 2048. Relative Representation of Wild Species by Fragment Count (excluding Worked Offcuts) 9 Hamwic, Six Dials. SOU31, Feature 2048 and SOU14. Worked Offcuts. Distribution over the Body 10 Hamwic, Six Dials. SOU31, Feature 2048. Relative Representation of the less common domestic species, by fragment count 11 Hamwic, Six Dials. SOU31, Feature 2048. Horse: Distribution over the Body 12 Northern Hamwic. A Selection of Worked Bone Assemblages 13 Some Comparisons of Size Factors 14 Hamwic Cow Radius Proximal Width 15 Hamwic Cow Metacarpus Proximal Width 16 Hamwic Cow Metacarpus Distal Width 17 Hamwic Cow Tibia Distal Width 18 Hamwic Cow Metatarsus Proximal Width 19 Hamwic Cow Metatarsus Distal Width Ian Riddler and Nicola Trzaska-Nartowski 1 Antler Waste from Holy Rood, Southampton 2 Antler Waste from Features 52, 95 and 96 3 Relative Quantities of Antler, by Category 4 Saw Blade Widths from Various Sites

v

Acknowledgements I would like to thank numerous individuals who have helped with the production of this volume. Nicola Trzaska-Nartowski assisted throughout, both with the study of material and with the compilation of the volume. I have benefited throughout from discussions with Robin Bendrey, Maria Bíró, Jennifer Bourdillon, Alice Choyke, Kordula Gostenčnik and Annick Thuet. M. Marc Talon kindly made me aware of Annick Thuet’s work and has helped enormously with a variety of archaeological matters. Will Foster and Beverley Leader advised on illustrations and provided one of those used in the volume. Andrew Russel generously searched archives and provided information about the Holy Rood excavations and Paul Bennett allowed the Monkton fallow deer remains to be included here, in advance of the full publication of that site. Gill Campbell and Polydora Baker of English Heritage kindly allowed Jennifer Bourdillon’s work to be published in this volume, and assisted with its editing. I am very grateful to English Heritage for their support of this project. Keith Parfitt and Barry Corke provided information on their site at Townwall Street, Dover and Dr. Cornelia Becker provided copies of her invaluable work. Dr Ingrid Ulbricht kindly responded to my enquiries over a long period of time. My own interests in bone and antler working began with the study of material from Hamwic, working alongside Jennifer Bourdillon. It is a great pleasure to acknowledge a little of my enormous debt with the publication of her text here, and to respectfully dedicate this small volume to her.

vi

Introduction The original impetus for this volume came from a small and informal conference held at the British Museum in 1997. The conference, in itself, was an inaugural meeting of the Worked Bone Research Group, an organisation set up shortly before to allow specialists in the study of worked bone, antler, ivory and horn to meet together and discuss assemblages, themes and work in progress. It was not intended that any papers from this first meeting should be published. Several of the participants, including Sabine Deschler-Erb, were in the process of finalising monograph publications, whilst others, like Bea de Cupere, Catherine Hills, Alice Choyke and Neil Stratford, had recently published their work. However, the interest both in the subject and in the dissemination of information within the Worked Bone Research Group exceeded initial expectations. It led to a second meeting, held in Budapest, and to an extensive monograph publication (Choyke and Bartosiewicz 2001). This small volume arose as a by-product of the same interest. Several texts were presented to the editor with a view to establishing a place for their publication. In collating these together, it became clear that a general theme of materials and their use could be established. Equally, a specific, if fairly broad, time frame of the first millennium AD could also be defined. With these parameters determined, it was possible to enlarge the volume a little by commissioning a few extra texts and updating several others, and to bring all of these together into a small monograph. Northern and central Europe witnessed enormous changes in the transition from the Roman to the medieval world across the first millennium AD. In one sense, these were two different worlds and two technologies, each markedly distinct, but with some elements in common. This volume pursues some of the common elements, as well as noting change over time. The first four papers are concerned with the Roman world, and with different aspects of raw material supply. Kordula Gostenčnik describes a number of objects from Magdalensberg, which are made from elk antler. The settlement of Virinum on the Magdalensberg was founded in the middle of the first century BC as a trading post for Italian merchants. They brought with them a new interest in the working of bone and antler (Gostenčnik 1996; 2001 and forthcoming). A new technology, involving the use of the lathe, also came to the area. Amongst the objects from the settlement are four discs, a tablet and waste material made from elk antler. In general, the antlers of elk were amongst the most favoured by bone and antler workers, but they were not available to all. Elk antler was readily available in Carinthia during the prehistoric period, but the animal had disappeared from the region by the end of the medieval period. It was not extensively used in central Europe at this time, however. Indeed, it is virtually absent from Augst and occurs only in small quantities elsewhere. In part, this may be a reflection of a preference for the working of bone, rather than antler (Deschler-Erb 1998), particularly during the early centuries of the first millennium AD. Equally, the Magdalensberg material shows that different raw materials were appreciated for their own, specific properties. The large palmate expanses of elk antler rendered it ideal for large discs and the bases of wickerwork baskets. Very few sections of animal skeletons would have been suitable for the same purpose. Indeed, this appreciation and understanding of the capabilities of the raw material permeates most of the contributions of this volume. The excavated portion of the Roman settlement at Monkton, Thanet, Kent, may have formed part of a large villa estate. The site is unusual in a number of respects. Almost all of the excavated structures are sunk into the ground; most are of first or second century date. Evidence from the faunal remains assemblages suggests that one of the occupations practiced there was the breeding vii

of horses (Bendrey forthcoming). There is also unequivocal evidence for the presence of fallow deer, represented both by antler and by post-cranial material. The assemblage is small but it is of considerable significance in terms of the vexed question of the distribution of fallow deer in northern Europe during the Roman period. As Bendrey notes, there is very little evidence for fallow deer in Roman Britain but recent excavations, at Monkton and elsewhere, have now firmly established its presence in southern and eastern England. At the same time, fallow deer are seen only in very small numbers and we now need to seek an explanation for this circumstance. The quantity of fallow deer from northern Europe has not risen appreciably from the days of earlier reviews. Does this, for example, represent the movement of live animals, or dead carcasses ? Bendrey notes that the faunal remains may represent trade in salted meat, although there are several metapodial bones within the Monkton assemblage, and these are not meat-bearing. Faced with a similarly small assemblage of fallow deer from Augst, Elizabeth Schmid noted the proximity of her material to the Temple area of the settlement, and argued that fallow deer were exotic animals imported from the Mediterranean for display purposes (Schmid 1965). The situation at Monkton is a little different and the animal remains come from a distinctly rural environment, located on the lower fringes of a series of extensive enclosures. Could the area have served as a Roman deer park ? Was the keeping of horses and deer the rationale behind the enclosures ? Robin Bendrey’s work raises interesting questions about the economic situation lying behind the faunal remains assemblage. Maria Bíró looks at raw material supply from an entirely different perspective. Building on her work on Roman bone and antler objects (Bíró 1994 and forthcoming) she examines a number of objects from late Roman contexts in Pannonia. She notes that the fourth century preference for objects in bone of the same types as those produced previously in metals may reflect the general impoverishment of the province. ‘A spoon carved from bone may be just as beautiful as a bronze one; however, its life span is significantly shorter’. This suggests that a deliberate choice was made for a specific object type, but it was tempered by a compromise over the material type. At the same time, the re-use of discarded bone objects increased, as did the utilisation of forms of waste (like antler burrs), which had previously been cast aside. In some cases objects were transformed and given new functions, most spectacularly with the extraordinary ivory comb from Tác. With the difficulties of raw material supply, the curation and transformation of existing pieces came to the fore, preserving vestiges of a material culture that was swiftly disappearing. The fifth century assemblage from Saint-Clair-sur-Epte examined by Annick Thuet is one of the most significant groups of antler of that century to have come from northern Europe. Indeed, it is one of very few late Antique to Merovingian assemblages and it forms an important counterpoint to Petitjean’s earlier study of contemporary comb forms and their manufacture (Petitjean 1995). Despite the pioneering work carried out by Cnotliwy, Ulbricht, MacGregor, Ambrosiani and Christophersen, amongst others, relatively few post-Roman assemblages of bone and antler have been fully published in the last twenty years. The techniques utilised at Haithabu in the ninth and tenth centuries may have been similar those seen at Saint-Clair-sur-Epte during the fifth century but as yet it remains difficult to define regional distinctions in bone and antler working at this time, even on a broad scale. The distinctions exist in the range and form of bone and antler artefacts but not, as yet, in the waste materials. The impressive sequence of monographs and articles on Roman boneworking in Europe may now allow such distinctions in working practices to be identified and understood for that period (Béal 1983; 1984; Bíró 1994; von Carnap-Bornheim 1994; Petkovic 1995; Mikler 1997; Obmann 1997; Deschler-Erb 1998; Gostenčnik forthcoming). For the fifth to ninth centuries the quantity of reports is now growing and material choices, techniques and products are becoming clearer. Thuet’s work emphasises the point that these assemblages need to be fully published. The raw material at Saint-Clair-sur-Epte is limited to red deer antler but the presence of a single bone pin hints at further working practices which otherwise are not represented within the surviving assemblage. Much less is known about bone and antler working in England at this time (Riddler 1996) and Saint-Clair-sur-Epte, alongside Mühlberg (Teichert 1983; Aufleger 1996), provides the best information yet available for northern Europe in the fifth and sixth centuries. The techniques of study are also important with these assemblages. Thuet has undertaken a thorough analysis of her material, which is viewed from an archaeozoological viewpoint, as well as incorporating the methods of a finds specialist. The requirement for these differing, comprehensive viewpoints has been stated on several occasions (Ulbricht 1980; Becker 1993). viii

Within studies of bone and antler waste elements of technology, typology and archaeozoology are combined to provide an overall image of each production episode. Complementary studies are concerned with specific object types, including in this case a study of objects made from roe deer antler. Roe deer remain widely distributed across Europe but their antlers are relatively small, in comparison with red deer or elk. It is not surprising, therefore, that they did not form a common raw material source. Across the period from the fifth to the twelfth century their use was more opportunistic perhaps, than planned. At the same time, when the raw material became available, it could be put to good use, as the evidence from Haithabu in particular indicates. The elaborate sequence of processes established for the dismemberment of red deer antler at Saint-Clair-sur-Epte, which recalls that indicated in detail by Ingrid Ulbricht (1978, 2530), is scarcely repeated on a small scale for roe deer antler. In contrast, objects of this material overtly display their origins, with the pronounced pearling of the antler often retained, alongside the shape of the beam and accompanying tines. Yet waste and unused roe deer antlers occur alongside deposits of red deer antler, suggesting that they were passed to bone and antler workers and were not widely available. Jennifer Bourdillon’s contribution to this volume originally formed a small part of a monograph on the environmental archaeology of Anglo-Saxon and medieval Southampton. That monograph has yet to be published, but it is hoped that this may happen in the near future. The first five contributions in this volume are concerned principally with the working of antler. From prehistoric times onwards bone was also worked, however, as well as antler. Indeed, worked bone is prominent in prehistoric and Roman deposits, but thereafter evidence for it is harder to find, at least until the medieval period. A few sites have revealed significant deposits of post-Roman worked bone and Middle Saxon Hamwic is one of the earliest and most important of these, both for the quantity of material and for the direct association of bone and antler waste (Driver 1984; Riddler 2001). At this site the antler worker was also the bone worker. Small quantities of bone working waste provide few problems in the analysis of faunal remains and this was the case with the pioneering study of the Melbourne Street assemblage at Hamwic (Bourdillon and Coy 1980). Subsequently, however, the material was discovered in greater quantities, and this led to a consideration of the selection of raw materials for working. John Driver’s work remains an important introduction to this question (Driver 1984). Jennifer Bourdillon has carried his analysis further in examining deposits of worked and unworked bone from a variety of contexts within Hamwic. For postcranial red and roe deer, and for sheep, there appears to be no appreciable bias, with the quantity of worked bones remaining insignificant when set against the faunal assemblage. For horse, the evidence is more equivocal, with a noticeable use of lower limb bones in working, but with no obvious bias stemming from this preference. The situation is quite different with cattle, and with cattle metapodia in particular, which form over 85% of the offcuts from most worked bone assemblages at Hamwic, a figure comparable with that seen at medieval Schleswig (Ulbricht 1984, table 4). With some rubbish pits, the quantity of worked cattle metapodia may easily exceed the unworked total. Distinctions can also be noted between the metacarpus and the metatarsus, the latter bone providing the clearest evidence for deliberate selection, leading to a bias in the faunal remains record. At any site with evidence for bone working, therefore, it is important to be careful about size factors in relation to cattle metapodia. At the same time, Bourdillon notes that at Hamwic during the eighth and ninth centuries there was no sense in which animals were deliberately reared for the craftsmen. Rather, the craftsmen made good use of the material available at any particular time, and often (but not invariably) selected the most appropriate bone, of the best size for working. The Middle Saxon site of Hamwic is well known but its late Saxon successor is less familiar. An assemblage of red deer antler, gathered from a small excavation at Holy Rood in late Saxon Southampton, forms the basis for the final contribution to the volume. Records of the site and its features are sparse and the surviving material derives entirely from hand recovery, with no sampling strategy undertaken during excavation. Nonetheless, by using techniques favoured also by Annick Thuet in this volume, the waste is described, quantified and placed within its late Saxon context. It is compared with material from Hamwic and other sites in Southampton, and the evidence supporting the idea of this period as the ‘Golden Age of the Antler Worker’ (MacGregor 1985, 48) is also considered. To a great extent, Haithabu dominates the image of antler working from the ninth to the eleventh century, ably supported by the evidence gathered also from Wolin, further to the east (Ulbricht 1978; Cnotliwy 1958; 1970; 1973). The quantity of ix

waste gathered from these settlements far exceeds that from most contemporary sites, but this may be, in part at least, a reflection of the location of excavations at other sites (following Riddler 2001), alongside the mechanisms of rubbish disposal. Sites of the seventh to ninth centuries include numerous rubbish pits, and sometimes consist of little else. Settlements of a later date, in some circumstances, make use of extensive middens and redeposit waste material in levelling layers that raise ground surfaces. Accordingly, waste material can be moved from the proximity of a workshop, and reinterred elsewhere. Small antler assemblages like that from Southampton (for which virtually all of the waste came from pits) provide an indication of production episodes, even if they are dwarfed by their Continental contemporaries.

Bibliography AUFLEGER, M., 1996, Beinarbeiten und Beinverarbeitung, Die Franken. Wegbereiter Europas. Vor 1500 Jahren: König Chlodwig und seine Erben, Mainz, 640-9 BÉAL, J. C., 1983, Catalogue des objets de tabletterie du musee de la civilisation Gallo-Romaine de Lyon, Lyon BÉAL, J. C., 1984, Les objets de tabletterie antique du musée archéologique de Nîmes, Cahiers des musée et monuments de Nîmes 2, Nîmes BECKER, C., 1993, Zur Aufdeckung von Kausalitäten zwischen Ernährungsgepflogenheiten und Knochenverarbeitung, in H. Friesinger, F. Daim, E. Kanelutti and O. Cichocki, Bio Archäologie +. Frühgeschichtsforschung, Archaeologia Austriaca Monograph 2, Vienna, 13357 BENDREY, R., forthcoming The Animal Bone, in, Rady J, editor, Roads to the Past. The Monkton Landscape. Prehistoric, Roman, Anglo-Saxon and Medieval sites along the duelling of the A253, Canterbury Archaeological Trust Occasional Papers, Canterbury BÍRÓ, M. T., 1994, The Bone Objects of the Roman Collection, Catalogi Musei Nationalis Hungarici, Series Archaeologica 2, Budapest BÍRÓ, M. T., forthcoming Combs and Comb Making in Roman Pannonia: Ethnical and Historical Aspects, in I. D. Riddler, A. MacGregor and N. Trzaska-Nartowski, editors, Combs and Comb Making, CBA Research Report, London BOURDILLON, J. and COY, J., 1980, The Animal Bones, in P. Holdsworth, editor, Excavations at Melbourne Street, Southampton, 1971-1976, CBA Research Report 33, London, 79-118 von CARNAP-BORNHEIM, C., 1994, Die beinernen Gegestände aus Kastell und Vicus in Niederbieber, Bonner Jahrbücher 194, 341-95 CHOYKE, A. M. and BARTOSIEWICZ, L., 2001, Crafting Bone: Skeletal Technologies through Time and Space, British Archaeological Reports, International Series 937, Oxford CNOTLIWY, E., 1958, Wczesnośredniowieczne przedmióty z rogu I kości z Wolina, ze stanowiska 4, Materiały Zachodnio-Pomorskie 4, 155-240 CNOTLIWY, E., 1970, Pracownie grzebiennicze na Srebrnym Wzgorzu w Wolinie, Materiały Zachodnio-Pomorskie 16, 209-87 CNOTLIWY, E., 1973, Rzemioslo rogownicze na Pomorzu wczesnosredniowiecznym, Gdansk DESCHLER-ERB, S., 1998, Römische Beinartefakte aus Augusta Raurica. Rohmaterial, Technologie, Typologie und Chronologie, Forschungen in Augst 27, Augst DRIVER, J. C., 1984, Zooarchaeological Analysis of Raw-Material Selection by a Saxon Artisan, Journal of Field Archaeology 2, 397-403 GOSTENČNIK, K., 1996, Die Kleinfunde aus Bein vom Magdalensberg, Carinthia 1/1996, 105-37 GOSTENČNIK, K., 2001, Pre- and Early Roman Bone and Antler Manufacturing in Kärnten, Austria, in: A. Choyke and L. Bartosiewicz, Crafting Bone: Skeletal Technologies through Time and Space. Proceedings of the 2nd Meeting of the (ICAZ) Worked Bone Research Group, Budapest, 31 August – 5 September 1999, British Archaeological Reports, International Series 937, Oxford, 383-97 x

GOSTENČNIK, K., forthcoming Die Beinfunde vom Magdalensberg, Archäologische Forschungen zu den Grabungen auf dem Magdalensberg 15, Klagenfurt MACGREGOR, A., 1985, Bone, Antler, Ivory and Horn. The Technology of Skeletal Materials since the Roman Period, London MIKLER, H., 1997, Die römischen Funde aus Bein im Landesmuseum Mainz, Monographies Instrumentum 1, Montagnac OBMANN, J., 1997, Die römischen Funde aus Bein von Nida-Heddernheim, Schriften des Frankfurter Museums für Vor- und Frühgeschichte 13, Bonn PETITJEAN, M., 1995, Les peignes en os à l’époque mérovingienne. Evolution depuis l’Antiquité tardive, Antiquités Nationales 27, 145-91 PETKOVIC, S., 1995, The Roman Items of Bone and Antler from the Territory of Upper Moesia, Monographs of the Archaeological Institute of Belgrade 28, Belgrade RIDDLER, I. D., 1996, The Antler Waste, in R. J. Williams, P. J. Hart and A. T. L. Williams, Wavendon Gate. A Late Iron Age and Roman Settlement in Milton Keynes, Buckinghamshire Archaeological Society Monograph Series 10, Aylesbury, 130-6 RIDDLER, I. D., 2001, The Spatial Organisation of Bone-working at Hamwic, in D. Hill and B. Cowie, editors, Wics: the Early Medieval Trading Centres of Northern Europe, Sheffield Archaeological Monographs 14, Sheffield, 61-6 SCHMID, E., 1965, Damhirsche im römischen Augst, Ur-Schweiz 29, 53-63 TEICHERT, M., 1983, Geweihreste aus der germanischen Siedlung bei Muhlberg, Kr. Gotha, Zeitschrift für Archaologie 17, 115-22 ULBRICHT, I., 1978, Die Geweihverarbeitung in Haithabu, Die Ausgrabungen in Haithabu Band 7, Neumünster ULBRICHT, I., 1980, Knochen- und Geweihfunde aus Alt-Schleswig. Aussagen zu Wirtschaft, Handel und Gesellschaft, Lübecker Schriften zur Archäologie und Kulturgeschichte 4, 211-14 ULBRICHT, I., 1984, Die Verarbeitung von Knochen, Geweih und Horn im mittelalterlichen Schleswig, Ausgrabungen in Schleswig. Berichte und Studien 3, Neumünster

xi

xii

K. Gostenčnik: Elk Antler as a Material of Manufacture

ELK ANTLER AS A MATERIAL OF MANUFACTURE Finds from Late Republican/Early Imperial ‘old Virunum’ on the Magdalensberg in Carinthia, southern Austria1 Kordula GOSTENČNIK

INTRODUCTION: SITE AND RAW MATERIALS

on the excavated site itself, nor in its vicinity; the localization will have to be proved by further field work. The excavations have focused on the town’s centre and the surrounding quarters, at the very beginning also on the cemetery and additionally on a stronghold on top of the mountain, the latter dating to between approx 15 BC and AD 15/20; a temple seems to have existed on top of the Magdalensberg, and from that area and a well in the settlement alone - the latter near the road to the mountain’s top - a few pieces of pottery also date into AD 2nd century (the well produced only one!). The fringes of the town are hitherto known only vaguely, indicating however an enormous expansion, which too will have to be investigated in years to come (Figure 1).

Bone, antler, and to some extent also ivory, horn and teeth are most common materials of manufacture among Roman small finds, but it is rather surprising to find elk antlers2 used for the production of artefacts with a typological provenance from the Mediterranean, especially in an area where objects of animals’ skeletal materials were used in a diminishing number before the Roman invasion. The finds dealt with in this paper were excavated on the Magdalensberg, where after the mid 1st century BC, about three decades before the Roman occupation of Noricum (in 15 BC), merchants from the south, especially from Aquileia, built up a trading post with Italican town planning structures (Figure 1)3. As supported by literal and epigraphical evidence, the town was named Virunum4, like its successor afterwards, when, by the time of Claudius in the AD mid 1st century, ‘old Virunum’ on the mountain was abandoned in favour of the new capital of Noricum, erected in the plains of the Zollfeld to the southwest of the Magdalensberg. A preceding, pre-Roman native Iron Age settlement has not been traced out yet on the mountain, neither 1

2

From the very beginning of the Magdalensberg settlement onwards, bone and antler were introduced as materials of manufacture by means of a vast variety of products unknown in Noricum before. Their appearance, along with an increased presence of Italicans after the mid 1st century BC, gives an impression of the daily use of bone objects, with which the southern immigrants were accustomed. The lack of any larger quantities of worked bone in our pre-Roman Iron Age settlements therefore remains an interesting, though not yet clearly resolved aspect5.

This paper is the revised and extended version of a recently published article, cf. Gostenčnik 1998. I wish to thank Ian Riddler for including the text in the present volume, and especially for perusing and correcting the manuscript. My thanks also go to Hubertus Mikler, Mainz, for securing literature otherwise unavailable for me, to Sabine Deschler-Erb, Hofstetten, for a short summary of her then forthcoming monograph, and, last but not least, to my colleagues from the Magdalensberg-team, who commented benevolently on an earlier draft. The drawings of the Magdalensberg-finds are all my own, the photographs were taken by U.P. Schwarz, Landesmuseum für Kärnten/Klagenfurt, Pl. 9 is my own.

Evidence of bone working is provided for the very beginning of ‘old Virunum’ already, both by raw materials and workshop waste, indicating both turning and carving of bone and antler. The need for raw materials was served by butchery and hunting, the town’s function as a trading base on the other hand facilitated marketing the products, simultaneously providing a demand for them - their production would not have started otherwise6.

Cf. gen. on elks RGA² vol. 7, s.v. Elch; Der Neue Pauly vol. 3, s.v. Elch; MacGregor 1985, 32ff.; Deschler-Erb 1998, 16ff.; a few references are also given in Catalogue München 1997; on current elk populations cf. Heptner et al. 1974, and Rülcker 1986; for archaeozoological data especially Chaix et al. 1981.

3

Piccottini 1977; id. 1989; id. et al. 1990; a prosopographical study was published by Vetters 1954. The top of the mountain measures 1058 m above sea-level, or 580 m above the surrounding plains; the air-line distance between Aquileia and the Magdalensberg is approx 130 km. The term Late Republic in preference to Late Latène outlines more clearly the exceptional rôle of the settlement in its beginning.

4

Cf. most recently Dobesch 1997. G. Dobesch interpreted literary sources from Antiquity referring to the etymology of the name Virunum on the mountain as a Continental Celtic lore: a wild boar devastated the country, but he was finally caught by a brave man; the hero was welcomed by acclamations of the Noricans as vir unus! Nollé 2001, 50, however, has drawn attention to the fact that the text reflects a close knowledge of Graeco-Roman cultural traditions and the topoi of Greek myths referring to the foundations of towns with a wild-boar-and-hero motive; the acclamation too is Latin, not Celtic.

On examining the raw materials, four discs and a tablet of unusual diameter attracted special attention. They turned out

1

5

This is not exclusive among our Carinthian sites, although it must be admitted that archaeological investigations on a larger scale in our Iron Age settlements and cemeteries have started anew only recently, and therefore this lack of evidence could be due to the lack of representative amounts of material. However, the Braunsberg oppidum in Lower Austria, near the Danube river and in the vicinity of the border to Slovacia, where excavations have been going on for a longer period of time, has provided us with one Stabwürfel (parallelepiped) and some astragali only, cf. Urban 1994, passim. The iron age sites in Slovenia too did not provide evidence for the use of bone or antler in such big quantities as any Roman settlement does. On the other hand, the so called SH-section on the Magdalensberg has produced approx 19,200 pieces of pottery, but only about 100 bone and antler artefacts; the relative balance of numbers is thus clearly visible.

6

Cf. Gostenčnik forthcoming, passim.

Materials of Manufacture

Figure 1. Magdalensberg. General outline of the excavations (after Piccottini)

to be made of elk antler, a most unusual situation among Roman finds. Elk bones and one piece of antler (Catalogue N°. 1) have been known, since the first comprehensive studies on the Magdalensberg’s animal bones were published7, but the artefacts of skeletal materials had not been examined then, and therefore the finds presented here escaped further notice8.

exactly the last Carinthian elks became extinct, is not yet revealed10. For Germany in the Middle Ages, the existence of elks to the west of the Elbe is currently doubted. In Europe today, elks are only found in Scandinavia and the north-east (Figure 2). of a newly excavated temple, cf. Mikler 2001; for southern Germany, von den Driesch 1979, Karlstein (undated); Kokabi 1982, 105 and id. 1988, 204-5, from Rottweil; id. et al. 1994, 285ff., from a Roman villa in Bondorf; Pichler 1976 reports finds from the Late Roman fortress Vemania/Isny. For Switzerland cf. Deschler-Erb 1998, 20f. n. 68-72 and Schibler et al. 1988, passim, 3 antlers and 11 bones from the extremities confirmed for Augst, 1st to 3rd centuries; a probable pendant in elk antler from Augst is given in Deschler-Erb 1998, pl. 37,3975. Würgler 1958, 258f., has 4th century evidence for elk bones from Schaan, Liechtenstein. However, Schmid 1963, 17ff. and ead. 1972, 30f. published 25 elk bones from Hüttenböschen/Switzerland, excavated in the area of a Roman temple and probably from sacrificed animals; the early Celtic Heuneburg in southern Germany, of mid 1st millennium BC, too provided for most astonishing data, as 128 elk bones were ascertained among the animal bones, cf. von den Driesch et al. 1989, 150. From Celtic Manching in southern Germany, on the other hand, only two elk bones are known, cf. Boessneck 1971, 98 (talus) and Hahn 1992, 235ff. (tibia). For Britain and the north-western Continent cf. e.g. MacGregor 1985, 35 and 37, with references, Roman period and Middle Ages.

The presence of elks in our region in ancient times is nothing unusual, as the country had swamps and marshes to a vast extent around the many Carinthian lakes, in the great plains of the Klagenfurter Becken, and in the river valleys; those wetlands were only drained in the Middle Ages. In similar environments, excavations in Switzerland and along the Danube river in southern Germany too provided us with data for those animals in prehistoric and Roman contexts9. When 7

Dräger 1964; Ehret 1964; Luhmann 1965; Fruth 1966; Hildebrandt 1966; Hornberger 1970. Investigations on animal bones from the Magdalensberg were started again only recently; any new data therefore are not available yet. Cf. also Benecke 1994a, 48ff., ibid. 278ff.

8

Fruth 1966, 23f. and pl. 1, fig. 3a-b, also including a review of literature published by then; he examined five bones of at least two elks - an adult and a young animal: three metacarpal fragments, one talus, one os centrotarsale. Supplementary to the former is Hornberger 1970, 71 and ibid. schedule 1, who added our Catalogue N°. 1 to the finds. Those investigations are limited to the finds from 1948-1966. Another roughly cut and burned elk antler-point remained unpublished. I do not agree with Gaggl 1997, who considers the Magdalensberg finds as indication of stray deer, perhaps from the Danube region; Fruth 1966, 24 looks upon the local natural habitats as more likely, besides the lowlands to the north or east of the Alps.

9

10

Elks are usually reported in very small numbers only; out of a number of publications I have chosen the following listing: the most recent finds of elk antlers were made in Mainz within a deposit dedicated to the deities

2

I was not able to find out whether or not elks were still known in the Middle Ages in Carinthia; a quest for data in the 1997 exhibition on hunting since Antiquity in Carinthia, cf. Catalogue Ferlach 1997, remained without results. In a different environment, elk antler was found among the workshop waste in an Early Middle Age (9th/10th century) Slavic bone carver’s workshop in Gars/Thunau am Kamp in Lower Austria, to the north of the Danube river; those finds are on display at Vienna University/Institute of Prehistory. Von den Driesch et al. 1988, 203 refer to the presence of elks along the Danube river in southern Germany in the Early Middle Ages. For northern Germany cf. gen. Ulbricht 1978, 16ff. and ead. 1984, 132-3, from Haithabu and Schleswig; RGA² vol. 7, s.v. Elch 128 doubts elks to the west of the Elbe river in the Middle Ages; for this discussion cf. also the references in MacGregor 1985, 35 and 37.

K. Gostenčnik: Elk Antler as a Material of Manufacture

therefore look after those resting-places and remove all the trees’ roots or cut their trunks so that the trees would fall over when an elk leans on them, the elk thus being an easy bag. Nowadays, this passage in Caesar’s text is supposed to be a later interpolation, partly derived from Pliny13. Pliny the elder (AD 23-79, died during the eruption of Vesuvius) in his Natural History (8,16) refers to the same tale, comparing alces with a similar northern animal called achlis, which moves backwards when grazing, due to its pronounced upper lip, which would inconveniently turn inside out otherwise. T. Calpurnius Siculus, orator under Nero (emperor 54-68) in Rome, mentions elks in his Bucolica (VII 56-60), and calls them rare animals even in the woodlands, where they come from; they were caught for venationes, and thus Siculus himself saw them in Rome. The Historia Augusta (created in Late Antiquity) proves the import of those animals to 3rd century Rome under the Gordians (SHA Gord. 3,68; 33,1). Ten elks were depicted ‘...in domo rostrata Cn. Pompei...’ in a painting referring to a celebration during the aedilitas of the later emperor Gordianus I (emperor in 238) under Severus (the latter emperor 193-211). For Gordian III (emperor 238-244), the SHA makes mention of another 10 elks (SHA Gord. 33,1), possibly meant for his future Persian triumph, which never took place as he was killed in his Persian war in Mesopotamia. Especially for Rome’s millenium under Philippus Arabs (emperor 244-249) in AD 248, again ten elks were brought to Rome, and those were even struck on the reverse of a special coin-issue celebrating the memorable event; those elks are sometimes considered as Gordian III’s animals, who were still on hand14. Aurelianus (emperor 270275) carried along some elks in his triumph (SHA Aurel. 33,4). It is inconceivable that the animals should have survived for long in the climate of Rome15.

Figure 2. Current distribution of elks in Northern Europe (after MacGregor)

ANCIENT WRITTEN SOURCES ON ELKS Literary sources on elks in Antiquity are rare11, which is far from being surprising, as the Mediterranean was not their natural environment. To Greeks and Romans, the elks (Latin alce or alces, Greek Ù ðkjg, -gr) were mythical animals, living in the north; a few passages are cited here, in order to give an impression of what could have been known to our ancestors in Antiquity. In Strabo’s Geography (IV 6,10) (approx 60 BC – AD 24), a reference is given to a lost passage by Polybius (approx 200-120 BC), mentioning an elk-like animal in the Alps. Although the localization of Polybius’ Alps is not undisputed at times, the presence of elk-bones in Roman settlements in the Alpine provinces and along the Danube river in southern Germany is an interesting factor, as the elks mentioned in the text are sometimes summoned as evidence for a more distant area in the north12.

For Early Middle Age Gaul, Venantius Fortunatus (later half of 6th century) seems to witness elks, if helix is to be understood as them. And, finally, the testament of a Gaul approx AD 100 may also be listed here, as it mentions ... et stellas omnes ex cornibus alcinis ..., which he wants to be burned together with his private possessions and especially hunting accessories on his funeral pile (CIL XIII 5708 II 28).

Perhaps the most renowned story is derived from Julius Caesar’s Gallic Wars (VI 27) (war in Gaule 58-51 BC; Caesar was murdered in 44 BC), which sounds like hunter’s yarn; in it Caesar narrates of a beast named alces as similar to a goat but taller, with motley coat and blunt antlers. For lack of knee-joints, the animal leans on a trunk to rest, because it is unable to stand up again after falling to the ground; hunters 11

12

Thessaurus Linguae Graecae, s.v. ðkjg, -gr, Thessaurus Linguae Latinae, ss. vv. alce, -es; achlis, -is; alcinus. Lenz 1856, 215f.; Keller 1909, 281ff.; Toynbee 1983, 132f.; perhaps still the best survey on literary sources is Prell 1941, 3ff. and 62ff. for Antiquity, and 66ff. an enormous collection of passages in Middle Age and later literature with references to elks; cf. also Der Neue Pauly vol. 3, s.v. Elch. Prell 1941, 3ff. had no knowledge then of archaeo-zoological investigations, which provided us with enough data during the last decades. He therefore tends to be over-critical as regards elks in middle Europe during the Roman period. For questions of localization of the geographical references in the Greek and Latin texts cf. gen. the commented text-editions, but mind that archaeozoological evidence usually goes unheeded there.

3

13

Der Neue Pauly vol. 3, s.v. Elch.

14

RIC Philip I n°. 19, pl. 6,10, mint of Rome; this piece definitely is an elk with heavy and palmated antlers and a tuft on the throat, whereas the antlers in pl. 6,11 are finely carved and stag-like; cf. also the comment ibid. 62. Prell 1941, pl. 3-4 has evidence for more coins of that issue, because according to him, goat-like animals on the reverses might be young elk bulls, whose antler started to grow in their early first season as long single points; a pronounced upper lip and a tuft on the throat are clearly discernible, but he might have over-interpreted them.

15

Further sources: Pausanias (V 12,1; IX 21,3) in AD 2nd century calls it an animal in the country of the Celts (as a Greek source, this might also include the Germania libera) which can be hunted only by encircling it at great length; C. Julius Solinus (20,6-7) in 3rd century compares the elk with a mule, and it comes from Gangavia insula; according to Isidorus from Sevilla (orig. XIV 4,4) (AD 570-636), elks are native in Germania; after Cicero’s freedman and secretary Tiro (1st century BC), inventor of shorthand, the Notae Tironianae 106,41 has the singular nominative form alces with an –s ending, which seems to be uncommon among Latin authors, who preferably write alce. Cf. also Der Neue Pauly vol. 3, s.v. Elch. Possible Greek sources are considered in Keller 1909, 282, or Prell 1941, 3ff.; I did not check all the Greek authorities of the 2nd half of 1st millennium BC quoted by them.

Materials of Manufacture

Figure 3. An elk in comparison to a red deer stag (after Catalogue München)

Whether or not this passage really refers to antler roundel pendants, as has been stated some decades ago by J. J. Hatt16 is not yet answered clearly; among published pieces at least, elk antler roundels do not seem to be reported17. However, on no account would I regard red deer and elk antler roundels to be easily distinguishable for archaeologists without a closer knowledge of archaeozoology.

on back, throat and belly may have been derived from fallow deer, which are more widespread, but the height of the latter inbetween roe deer and stags would not go together with the depiction of perilous game. Fallow deer are rather small and more suitable for hunting with hounds20, whereas male elks are reported to reach shoulder-heights up to 2,0 m and lengths of nearly 3,0 m, with antlers weighing 30,0 kg (Figures 3 and 4)21 Evidence for the domestication of those animals in Antiquity derives from regions to the north of the Black Sea, indicated by a milking scene dated to around 200 BC22.

A very rare depiction of elks is given on the panel of a diptych of Late Antiquity, approx AD 400 to early 5th century, in Liverpool Museum, showing a venatio with five elks (Figure 4)18. A pronounced upper lip, a hump above the shoulders, a tuft on the throat, and palmated antlers are clearly discernible, though the latter are not, as would usually be expected, growing from the side of the skull19. But when considering those animals to be most unusual in southern environments, a lack of familiarity in their appearance to the artisan may well account for discrepancies in their depiction. The dots 16

Hatt 1955.

17

Cf. e.g. Greep 1994.

18

Cf. Delbrueck 1929, 223ff. and pl. 58, suggesting stag, also Keller 1909, 283. Cf. Toynbee 1983, 132f., where fallow deer from Mesopotamia is named besides the usual interpretation as elks. For a more recent publication of the diptych and an updated bibliography cf. Gibson 1994, 17f. and fig. VIIa, identified as stag-fight. The dimensions of the panel are 29,4/12,0/0,6 cm.

19

THE MAGDALENSBERG ELK ANTLER ARTEFACTS Hunting was the source for supplying workshops with elk antlers on the Magdalensberg. Our elk bones, being fragments of tali and metapodials, might also indicate the workshop waste

But cf. e.g. Heptner et al. 1974, passim, and Rülcker 1984, passim, stating much differing growth of antlers.

4

20

Sources for hunting scenes with fallow deer are given in Toynbee 1983, 130ff. and footnotes.

21

Cf. RGA² vol. 7, s.v. Elch, 127; Heptner et al. 1974, passim; Rülcker 1986, passim; Mikler 2001.

22

Benecke 1994b, 441f. and fig. 256; cf. ibid. also references for experimental domestication of elks in the former Soviet Union. For elks in Scythian art cf. e.g. Prell 1941, pl. 2-3, ibid. pl. 1 also in cave art from the Stone Age.

K. Gostenčnik: Elk Antler as a Material of Manufacture

have not yet yielded structures or an increased amount of lower limb bones to indicate the localization of a tannery. But bearing in mind the enormous size an elk bull might attain (cf. above), the main reason for the rarity of elk bones in our settlement probably derives from the fact that the hunted elks were dismembered, and meat, hides, and antler alone were carried away from the plains up to the mountain, leaving the rest of the carcasses behind mainly for logistical reasons. It would have been a strenuous task to toil up the mountain with such an enormous prey, keeping in mind also the distances; besides, Pausanias’ description of an elk hunt (cf. n. 15) also points at the difficulties in this task. I even consider this most positively as the real reason for the interesting fact that, in most cases, limb bones and/or antler are the only hints as to the presence of elk in the archaeozoological record of most Roman sites, where the bones of those animals are never recorded in vast numbers25, although they appear with regular frequency on different sites in the Alps and the adjacent plains in southern Germany until the Early Middle Ages. Taking this into account, the archaeozoological records in my opinion indicate the presence of elks in the various sites’ nearer or farther environs rather than the import of raw materials or the hint as to possible stray deer. Mentioning above the typological provenance of our bone and antler small finds from the Mediterranean, questions also arise as to the origin of the craftsmen who manufactured these goods. As production on the Magdalensberg started just after the mid 1st century BC, the first artisans are to be regarded as coming from the south rather than being native inhabitants of Noricum, although any skilled artisan on the other hand should have been able to adopt the techniques without difficulties. Any use of elk antlers or antler in general may be concerned with experiments with raw materials at first, in order to find out their workability26. The compacta of elk antlers is comparatively thick, so decorative elements could be cut deeply into the surface. Being only slightly curved, the material is ideal for carving or for turning large and flat objects27. Our finished products (Catalogue N°s 2-5) are lathe-turned, whereas Catalogue N°. 1 is prepared for further use. Aside from the turned pieces, carvings in elk antler are not reported yet from our site. 25

Cf. n. 9-10, especially also for Augst.

26

Deschler-Erb 1998, 90 considers Celtic influence to be a reason for increased use of antlers in AD 1st century Augst; for lack hitherto of any considerable numbers of worked bones and antler from a pre-Roman settlement, this can neither be supported nor denied with our Carinthian material, cf. Gostenčnik 2001a. Although data as to worked bones from Roman Italy are rare, one should consider the vast number of red deer antlers reported from the Vesuvian area in Catalogue Naples 1999, 49, even if they have not emerged yet as artefacts. One reason could be the fact that bovine metapodials from Italy surpass Alpine stock in their length and thickness of the compacta, and therefore antlers could have been neglected, cf. the results as to native and imported cattle at Augst in Deschler-Erb 1998, 80f., especially the hinges, and Catalogue Naples 1999, 118, fig. 68. Exploitation of red deer antlers in Italy is proved e.g. for Etruscan workshops and again for Rome herself in the Early Middle Ages, cf. Nielsen 1995, 7th/6th century BC at Poggio Civitate in Etruria, and Catalogue Rome 2001, 64 middle, a vast quantity of antler waste from 7th century AD Rome.

27

Larger decorative fittings for a Roman couch now housed in Cambridge’s Fitzwilliam Museum were i.a. manufactured of horse’s scapulae, cf. Nicholls 1979, 1.

Figure 4. Diptych-panel of Late Antiquity from Liverpool Museum, venatio with five elks (after Delbrueck)

of a tannery besides their provenance from a kitchen23, as skulls or extremities were often left with the raw hides when delivered to the tanner; so these bones could suggest the presence of such a specialized workshop, which would otherwise go unheeded for lack of evidence24. Our excavations 23

Kokabi 1988, 204f. and fig. 30 gives evidence of butchered elk bones, meat being removed with a knife.

24

Cf. MacGregor 1985, 30; Berke 1989, passim clearly outlines, which regions of animals’ skeletal remains are typical for workshops associated with their exploitation.

5

Materials of Manufacture

Pl. 1 Catalogue N°s 1-2; scale 1:0,7

Pl. 2 Catalogue N°s 3-4; scale 1:0,7

Catalogue N°. 1 is a tablet with smoothed surfaces, showing clearly the use of a knife or drawing knife28 for this purpose29. As far as traces of production are discernible, many a bone 28

Salaman 1989, 175ff.

29

Cf. Barbier 1987, where the smoothing of a bone artefact with a knife is depicted. A Lapp carving and decorating a reindeer-artefact by means of a knife only, is shown in Ulbricht 1978, pl. 54; this artisan did not use any further tools during production, which indicates that data gained in experiments should be regarded with caution - an artisan may against all experimentally proved techniques or theoretical doubts behave in an ‘unorthodox’ fashion.

or antler artefact or workshop waste from the Magdalensberg bears distinct marks of having been treated with a knife. Only Catalogue N°. 5 has deep imprints of a chisel on its reverse, the single obvious indication for the use of this tool in the working of bone and antler on our site30. The edges of Catalogue N°. 1 are sawn; the reverse surface still shows its

6

30

Cf. Barbier 1987, who by way of experiment carved a belt buckle with a chisel.

31

In comparison with red deer, the spongy tissue of our elk antlers seems to be finer; in MacGregor 1985, 14, with references, it is stated to be the

K. Gostenčnik: Elk Antler as a Material of Manufacture

products. However, the room where those pieces were unearthed did not form part of a workshop; it belongs to a complex of rooms around the older baths on the western side of the forum. The material had been taken from some unknown place. The four discs (Catalogue N°s 2-5) are lathe-turned, though not always on both sides. In order to turn the design into the surface, the discs had to be handled from the front, and fixed to the lathe by means of special auxiliary facilities33. Thus, on the reverse of Catalogue N°. 4, the central imprint of a point is visible still, permitting both the fixing of the object and a smooth running on the lathe. As the edges are turned as well, the workpieces do not seem to have been flush-fitted within a wooden frame; the latter would not have allowed working the edges. Catalogue N°. 2 is turned on both sides, showing fine marks of the turning chisel even around the bore-hole. All four finds were carefully centred, but the reverses of Catalogue N°s 3-5 are more or less neglected, which indicates that they were not meant to be visible. Catalogue N°. 2 might have been part of a couch-leg or a leg of some other piece of furniture34, which consists of several elements mounted on an iron rod. Around its central hole, one side has an edge for fitting together with the next piece, and the opposite side shows imprints around the hole, which seem to have been caused by friction of the following element. I do not consider Catalogue N°. 2 to be a flywheel for the cord of a bow on a bow-lathe, as has been suggested during the discussion of the finds, because the cord would have left deep grooves in the spongy tissue, and there are obviously none. On the other hand, a couch leg normally consists of a wooden core, supported by an iron rod, with bone carvings applied to the surface of the lathe-turned wood. But discoid elements of considerable diameter and number formed part of furniture-legs depicted in wall-paintings, giving an impression of how to reconstruct the mounting of this find. Due to the raw material, this decorative element must have been produced on the site itself, indicating perhaps repairs in a local workshop. Fragments of couches or furniture in general are not much in evidence among our small finds, but considering that they were objects of value, damage may well have been mended; all household effects and furniture were moved to Virunum on the Zollfeld, after the town was abandoned by AD mid 1st century, which explains this lack in the archaeological record. Pl. 3 Catalogue N°. 5; scale 1:0,7

33

Cf. for this purpose Spannagel 1948; Mutz 1972; Barbier 1987; Earwood 1993. The shapes of chisels used for turning wood, bone or metals in Antiquity are, with the exception of a few suggestions, still unrecognized among published metal-objects; a typology therefore is still lacking. Spannagel 1948, or Salaman 1989, s.v. Chisel, and others besides, are, although based on technology of recent date, most helpful for anyone, who is simply trying to make himself familiar with the tools that might have been used in Antiquity also. Save for Mutz 1972, authors keen on experimenting tend to forget to enclose also detailed drawings or photographs of the tools they used for turning, although the reconstructed lathes are frequently published. It would be most welcome to have an impression of how the cutting edges of chisels and the like are set.

34

For the construction of a couch-leg cf. gen. Nicholls 1979; Letta 1984. Couches from various places have been published to a vast extent since these two papers were written; for references cf. Mikler 1997, 64ff., Deschler-Erb 1998, 181ff.

fine spongy tissue, which was not removed31. The stratum from which the tablet derived, filled in beneath two floors, dates into the last decade BC32, and, what is more, the piece was associated with a variety of workshop waste, mostly red deer antlers with traces of sawing and removing tissue with a knife, besides some broken artefacts, but also finished other way round. As my comparative material is limited to Magdalensberg, however, I can not decide here. 32

For the whole complex cf. Gostenčnik forthcoming, part 3, on evidence for workshops.

7

Materials of Manufacture

A

B

Pl. 4 Catalogue N°. 1

A

B

Pl. 5 Catalogue N°. 2

A

B

Pl. 6 Catalogue N°. 3 8

K. Gostenčnik: Elk Antler as a Material of Manufacture

A

B

Pl. 7 Catalogue N°. 4

A

B

Pl. 8 Catalogue N°. 5

A

B

Pl. 9 Catalogue N°. 6 6 9

C

Materials of Manufacture

Figure 5. Wickerwork vessels from the Kertch-region (after Rieth and Vaulina et al)

finds37, although with dark brown it is difficult, as this seems to be caused by the conditions of the immediate matrix within a stratum rather than deriving from an intentional dyeing process during the course of production. The matrix for N°. 3 derives from enclosed charred debris, and the piece is also fragmented, possibly due to fire.

Catalogue N°.s 3 and 4 are elements of pyxides, N°. 3 the bottom and N°. 4 the lid of considerable diameter, similar in shape to their well known smaller parallels35 Both pieces are decorated by means of grooves. Due to the unusual diameter it may be considered that the cylindrical elements of those pyxides were made of wood. However, most pyxides with an ivory cylinder, especially from Late Antiquity, seem to have wooden lids or bottoms of more recent date, i.e. from mending them36. It is interesting to note that N°. 3 is of a dark brown colour right through, which might result from dyeing, which was applied to some of our bone and antler 35

Cf. Béal et al. 1983, passim.

36

Cf. gen. Volbach 1976; Catalogue Schallaburg 1998, 199-201.

37

10

Cf. Gostenčnik forthcoming, chapter 3.1.3.4., on dyeing; for an additional study cf. also Gostenčnik in preparation. Most usual colours are red or a reddish and sometimes brownish shade, which always appears in patterns on the surface only and never penetrates the whole bone, furthermore green and dark brown, both penetrating the entire artefact when applied to it; black might have been used as a colour also, although it is hard to distinguish whether or not this was intended or caused by mere chance, i.e. through fire or circumstances of the context. Cf. also MacGregor 1985, 67ff.; Deschler-Erb 1998, 81 ff.

K. Gostenčnik: Elk Antler as a Material of Manufacture

Artefact Catalogue

Catalogue N°. 5 only revealed its intended purpose by an examination of wooden counterparts from the Kertch-region, from the northern coast of the Black Sea (Figure 5). Among grave goods in this area, wickerwork baskets with wooden discs as centres both for lid and bottom were found in different places, now housed in the Ermitage/St. Petersburg and the Louvre/Paris38. The edges are grooved twice, the wickerwork starts within the grooves with thin rods, obviously without an additional frame of thicker rods. A most interesting find from AD 1st half of 2nd century from Knossos indicates a different mounting of rods, the central disc being bored several times around the groove, in order to make the frame look like the spokes of a wheel, around which the weaving starts39. A marble urn in the Metropolitan Museum/NY, imitating a cylindrical wickerwork vessel in every detail, shows a separate central piece on the lid too40, a slightly different model for the reconstruction of small baskets with wooden or bone elements. Catalogue N°. 5 is not the only such find from the Magdalensberg; two more similar pieces, although smaller in diameter (3,5 cm and 2,8 cm )41, were found on our site as well, both with a deep groove around the edge. Wickerwork was nothing unusual among objects of daily use42, but organic materials did not survive over time on our site, except for a few waterlogged timbers, and barrels forming frames in wells. A bonecarver’s workshop and an osier’s may well have cooperated at our location.

Abbreviations: Pl Prov MB Dat L W H D We Ref

Plate Provenance Magdalensberg Dating Length Width Height Diameter Weight References

1. (Pl. 1,1 and 4): Prov.: MB 1957, Q/1; Dat. last decade BC; L. 8,8 cm, W. 6,3 cm; We. 32 g. Ref. Hornberger 1970, 71 no. 7. Tablet, prepared for further use; on the outer surface traces of smoothing with a knife and remains of the antler structure, the edges are sawn; on the reverse fine spongy tissue. 2. (Pl. 1,2 and 5): Prov. MB 1973, NG 37; Dat. no closer dated context (or Claudian?); D. 8,4 cm; We. 54 g. Element for a furniture leg (?), lathe-turned all over. 3. (Pl. 2,3 and 6): Prov. MB 1981, SH/7; Dat. Augustan; D. approx 11,0 cm; We. 12 g. Bottom element of a pyxis, probably dyed dark brown (?); front side and edges lathe-turned; reverse smoothed, with fine spongy tissue. 4. (Pl. 2,4 and 7): Prov. MB 1967, OR/28; Dat. between approx AD 35/40 and 50;. D. 9,2 cm, We. 32 g. Pyxis lid; front side and edges lathe-turned, reverse with central point, spongy tissue and remains of antler surface. 5. (Pl. 3 and 8): Prov: MB 1968, Room S, 2,0 m below Niveau I; Dat. probably early Augustan; D. 11,9 cm; We. 128 g. Central bottom element of a wickerwork vessel; front side and edge as well as partly the reverse are lathe-turned; groove around the edge; reverse with marks caused by removing the surface with a chisel.

Four finds out of our five are closely dated (cf. catalogue below), indicating use of elk antler products from the period of Augustus to the end of the settlement in AD mid 1st century. The presence of elk bones together with the artefacts in elk antlers on our site not only bear evidence of hunting for those animals both in 1st century BC and AD in Carinthia, but also of the use of their antlers in local workshops. Although the volume of production in our town can hardly be reconstructed, a preference for antler against bone is clearly indicated by a much larger quantity of antler-waste than that of bone43. Writing equipment, pyxides, casketmounts, gaming pieces, and many a knife-handle was produced in ‘old Virunum’; elk antler as a material of manufacture adds an interesting aspect to the working of animals’ skeletal materials in Late Republican/Early Imperial Carinthia, and gives evidence of the presence of elks in our local fauna for this period.

References 38

39

Cf. Vaulina et al. 1974; 151ff.; Pinelli et al. 1986; Rieth 1955; with slightly different reconstruction Sokol’skij 1971, 247ff.

The listing below comprises literature up to 1998; a few additions were made in the summer 2001, in course of editorial work

Sackett 1992, pl. 318,17, and a parallel from Tarquinia in SerraRidgeway 1996, pl. 173, T.116,8, both not identified as lids; cf. Gostenčnik 2001b, fig. 3,3-4.

40

Cf. Catalogue New York 1987, 131, fig. 100, AD 1st century.

41

Cf. Gostenčnik 1997; for further references cf. Gostenčnik 2001b: 57ff. I was not yet able to confirm a parallel from Northern Italy, which has recently been announced to me.

42

Cf. gen. Gaitzsch 1985; White 1975, 56ff.

43

Cf. Gostenčnik forthcoming, chapter 3.1.2., raw materials and workshop waste.

BARBIER 1987: M. Barbier, Le travail de l’os à l’époque galloromaine, Dossiers Histoire et Archéologie 126, 48-55 BÉAL et al. 1983: J.-C. Béal, M. Feugère, Les pyxides gallo-romains en os de Gaule méridionale, Documents d’Archéologie Méridionale 6, 115-26 BENECKE 1994a: N. Benecke, Archäozoologische Studien zur Entwicklung der Haustierhaltung in Mitteleuropa und Südskandinavien von den Anfängen bis zum ausgehenden Mittelalter, Schriften zur Ur- und Frühgeschichte 4, Berlin 11

Materials of Manufacture FRUTH 1966: M. Fruth, Tierknochenfunde aus der Stadt auf dem Magdalensberg bei Klagenfurt 4. Die Wiederkäuer ohne die Bovini Naturkundliche Forschungen zu den Grabungen auf dem Magdalensberg 6. Kärntner Museumsschriften 41, Klagenfurt

BENECKE 1994b: N. Benecke, Der Mensch und seine Haustiere. Die Geschichte einer jahrtausendealten Beziehung, Stuttgart BERKE 1989: H. Berke, Funde aus einer römischen Leimsiederei in Köln, Kölner Jahrbuch für Vor- und Frühgeschichte 22, 87992

GAGGL 1997: G. Gaggl, Haustierhaltung und Jagd in Noricum, in: Catalogue Ferlach 1997, 135-40

BOESSNECK 1971: J. Boessneck, Die Tierknochenfunde aus dem Oppidum von Manching, Die Ausgrabungen in Manching 6, Wiesbaden

GAITZSCH 1986: W. Gaitzsch, Antike Korb- und Seilerwaren, Schriften des Limesmuseums Aalen 38, Stuttgart GIBSON 1994: M. Gibson, The Liverpool Ivories. Late Antique and Medieval Ivory and Bone Carving in the Liverpool Museum and the Walker Art Gallery, London

Catalogue Ferlach 1997: Alles Jagd. Eine Kulturgeschichte, Katalog der Kärntner Landesausstellung Ferlach, Klagenfurt Catalogue München 1997: M. Kokabi, B. Schlenker, J. Wahl, L. Wamser (eds.), Schmuck und Gerät aus “Bein”. Vom Eiszeitalter bis zur Gegenwart, München

GOSTENČNIK 1997: K. Gostenčnik, Beinscheiben als Zentren von Deckeln und Böden geflochtener Behälter?, Bulletin Instrumentum 6, 13

Catalogue Naples 1999: A. Ciarallo, E. De Carolis (eds.), Homo Faber. Natura, scienza e tecnica nell’antica Pompei, Milano

GOSTENČNIK 1998: K. Gostenčnik, Die Kleinfunde aus Elchgeweih vom Magdalensberg, in: G. Piccottini (ed.), Die Ausgrabungen auf dem Magdalensberg 1980-1986, Magdalensberg-Grabungsbericht 16, Klagenfurt, 577-93

Catalogue New York 1987: J. A. Mertens, The Metropolitan Museum of Art. Greece and Rome, New York Catalogue Rome 2001: M. St. Arena (ed.), Roma dall’Antichità al Medioevo: archéologia e storia nel Museo Nazionale Romano Crypta Balbi, Milano

GOSTENČNIK 2001a: K. Gostenčnik, Pre- and Early Roman Bone and Antler Manufacturing in Kärnten, Austria, in: A. Choyke, L. Bartosiewicz (eds.), Crafting Bone: Skeletal Technologies through Time and Space. Proceedings of the 2nd Meeting of the (ICAZ) Worked Bone Research Group, Budapest, 31 August – 5 September 1999, BAR S. 937, Oxford, 383-97

Catalogue Schallaburg 1998: Ägypten. Schätze aus dem Wüstensand. Kunst und Kultur der Christen am Nil, Wiesbaden CHAIX et al. 1981 : L. Chaix, J. Desse, Contribution à la connaissance de l’élan (Alces alces L.) postglaciaire du Jura et du Plateau suisse. Corpus des mesures Quartär 31-32 13990

GOSTENČNIK, K. 2001b: Flechtwerk und Korbwaren im römischen Kärnten, Carinthia I 191, 53-74 GOSTENČNIK forthcoming: K. Gostenčnik, Die Beinfunde vom Magdalensberg, Archäologische Forschungen zu den Grabungen auf dem Magdalensberg 15, Klagenfurt, forthcoming

DELBRUECK 1929: R. Delbrueck, Consulardiptychen, Studien zur spätantiken Kunstgeschichte 2, Berlin Der Neue Pauly vol 3, s.v. Elch: H. Cancik, H. Schneider (eds.), Der Neue Pauly vol. 3, Stuttgart, 1997

GOSTENČNIK in preparation: K. Gostenčnik, Coloured bone artefacts from the Magdalensberg-excavations in Kärnten, Austria; paper presented at the 3rd International Bone Tool Conference in Augst (CH), September 2001. Internationale Archäologie, Espelkamp.

DESCHLER-ERB 1998: S. Deschler-Erb, Römische Beinartefakte aus Augusta Rauriapprox Rohmaterial, Technologie, Typologie und Chronologie, Forschungen in Augst 27/1-2, Augst DOBESCH 1997: G. Dobesch, Zu Virunum als Namen der Stadt auf dem Magdalensberg und zu einer Sage der kontinentalen Kelten, Carinthia I 187, 107-28

GREEP 1994: St. Greep, Antler roundel pendants from Britain and the north-western Roman provinces, Britannia 25, 79-97 HAHN 1992: E. Hahn, in: F. Maier, U. Geilenbrügge, E. Hahn, H.J. Köhler, S. Sievers, Ergebnisse der Ausgrabungen 1984-1987 in Manching, Die Ausgrabungen in Manching 15, Stuttgart, 23569

DRÄGER, N. 1964: Tierknochenfunde aus der Stadt auf dem Magdalensberg bei Klagenfurt in Kärnten 1. Die Vogelknochen, Naturkundliche Forschungen zu den Grabungen auf dem Magdalensberg 1. Kärntner Museumsschriften 33, Klagenfurt

HATT 1955: J. J. Hatt, ‘Et omnes stellas ex cornibus alcinis’. Talismans gallo-romains en bois de cerf ou d’élan trouvés dans les tombes, Revue archéologique de l’Est et du Centre-Est 6, 559

von den DRIESCH 1979: A. von den Driesch, Tierknochenfunde aus Karlstein, Ldkr. Berchtesgadener Land, Bayrische Vorgeschichtsblätter 44, 149-70

HEPTNER et al. 1974: W. G. Heptner, A. A. Nasimowitsch, Der Elch: Alces alces, Die neue Brehm-Bücherei 386, Wittenberg²

von den DRIESCH et al. 1988: A. von den Driesch, J. Boessneck, Haustierhaltung, Jagd und Fischfang bei den Bajuwaren, in: Die Bajuwaren. Von Severin bis Tassilo 488 - 788, Gemeinsame Landesausstellung des Freistaates Bayern und des Landes Salzburg, Rosenheim/Bayern und Mattsee/Salzburg 1988, München/Salzburg, 198-207

HILDEBRANDT 1966: K. Hildebrandt, Tierknochenfunde aus der Stadt auf dem Magdalensberg bei Klagenfurt in Kärnten 5. Rinderknochen, Naturkundliche Forschungen zu den Grabungen auf dem Magdalensberg 7. Kärntner Museumsschriften 42, Klagenfurt

von den DRIESCH et al. 1989: A. von den Driesch, J. Boessneck, Abschlußbericht über die zooarchäologischen Untersuchungen an Tierknochenfunden von der Heuneburg, in: E. Gersbach, Ausgrabungsmethodik und Stratigraphie der Heuneburg, Heuneburgstudien 6. Römisch-Germanische Forschungen 45, Mainz, 131-57

HORNBERGER 1970: M. Hornberger, Gesamtberurteilung der Tierknochenfunde aus der Stadt auf dem Magdalensberg in Kärnten (1948-1966), Naturkundliche Forschungen zu den Grabungen auf dem Magdalensberg 10. Kärntner Museumsschriften 49, Klagenfurt

EARWOOD 1993: C. Earwood, Domestic Wooden Artefacts in Britain and Ireland from the Neolithic to Viking Times, Exeter

KELLER 1909: O. Keller, Die antike Tierwelt Bd. 1. Säugetiere, Leipzig

EHRET 1964: R. Ehret, Tierknochenfunde aus der Stadt auf dem Magdalensberg bei Klagenfurt in Kärnten 2. Carnivora, Lagomorpha, Rodentia und Equidae, Naturkundliche Forschungen zu den Grabungen auf dem Magdalensberg 2. Kärntner Museumsschriften 34, Klagenfurt

KOKABI 1982: M. Kokabi, Arae Flaviae II. Viehhaltung und Jagd im römischen Rottweil, Forschungen und Berichte aus BadenWürttemberg 13, Stuttgart KOKABI 1988: M. Kokabi, Viehhaltung und Jagd im römischen Rottweil. B. Wildtiere. 3. Elch, Alces alces, in: M. Klee, M. 12

K. Gostenčnik: Elk Antler as a Material of Manufacture Kokabi, E. Nuber, Arae Flaviae IV, Forschungen und Berichte aus Baden-Württemberg 28, Stuttgart, 204-5

RÜLCKER 1986: J. Rülcker, Das Elchwild: Naturgeschichte, Ökologie, Hege und Jagd des europäischen Elches, Hamburg

KOKABI et al 1994: M. Kokabi, G. Amberger, J. Wahl, Die Knochenfunde aus der villa rustica von Bondorf, in: A. GaubatzSattler, Die Villa rustica von Bondorf, Forschungen und Berichte zur Vor- und Frühgeschichte in Baden-Württemberg 51, Stuttgart, 285-335

SACKETT 1992: L.H. Sackett, Knossos. From Greek City to Roman Colony. Excavations at the Unexplored Mansion 2, The British School of Archaeology at Athens Suppl. Vol 21, London SALAMAN 1989: R. A. Salaman, Dictionary of Woodworking Tools c. 1700-1970, and Tools of Allied Trades, revised edition by P. Walker, London/Sydney/Wellington

LENZ 1856: H. O. Lenz, Zoologie der alten Griechen und Römer, Wiesbaden

SCHIBLER et al. 1988: J. Schibler, A. R. Furger, Die Tierknochenfunde aus Augusta Raurica (Grabung 1955-1974), Forschungen in Augst 9, Augst

LETTA 1984: C. Letta, Due letti funerari in osso dal centro romanoetrusco della valle d’Amplero (Abruzzo), Monumenti Antichi. Serie Miscellanea 3,3, Roma

SCHMID 1963: E. Schmid, Die Tierknochen, in: R. Laur-Belart, Hüttenböschen. Ein gallorömischer Vierecktempel am unteren Ende des Walensees, Jahrbuch des Historischen Vereins Kanton Glarus 60, 17-19

LUHMANN 1965: F. Luhmann, Tierknochenfunde aus der Stadt auf dem Magdalensberg bei Klagenfurt in Kärnten 3. Die Schweineknochen, Naturkundliche Forschungen zu den Grabungen auf dem Magdalensberg 4. Kärntner Museumsschriften 39, Klagenfurt

SCHMID 1972: E. Schmid, Knochenatlas. Für Prähistoriker, Archäologen und Quartärgeologen, Amsterdam/London/New York

MacGREGOR 1985: A. MacGregor, Bone, Antler, Ivory and Horn. The Technology of Skeletal Materials since the Roman Period, London/Sydney

SERRA-RIDGEWAY 1996: F. R. Serra-Ridgeway, I corredi del fondo scataglini a Tarquinia, Milano

MIKLER 1997: H. Mikler, Die römischen Funde aus Bein im Landesmuseum Mainz, Monographies Instrumentum 1, Montagnac

SOKOL’SKIJ 1971: N. J. Sokol’skij, Derevoobrabatyvajuščee remeslo v antichnych gosudarstvach Severnogo Pricernomor’ja (Woodworking handicraft in the Antique States of the regions to the north of the Black Sea), Materialy i Issledovanija po Archeologii SSSR 178, Moskva

MIKLER 2001: H. Mikler, Mainz (D). Geweihte Geweihe vom Elch (Alces alces), Bulletin Instrumentum 13, 37

NICHOLLS 1979: R. V. Nicholls, A Roman couch in Cambridge, Archaeologia 106, 1-32

SPANNAGEL 1948: F. Spannangel, Das Drechslerwerk. Ein Fachbuch für Drechsler, Lehrer und Architekten. Auch ein Beitrag zur Stilgeschichte des Hausrats, Ravensburg²; repr. Hannover, 1990

NIELSEN, E. 1995: Aspetti della produzione artigianale a Poggio Civitate, in: E. Formigli (ed.), Preziosi in oro, avorio, osso e corno. Arte e tecniche degli artigiani etruschi, Siena

TOYNBEE 1983: J. M. C. Toynbee, Tierwelt der Antike, Mainz ULBRICHT 1978: I. Ulbricht, Die Geweihverarbeitung in Haithabu, Die Ausgrabungen in Haithabu 7, Neumünster

NOLLÉ 2001: J. Nollé, Die Abwehr der wilden Schweine. Schwarzwildjagd im antiken Lykien, München

ULBRICHT 1984: I. Ulbricht, Die Verarbeitung von Knochen, Geweih und Horn im mittelalterlichen Schleswig, Ausgrabungen in Schleswig. Berichte und Studien 3, Neumünster

MUTZ 1972: A. Mutz, Die Kunst des Metalldrehens bei den Römern, Basel/Stuttgart

PICCOTTINI 1977: G. Piccottini, Die Stadt auf dem Magdalensberg - ein spätkeltisches und frührömisches Zentrum im südlichen Noricum, in: H. Temporini, W. Haase (eds.), Aufstieg und Niedergang der römischen Welt II 6, Berlin/New York, 263301

URBAN 1994: O. H. Urban, Der Braunsberg, Keltische Höhensiedlungen an der mittleren Donau vom Linzer Becken bis zur Porta Hungarica 2. Linzer archäologische Forschungen 23, Wien VAULINA et al. 1974: M. Vaulina, A. Wąsowicz, Bois grecs et romains de l’Ermitage, Wrocław

PICCOTTINI 1989: G. Piccottini, Bauen und Wohnen in der Stadt auf dem Magdalensberg, Studien zur Ur- und Frühgeschichte des Donau- und Ostalpenraumes 4. Denkschriften der österreichischen Akadamie der Wissenschaften in Wien, phil.-hist. Kl., Vol. 208, Wien

VETTERS 1954: H. Vetters, Die Personennamen vom Magdalensberg, in: Beiträge zur älteren europäischen Kulturgeschichte. Festschrift für Rudolf Egger vol. 3, Klagenfurt, 32-45

PICCOTTINI et al. 1990: G. Piccottini, H. Vetters, Führer durch die Ausgrabungen auf dem Magdalensberg, Klagenfurt4

VOLBACH 1976: W. F. Volbach, Elfenbeinarbeiten der Spätantike und des frühen Mittelalters, Kataloge vor- und frühgeschichtlicher Altertümer 7, Mainz³

PICHLER 1976: W. Pichler, Die Knochenfunde aus dem spätrömischen Kastell Vemania, PhD Dissertation München University; non vidi

WHITE 1975: K. D. White, Farm Equipment of the Roman World, Cambridge

PINELLI et al. 1986: P. Pinelli, A. Wąsowicz, Musée du Louvre. Catalogue des bois et stucs grecs et romains provenant de Kertch, Paris PRELL 1941: H. Prell, Die Verbreitung des Elches in Deutschland zu geschichtlicher Zeit, Leipzig

WÜRGLER 1958: F. E. Würgler, Die Knochenfunde aus dem spätrömischen Kastell Schaan (4. Jahrhundert n.Chr.), Jahrbuch des Historischen Vereins für das Fürstentum Liechtenstein 58, 255-282

RGA² vol. 7, s.v. Elch: H. Beck, H. Jankuhn, K. Ranke, R. Wenskus (eds.), Reallexikon der germanischen Altertumskunde vol. 7, Berlin/New York², 1989

Addendum

RIC Philip I: H. Mattingly, W. A. Sydenham, C. H. V. Sutherland (eds.), The Roman Imperial Coinage Vol. IV, Part III, London, 54-95

In the course of the excavations in an Imperial fabrica on the Magdalensberg (production of gold ingots for Caligula, emperor 37-41, as revealed by ingot-casts with his inscriptions, and furnaces for melting gold, cf. now G.

RIETH 1955: A. Rieth, Antike Holzgefäße, Archäologischer Anzeiger, 1-26 13

Materials of Manufacture

Piccottini, Anzeiger der Österreichischen Akademie der Wissenschaften, phil.-hist. Kl. 136, 2001, 41-67), the uppermost slope-wash produced a new piece of elk antler (Pl. 9). The find is sawn at its bottom (Pl. 9, right), and has traces of an animal gnawing at it (Pl. 9, middle). It has been removed close to the edge of an antler, and therefore it is discarded refuse rather than raw material for any further use; one edge is broken.

worked bones, dark brown with finds from the Magdalensberg is caused by the conditions of the immediate matrix in the strata rather than by a colouring process. As dark brown always penetrates the artefact to the core, as does green, pieces like Catalogue N°. 3 received their colouring most positively by means of charred debris inter alia rather than through a dyeing process, or were exposed to fire. I meanwhile doubt artificial colouring whenever the colour penetrates the whole piece (cf. also Gostenčnik in preparation). One might consider an intentional dyeing of raw materials before their processing, but then the bones must have remained in a dyeing bath long enough to be soaked with colours all through. With what is left as workshop waste at the Magdalensberg, this can, however, not be confirmed. On the other hand, it also does not mean that it could not have been applied with the material of other sites. Colouring is reported only vaguely, so we are here more or less at the very beginning of our researches. As far as I can tell from the Magdalensberg material, only bone has been dyed, with antler, however, I could not yet confirm it positively.

Artefact Catalogue 6. (Pl. 9): Prov. MB 2001, eastern AA buildings, slope wash; Dat. no closer dated context; L. bottom 8 cm, H. point 9 cm; We. not available. Sawn piece of an elk antler with imprints of animal teeth; few traces of antler structure preserved on the surface. For Austria, an elk talus (again!) is now also confirmed in a Late Roman (AD 3rd/4th century) hillfort in the Styrian Ennstal in Ramsau (S. Czeika, Tierreste aus einer spätrömischen Rückzugssiedlung in der Gemeinde Ramsau am Dachstein, Steiermark, Anisa 21, 2000, 11-14).

As to wickerwork, in the1950s, a storage vault on the Magdalensberg seems to have produced traces of wickerwork dating immediately after 50 B.C., as has been noted in an old field-journal (Eleni Schindler-Kaudelka, personal comment).

Since I wrote the manuscript, my investigations partly made me change my views stated above. As to the colouring of

14

R. Bendrey: The Identification of Fallow Deer (Dama dama) Remains from Roman Monkton...

THE IDENTIFICATION OF FALLOW DEER (DAMA DAMA) REMAINS FROM ROMAN MONKTON, THE ISLE OF THANET, KENT Robin BENDREY

INTRODUCTION

is therefore of some importance. The deer remains will be first described, and then placed into the context of other fallow deer finds from archaeological sites in Kent, Britain and Europe.

The purpose of this report is to detail the identification of a number of fallow deer bones recorded in an animal bone assemblage from a rural Roman settlement from Monkton, Kent. Canterbury Archaeological Trust excavated the site during 1994-1995 along the line of the proposed MonktonMount Pleasant dual carriageway.

METHODOLOGY Similarities between the skeletons of fallow deer, Dama dama (L.), and red deer, Cervus elaphus L., can cause problems of identification. The modern red deer is a larger animal than the fallow. Size alone is not a sufficient criterion for identification, as the range of size variation for each species creates a degree of overlap. Lister (1996) has defined morphological characters for the separation of the bones of the two species. The deer bones from Monkton have been identified using Lister’s characters, and if no characters are present the bones have been identified on the basis of size.

Roman deposits at Monkton produced over five thousand bone fragments through hand-excavation (Bendrey forthcoming a). Sheep is the most commonly recorded animal, followed by cattle, horse and then pig. Cattle contribute the most by bone weight. Small numbers of dog, cat, goat, hare, red deer, and whale were also recorded. Nine fragments of fallow deer were identified from the assemblage. The fallow deer was present in Britain during the Hoxnian and Ipswichian inter-glacials (Yalden 1999). The reintroduction of fallow deer to Britain is variously attributed to the Romans (e.g. Clutton-Brock 1987, 182) and the Normans (e.g. Rackham 1986, 49). A number of archaeological sites have produced evidence of fallow deer from Roman contexts, though several of these suffer from problems of misidentification and contamination. The identification of this species in securely dated Roman contexts

Table 1 details the expression of Lister’s (1996) identification characters in the post-cranial elements, and the measurements obtained. All measurements given follow the method of von den Driesch (1976). Lister (1996, 123-34) gives full descriptions and diagrams of all the characters discussed below. Scoring of the different characters in Table 1 uses the following structure (ibid., 121):

Table 1. Scoring of Lister’s (1996) identification characters for the fallow deer remains, and measurements. Element

Context

Character

Character State

Measurements (mm)

scapula

967

3

d

SLC = 27.1

5

D

scapula

933

1

D

2

D

BG = 30.5

3

=

SLC = 25.1

4

c

LG = 35.3

5

D

GLP = 43.6

radius

6176

-

no data, identified on basis of size

Bd = 31+/-2

metacarpal

849

1

D

Bp = 27.3

2

D

SD = 15.9

3

D

4

c

-

no data, identified on basis of size

6289

-

no data, identified on basis of size

SD = 16+/-1

800

1

D

Bp = 30+/-1

2

D

SD = 17+/-1

3

C

metacarpal

849

metacarpal metatarsal

15

Materials of Manufacture

‘D’ specimen clearly shows the putative ‘Dama’ condition

generally more friable bone condition. This specimen may be intrusive.

‘d’ specimen more like D than like C ‘=’ specimen shows state intermediate between D and C, or unlike either

Context 933 produced a nearly complete right scapula. Its morphology and size clearly identify it to fallow deer (Table 1).

‘c’ specimen more like C than like D ‘C’ specimen clearly shows the putative ‘Cervus’ condition

Damage to the distal radius from context 6176, has obliterated the morphological features that might have proven diagnostic. The small size of the bone has been used to identify it to fallow deer.

OSTEOLOGICAL DESCRIPTION

Two fragments of metacarpal were identified from context 849. The larger fragment, consisting of the proximal half of a left metacarpal, is identified to fallow deer on both its size and morphology (Table 1). The other specimen from this context is a small anterior shaft fragment from a proximal metacarpal. The specimen bears no diagnostic characters or metrical data, but is comparable in size to the other fragment from the same context.

Nine fragments of fallow deer have been identified: two shed antlers, three metacarpals, one metatarsal, two scapulae, and one radius. The location on site, and pottery dates, of the finds are given in Table 2. A minimum number of three animals are represented by the bones, counting the three left metacarpals. The epiphyses of the two scapulae and the distal radius were fused. Antlers are very variable in size depending on age, condition and genotype (Chapman, 1977, 431), this variation means that age cannot be accurately attributed from their size and shape (Chapman, 1984, 8). The shed specimens from Monkton are incomplete, but are from a second head of antlers or older. Antlers are shed April to June, with older animals casting before younger ones (Chapman, 1977, 431).

Context 6289 produced the shaft from another left metacarpal. The articular ends are missing and the surface is slightly eroded, but it is comparable in size to the other fallow deer fragments (see SD measurements in Table 1). Finally, the proximal three-quarters of a right metatarsal is recorded from context 800. Erosion of the bone surface obscures the analysis of characters from this specimen. Characters 1 and 2 suggest that the specimen is fallow deer, in that the proximal articular facets are level where they meet and appear to meet or nearly meet in proximal view (Lister, 1996, 133). Character 3, appertaining to the foramen in the centre of the proximal articulation of the metatarsal, appears to indicate red deer (Table 1): red deer has a single, large foramen compared to fallow, which has ‘a network of pores in thin bone’ (ibid.). The surviving length of the bone is 18 cm, and it probably would have had a total length, when complete, of about 21 cm. This specimen has been identified to fallow deer on the basis of characters 1 and 2, and according to its size.

A shed left antler fragment derived from context 1345. The absence of a bez tine, and the oblique angle of the brow tine with the beam identifies this fragment to fallow (Lister, 1996, 122-3). The trez tine is present as also is the proximal part of the palmation. The pearling is very weak at the base. Context 967 produced a shed right antler fragment. The absence of a bez tine, the oblique angle of the brow tine with the beam and the absence of pearling identifies this specimen to fallow. Only characters 3 and 5 were available from the left scapula fragment from context 967. This specimen had numerous cut, and a few chop marks transversely across the neck, below the post-spinous fossa. This specimen, however, differs from the rest of the deer bones from Monkton in the quality of bone preservation. The surface preservation is pristine and the bone itself appears denser compared to the other specimens which exhibit a slightly eroded surface and a

The evidence for Roman Fallow Deer in Britain and NorthWestern Europe The re-introduction of fallow deer to Britain is most commonly ascribed to the Romans or the Normans, though

Table 2. Distribution and date of fallow deer finds from Roman Monkton. Group

Context

Type

Pottery date

Ditch/Gully Group 7

6289

Fill of 679

300-400

1 metacarpal.

Pit/Feature 37

849

Fill of 581

175-200/250

2 metacarpals.

967

Fill of 581

175-200/250 nd

Fallow deer remains

1 antler, 1 scapula. rd

Pit/Feature 44

6176

Fill of 6080

Late C2 - C3

Structure 3

1345

PH (?cut)

-

1 antler.

Structure 25

933

SFB infill

300-350/400

1 scapula.

Structure 26

800

SFB infill

275-350/400

1 metatarsal.

16

1 radius.

R. Bendrey: The Identification of Fallow Deer (Dama dama) Remains from Roman Monkton...

DISCUSSION

it has also been attributed to the Phoenicians, Gauls and Ancient Britons (Yalden, 1999, 153). Yalden (1999, 153-6) lists archaeological sites that have produced finds of fallow deer and shows that they were not common until the medieval period. Naomi Sykes (2001) has extensively reviewed the evidence for the re-introduction of fallow deer to Britain, and has shown many of the pre-Norman finds to be misidentified or probably intrusive. The former includes the Roman fallow deer remains from Portchester Castle (Grant 1975), which are often quoted in arguments for a Roman introduction. Amongst the secure finds which remain valid are shed antlers from St Albans, and Scole-Dickleburgh, although these may be accounted for by trade rather than a resident population (Sykes 2001). The shed antler from ScoleDickleburgh has been radio-carbon dated to 3rd-6th centuries AD (Baker 1998, 16). Baker also refers to post-cranial bones from Roman deposits in London that may suggest the importation of venison to England (West 1983; referenced in Baker 1998, 17).

The discovery of nine fallow deer bones and the range of skeletal elements represented from Roman Monkton stand out in the archaeozoological record of Roman Britain and north western Europe. Other sites from Kent concur with the general picture from the rest of Britain that fallow deer are not common until the medieval period (Bendrey forthcoming b; Driver 1990; Wall 1980). The identification of fallow deer from Roman Monkton is secure, although the strengths of the individual identifications vary. The pottery evidence gives a variety of dates covering the second, third and fourth century (Table 2). The deer bones form no spatial pattern, although the largest concentration, four fragments, came from pit/feature 37. Pit/feature 37 contained over 1,200 hand-recovered bone fragments, approximately one quarter of the total site assemblage. King’s (1978) survey of Roman bone assemblages found that the percentage of cervid bones was usually less than 5% of total cattle, sheep and pig bones. This is certainly true for Roman Monkton. In fact there are few other wild mammals present; including only two bones of hare, one red deer antler and one fragment of whale.

Antler may have been imported as a raw material for working (there is some evidence for butchery on the antler from ScoleDickleburgh, although the specimen is almost complete (Baker 1998, 16)). Alternatively, shed antler may represent trade in religious relics (Sykes 2001). The presence of fallow deer foot bones on Roman sites may be associated with the adoption of Greek religious beliefs by the Romans (astragali have been found in religious contexts from Greece) (ibid.). Fallow deer were often in Roman art as the companion of Diana and they may have been used as offerings to her in Temples (Schmid 1965, 60).

The range of elements represented do not fit a pattern of foot bones and antlers associated with trade in raw materials or religious relics. Also, the fallow deer finds are not derived from any ‘special’ context, but rather from infilling of sunkenfeature buildings and fills incorporating waste from pits. Two explanations can be put forward for their presence at Monkton. Firstly, they may represent trade in salted meat. This would probably not include the foot elements which do not bear much meat, though these may be present if trade in raw materials or relics was occurring. Secondly, they may represent live animals from the locality. The presence of live animals would seem unique in Roman Britain from the archaeological evidence. The large number of Roman bone assemblages now published which do not contain fallow is testimony to their sparse existence. If, as it appears, fallow deer were not introduced into the native fauna, this does not mean that they were not introduced into parks. Kent, with its proximity to continental Europe, would seem a logical place through which an introduction might have occurred. The transport of fallow deer across the English Channel would have been a fairly simple process. Clutton-Brock (1987, 167) illustrates this with a photograph of spotted deer, a similarsized species, being transported in boats.

Finds from the continental European countries closest to Britain are also sparse. The sole fallow deer bone from the Roman Netherlands is a worked metatarsal from Valkenburg (Prummel 1975, 227). There are no fallow deer bones known from Roman Belgium (Ervynck et al, 1999; referenced in Sykes 2001). Four fallow deer bones have been recovered from separate sites from northern France; including three antler fragments and an astragalus (Lepetz et Yvinec, sous presse). There is limited documentary evidence for Roman fallow deer. The Romans established large hunting parks in mainland Europe. Columella refers to one in Gaul, covering 4,000 hectares, that contained fallow deer (Chapman and Chapman 1975, 47). Although fallow deer were kept in parks in Italy (Rackham, 1986 49; Yalden 1999, 128), it is not known whether any such parks were established in Britain. There is uncertain evidence of fallow deer appearing in the shows of the emperors (Chapman and Chapman 1975, 46). Toynbee (1973 144) refers to the emperor Gordian I’s collection, that ‘contained dammae ducenti and two hundred stags with antlers shaped like the palm of the hand, including some British stags (cervi palmate ducenti mixtis Britannis)’. Yalden (1999 153) however suggests that these animals are more likely to be elk than fallow deer (see Gostenènik, this volume). Schmid (1965, 60) proposes, as a possibility, that fallow deer bones identified from Roman Augst, Switzerland, may have been from show animals; exotic animals kept as a statement of authority and wealth.

CONCLUSIONS Nine fragments of fallow deer, including two antlers, two scapulae, one radius, three metacarpals, and one metatarsal, have been identified from a rural Roman settlement from Monkton, Kent. Monkton differs from the few other sites from north western Europe producing fallow deer bones, in the variety of bones represented. The presence of these bones at Monkton could represent either traded salted meat and 17

Materials of Manufacture

raw materials, or the introduction of live fallow deer to Britain during the Roman period. It is clear from their sparse presence in the archaeological record that fallow deer were not introduced into the native fauna of Britain, but the possibility remains that the introduction of live animals into controlled parks may have occurred.

ERVYNCK A., VAN NEER W., LENTACKER A., 1999, Introduction and extinction of wild animal species in historic times: the evidence from Belgium, in, Benecke N. (ed.), The Holocene history of the European vertebrate fauna, Modern aspects of research workshop, 6th-9th April 1998, Berlin GRANT A., 1975, The animal bones, in Cunliffe B W (ed.), Excavations at Portchester Castle I; Roman, London, 378-408 KING A., 1978, A comparative survey of Bone Assemblages from Roman sites in Britain, Bulletin of the Institute of Archaeology 15, 207-32

Acknowledgements I would like to thank Polydora Baker, Naomi Sykes and JeanHervè Yvinec for sending me information and reports on Roman finds of fallow deer, and Ian Riddler and Bev Bendrey for reading and commenting on earlier versions of this text.

LEPETZ S. et YVINEC J.-H., sous presse Présence d’espèces animales d’origine méditerranéennes en France du nord aux périodes romaine et médiévale: actions anthropiques et mouvements naturels, in, Gardeisen A. éd., Mouvements ou déplacements de populations animales en Méditerranée au cours de l’Holocène, British Archaeological Reports, International Series, Oxford

Bibliography

LISTER A.M., 1996, The Morphological Distinction Between Bones and Teeth of Fallow Deer (Dama dama) and Red deer (Cervus elaphus), International Journal of Osteoarchaeology 6, 119-43

BAKER P., 1998, The vertebrate remains from Scole-Dickleburgh, excavated in 1993 (Norfolk ands Suffolk), A140 and A143 Road Improvement Project. Anc. Monuments Lab. Rep. 29/98, London

PRUMMEL W., 1975, Some reflections on the faunal remains of the Roman castellum Valkenburg, excavations 1962, in, Clason A T (ed.), Archaeozoological studies, Groningen, 225-9

BENDREY R., forthcoming a The Animal Bone, in, Rady J., editor, Roads to the Past. The Monkton Landscape. Prehistoric, Roman, Anglo-Saxon and Medieval sites along the duelling of the A253, Canterbury Archaeological Trust Occasional Papers, Canterbury

RACKHAM O., 1986, The History of the Countryside, London SCHMID E., 1965, Damhirsche im römischen Augst, Ur-Schweiz 29, 53-63

BENDREY R., forthcoming b, The Mammal Bone, in Parfitt K., Corke B. and Cotter J., Excavations off Townwall Street, Dover, 1996, Canterbury Archaeological Trust Occasional Papers, Canterbury

SYKES N., 2001, The Norman Conquest: A Zooarchaeological Perspective, PhD thesis, University of Southampton TOYNBEE J.M.C., 1973, Animals in Roman Life and Art, London

CHAPMAN D., 1977, Fallow deer Dama dama, in, Corbet and Southern (ed.), The Handbook of British Mammals, Oxford, 42937

von DEN DRIESCH A., 1976, A Guide to the Measurement of Animal Bones from Archaeological Sites, Peabody Museum Bulletin 1, Washington

CHAPMAN D. and CHAPMAN N., 1975, Fallow deer, Lavenham

WALL S.M., 1980, The animal bones from the excavation of the hospital of St. Mary of Ospringe, Archaeologia Cantiana 96, 22766

CHAPMAN N., 1984, Fallow deer, Oswestry CLUTTON-BROCK J., 1987, Domesticated animals from early times, London

WEST B., 1983, The Roman buildings west of the Walbrook Project: Human, animal and bird bones, Level III. Unpublished manuscript, Department of Urban Archaeology, Museum of London, London

DRIVER J.C., 1990, Faunal Remains, in Driver J.C., Rady J., and Sparks M., Excavations in the Cathedral Precincts, 2 Linacre Garden, ‘Meister Omers’ and St Gabriels Chapel, The Archaeology of Canterbury IV, Maidstone, 228-57

YALDEN D., 1999, The History of British Mammals, London

18

M.T. Bíró: Recycling worked bone in Pannonia: Data on the curation of workshop debitage and worn/damaged objects...

RECYCLING WORKED BONE IN PANNONIA: Data on the curation of workshop debitage and worn/damaged objects in the Roman bone manufacturing industry Maria T. BÍRÓ

Recently, an increasing number of carved bone objects have been encountered in 4th-5th century settlements and cemeteries of the former province of Pannonia (Bíró 1994). The forearms of women found in cemeteries are frequently decorated with eight to ten bone bracelets. Double-sided, fine combs are found by their heads or feet, and single-sided, ‘beetle-backed’ combs were commonly placed on the knees of buried women and children. Bone hairpins of varying sizes and shapes held the hair or fastened headscarves. Bone needles with zoomorphic designs, commonly found in the graves of men, held the clothes together. The quantum leap in the number of bone artefacts in Late Roman grave inventories may also be explained by a change in burial rites. From the 3rd century onwards, the custom of cremation was gradually replaced by placing the deceased into coffins and sarcophagi. The change in ritual behaviour is probably the result of the spread of eastern mystery religions which laid emphasis on the hope of resurrection. This shift accelerated and ultimately became obligatory as Christianity was recognized as a state religion. Consequently, objects of attire that would have been incinerated during the cremation process were well preserved. This holds especially true for bone artefacts. In cremation burials where human bones were also turned into ashes, at best the calcined remains of ivory inlay caskets could be identified. Although the custom of cremation rites was abandoned throughout the European provinces of the Roman Empire, the fashion in bone carving seems to cease at the western border of Pannonia. It may be hypothesized, therefore, that the mass appearance of bone objects was related to the increasing poverty of the limes regions of the frontier province.

empire. It is perhaps the economic mechanisms of the Roman Empire which were most reminiscent of those in modern consumers’ societies. The enormous market that developed within the empire created a demand for more-or-less identical products between the Strait of Gibraltar and Persia. This largely urbanized market was supplied by mass produce from factories and workshops. However, while the overall level of demand was comparable, it was backed up by different capital power. Since, in the absence of mechanization, workshops in the empire could not produce series of goods, the masses of poor customers were supplied with products that were made from cheap raw materials but with the formal resemblance retained. That is, many centuries ago the tendency during the times of regression was for fashionable items to be made from cheap materials in order to satisfy the demands of poor customers as well. A spoon carved from bone may be just as beautiful as a bronze one; however, its life span is significantly shorter. In late Roman period provinces emerging bone manufacturing industries simply applied a merchandising system that has worked perfectly throughout the centuries. On the other hand, a new phenomenon also occurred: the numbers of curated utilitarian and decorative bone objects increased. Pieces that had been discarded in previous periods were re-used. On the basis of archaeological finds, two forms of curation may be reconstructed. First, re-use was made of manufacturing debris produced in the workshops. These workshops can be easily recognized both in Roman period and medieval excavations. They may be found in areas where unusually high numbers of defective carvings and blanks come to light. Masses of bone fragments, not suitable for further manufacturing, form the debris deposited around all of these of workshops. These include sawn-off antler tines of red deer and forked antler fragments, as well as the chopped off parts of antler burrs. In the refuse pit of the bone-manufacturing workshop excavated in the medieval Royal Palace of Buda all these antler fragment types were identified as useless debitage unsuitable for additional processing. At late Roman period settlements, however, even such pieces were utilized. Seemingly useless antler tine tips were decorated with tiny circles and perforated at their thicker end in order to serve as toggles used in tying ropes during the loading of pack animals (Figure 1). A similar piece of antler tine, drilled through in the middle, may have served as a cheek-bit. Straighter and sturdier tine fragments could be turned into massive awls or perforators (Figure 2). Forked antler segments could be turned into handles for drills and chisels in domestic circumstances (Figure 3).

The fourth century richness of bone industries may also be related to a decrease in exports resulting from existential insecurity, as well as to the radical decline in monetary transactions which determined the history of the province during its last century. Expensive, noble materials (silver, bronze etc.) were increasingly replaced by cheaper bone. This material could be used in making the same types of objects as precious metals and bronze. Fourth century bone pins are carved into exactly the same shape as their 3rd century metal precursors. Bone spoons were also shaped to imitate their metal counterparts. Surprisingly enough, the range of motives used in the surface decoration of other types of bone artefacts imitates designs seen on glassware. The hypothesis that the astonishing boom in bone manufacturing was caused by general impoverishment in fourth century Pannonia is also supported by trends that were characteristic of the production structure over the entire 19

Materials of Manufacture

A

B

C

Figure 1 Cut off red deer antler tines: A: Tác - Gorsium; B: Dunaújváros/Dunapentele - Intercisa and C: rope-tying toggle: Esztergom - Solva

Figure 2 Awl (Tác - Gorsium)

Figure 3 Drill handle (Szombathely - Savaria) 20

M.T. Bíró: Recycling worked bone in Pannonia: Data on the curation of workshop debitage and worn/damaged objects...

The use of antler burrs is even more special. The oval shape and ornate natural rim of this antler element practically invited manufacturing into medals or pendants. One such bone disc was perforated at its top, while its centre was decorated with a phallic design. Many such domestically made amulets are known, all of them prepared from burrs that would be otherwise considered workshop debitage. Due to the belief among Roman women that wearing images of the male organ would have a beneficial effect on their fertility, European museums are stocked with masses of phallic images made from gold, bronze etc. which occur in varying sizes and were decorated with various designs. They were worn on the necklaces of high ranking Roman ladies. Primitive imitations of these amulets were carved from debitage and bone epiphyses by the poor (Figure 4).

archaeological sensation of European dimensions. An Indian import comb fragment came to light during the course of excavations at the site of Tác - Gorsium at the beginning of the 1970’s (Bíró 1985). It was decorated with the bust of a woman playing an ancient Indian instrument (biwa). This representation belongs to the Gandhara stylistic sphere and is related to the worked ivory finds from Taxila and Bégram (Marshall 1951; Hackin 1954). Aside from this Tác Gorsium comb only a single original Indian artefact is known from Classical times: an ivory statuette of Lakshmi that was found in Pompeii (Wheeler 1955). However, the ivory plate recovered at Tác - Gorsium did not find its way to Pannonia through trade routes. It must have been brought back to Europe as booty by soldiers in the army divisions (Leg. I. and II. Adiutrix) that took part in the Parthian wars. Originally, this representation must have decorated an ivory casket, and the comb with its several figures must have been cut out of it (Figure 5). Its raw material, a re-used pictorial tablet, must have been of a genre similar to the dressing scene depicted on the Bégram piece (Figure 6). The comb must have been made from the lid of a casket similar to that shown in this latter illustration. The image of the woman in the comb’s decoration is damaged even at the otherwise intact

The fourth century impoverishment, production lulls and diminishing market demand not only inspired the utilization of refuse materials. It also encouraged the repair of damaged bone artefacts as well as their transformation into other objects for secondary use. Antecedents to this may also be followed in this case and one example of recycling from the early period of Pannonia may be considered an

A

B

Figure 4 Amulets made from antler burrs: A: Dunaújváros/Dunapentele - Intercisa; B: Szombathely - Savaria 21

Materials of Manufacture

been the high value raw material and the exotic design that inspired the re-using of the ivory plate. Curation of bone artefacts in the fourth century, on the other hand, was brought about by economic decline. Fibulae, widely used in provincial attire never entirely replaced the undoubtedly more archaic clothing pins. Although such pins that have survived from provincial sites are smaller than their precursors shown on Greek vases they are just as beautifully decorated. The heads of these pins were often carved with zoomorphic motives such as red deer and bear figurines. The most commonly depicted animals include birds (roosters or doves). In Figure 7 such a specimen is shown. The pin was found in a burial in Tác - Gorsium (Bíró 1987) and is worn almost beyond recognition. This dramatic degree of erosion shows how much such small pieces of jewellery were appreciated. During the last century of the empire, damaged or worn jewellery did not end up in the garbage. Their owners wore them until their deaths. The third example is even more expressive. This rooster-decorated pin (Figure 8) was found by the upper arm bone of a man’s skeleton in Grave 95 of the aforementioned Tác - Gorsium cemetery. Although the original small, carved rooster shape had been completely eroded, the pin was not discarded. Instead, its decoration was renovated by carving a secondary grid-pattern on it. Intersecting lines entirely cover the original surface, ignoring the previous design, and cover the surface from the rooster’s crest to its feet on the pin’s handle. A similarly renovated pin is known from Dunaújváros/Dunapentele - Intercisa (Figure 9). It is at the time of the fourth century that this simple motif starts appearing on the side-plates of combs and the cylinders of bone pins and even begins to dominate in pottery design. It corresponds in practice to the decoration applied on the smooth surfaces of Hunnic Period ceramics. On some pins, the cylindrical head is decorated by nothing but rough abrasion above the transversal ribs carved on the pin’s neck. Such pieces look like primitive, homemade carvings. A lucky coincidence helped in interpreting this pin type when a large pin with a damaged turban head decoration was found. The separated head part could be easily put back to its place since it had been made of a separate piece of bone into which the pin itself had been carefully inserted. During use, however, this decorative head could be easily caught in things and was lost. Discoloration on the butt end of surviving pins without heads recovered from burials indicates that these artefacts remained in use long after the loss of the pin’s head. Given the long temporal perspective, the small micro historical data presented in this paper may seem like petty problems. Impoverished provincial people in Pannonia, however, deserve respect for the faithful care and curation applied to the objects they used. These were not only manifestations of poverty but also indicate an appreciation for crafts people’s work. One should not forget about the fact that by taking good care of these simple, mundane belongings in the war torn limes region, people tried to hold on to a material culture which was hopelessly vanishing. All their personal objects that were saved, curated and re-used prolonged the agony of an empire. Wherever archaeologists discover such pieces they can only say: ‘These are still the graves of Romans...’.

Figure 5 Fragment of an Indian comb (Tác - Gorsium)

edge of the comb. Her right elbow and the top of her turban were left out of the comb’s decoration. Consequently, the original picture must have been larger than the section included within the present decoration on the comb. Unfortunately, it is not possible to tell whether the original fragment of an ivory casket inlay or some other broken furniture mount was turned into a comb in India or Pannonia. Should this latter have been the case, however, it is unlikely that at the early date associated with this find, curation would have been motivated by economic necessity. It would have 22

M.T. Bíró: Recycling worked bone in Pannonia: Data on the curation of workshop debitage and worn/damaged objects...

A

B

Figure 6 Lid of ivory casket (A: Bégram); Ivory carving (B: Bégram) 23

Materials of Manufacture

Figure 7 Heavily worn bird-shaped pin (Tác - Gorsium)

Figure 8 Renewed bird-shaped pin (Tác - Gorsium)

Bibliography BÍRÓ, M.T., 1985, The Indian ivory comb from Gorsium. Acta Arch. Hung. 37, 419-30 BÍRÓ, M.T., 1987, Gorsium bone carvings, Alba Regia 13, 25-63 BÍRÓ, M.T., 1994, The Bone Objects of the Roman Collection, Catalogi Musei Nationalis Hungarici II, Budapest BÍRÓ, M.T., 1995, Pannóniai csonthulladékok (Bone debris in Pannonia), História 5-6, 49-50 HACKIN, J., 1954, Nouvelles Recherches Archéologiques á Bégram, Paris MARSHALL, J., 1951, Taxila I-III, Cambridge WHEELER, R.E.M., 1955, Rome Beyond the Imperial Frontiers, Harmondsworth

24

Figure 9 Renewed bird-shaped pin (Dunaújváros/Dunapentele - Intercisa)

A. Thuet: Un Atelier de Production de Peignes en Bois de Cerf de la fin de l’Antiquité tardive...

UN ATELIER DE PRODUCTION DE PEIGNES EN BOIS DE CERF DE LA FIN DE L’ANTIQUITÉ TARDIVE À SAINT-CLAIR-SUR-EPTE (EURE) Annick THUET1

CONTEXTE DE LA DÉCOUVERTE

d’autant plus que la quantité de pièces inventoriées est importante. A titre de comparaison, seul un artisanat tardif de ce type a été découvert à Drevant, dans le Cher (Béal 1984). On y produisait différentes pièces comme des fusaïoles principalement mais aussi des clés, des plaques et d’autres objets difficilement identifiables. Un peu plus tardif, le site mérovingien de Huy, en Belgique, a livré seulement 76 éléments en bois de cerf, vestiges d’atelier de «production de plaquettes de fixation ou de garniture de peigne» (Gob 1986, p. 195). Enfin, plus éloigné de nous, on peut mentionner la découverte en Suisse, à Rheinau, d’un atelier de fabrication de peigne triangulaire en bois de cerf (93 éléments recensés) datant de la fin du IVè ou du début du Vè s. ap. J.-C. (Hedinger 2000).

De nombreux vestiges d’une production artisanale sur bois de cerf ont été découverts dans un fond de cabane, St. 50, daté par la céramique du milieu du Vè s. apr. J.-C. Il s’agit du contexte mobilier le plus ancien découvert sur le site. De forme quadrangulaire, mesurant un peu plus de 3 m de long sur 2,20 m de large, la fosse est percée de 8 trous de poteaux régulièrement répartis près des parois. Peu profonde, elle recélait deux niveaux de remblai, dont le premier, le plus ancien, pourrait s’apparenter à une phase d’occupation mesurant environ 4 cm d’épaisseur et remontant le long des parois. Dans cette couche, quelques éléments en bois de cerf ont été découverts. Mais ils étaient essentiellement concentrés dans le niveau de remblais qui la scellait, mesurant 20 cm d’épaisseur et correspondant à l’abandon de la structure.

D’après Mme Poulain-Josien, l’existence d’un atelier de tabletterie est attestée par la réunion de quatre types de vestiges : une réserve de matière première, des fragments de bois ayant subi un début de façonnage, des résidus de travail et des objets fabriqués (Poulain-Josien 1976). A priori, il semble qu nous soyons en présence d’un atelier de traitement du bois de cerf. L’étude des vestiges pourra confirmer l’usage de la structure 50 comme local de production.

La céramique commune découverte dans ces niveaux présente un intérêt particulier dans la mesure où elle associe des formes gallo-romaines tardives et les premiers éléments d’une production mérovingienne. Le fond de cabane peut être alors attribué à une phase de transition, conférant aux vestiges en bois de cerf un intérêt chronologique certain.

La technique de fabrication de peignes est désormais bien connue grâce au travail de synthèse de Michel Petitjean (Petitjean 1995) portant sur l’Antiquité tardive et l’époque mérovingienne. L’analyse de la chaîne opératoire du site de Saint-Clair-sur-Epte reprend t’elle exactement le même procédé ou existe t’il des variantes significatives ?

Mélés à quelques éléments osseux correspondant à des déchets alimentaires, 453 fragments, rejets de taille et objets semi-finis ont été recensés dans cet espace. Ils permettent de suivre l’ensemble d’une chaîne opératoire de débitage sur bois de cerf menant à la fabrication de peignes décorés à double denture. La présence d’une épingle à cheveux peutêtre en os, en cours de façonnage, pourrait indiquer un travail secondaire (voire occasionnel) de fabrication de pièces osseuses.

Le bois de cerf est employé du Ier au IIIè s. ap. J.-C. de façon secondaire dans les ateliers de tabletterie, l’os étant de loin la matière principale. Toutefois, au IVè s, ces données s’inversent à Rouen où le bois de cerf devient le matériau principal (Rouen 1982, p. 79). Situé sur un axe de communication important entre Rouen et Paris, les vestiges de ce site pourraient s’inscrire dans cette tendance.

INTÉRÊT D’UNE TELLE ÉTUDE Peu d’ateliers de production de tabletterie utilisant le bois de cerf comme matériau principal ont été mis au jour et étudiés, conférant à cette découverte un caractère peu ordinaire, 1

TERMINOLOGIE RELATIVE AU BOIS DE CERF (BILLAMBOZ, 1979)

INRAP Nord-Picardie, 518, rue Saint-Fuscien, 80000 Amiens. Je tiens à remercier chaleureusement M. Riddler, à l’origine de cette publication, pour son entière disponibilité et son écoute attentive, ainsi qu Mr. Léon, responsible de l’opération archéologique, qui m’a confié l’étude de cet atelier. Mes remerciements sont également à ceux qui sont à l’origine des ces collaborations, MM. Lepers et Mantel du SRA de HauteNormandie et M. Talon, directeur de l’antenne INRAP Nord-Picardie. Merci enfin aux archéozoologues du CRAVO de Compiègne et à M. Ranson, taxidermiste, pour l’aide qu’ils m’ont apportée dans la determination des bois.

Il est important de rappeler en préambule la signification de quelques termes propre au bois de cerf, que nous utiliserons souvent par la suite (Figure 1): - bois de mue : bois séparé naturellement du crâne lors de la chute annuelle (entre février et mai, suivant l’âge de l’animal). 25

Materials of Manufacture

Figure 1 Terminologie relative au bois de cerf, d’après Billamboz 1979, 96

1 Classement des Vestiges selon leur position d’origine dans la Ramure et/ou leur état d’avancement dans la Chaîne Opératoire

- bois de massacre : bois prélevé artificiellement sur une bête abattue ou morte naturellement. - pédicule : appendice frontal, non caduc, intermédiaire entre le bois et le crâne. - médaillon : surface de séparation à la base du bois de mue, présentant un aspect piqueté.

Les différents fragments présentés sont dessinés à l’echelle 1/1 par A. Thuet sauf mention contraire.

- meule : bourrelet circulaire à la base du merrain, formé de petites bosses plus ou moins saillantes ou soudées entre elles, les pierrures.

1.1. Eléments détachés du bois de cerf

- merrain : tige centrale du bois sur laquelle se greffent les futurs andouillers.

1.1.1. Origine des ramures

- andouiller : ramifications annuelles partant du merrain.

11 bases de bois ont été découvertes dans cette structure. Bois de mue et de massacre sont assez équitablement répartis. Seul un petit fragment de merrain lié à une meule très abîmée (n° 11) n’a pu être classé dans l’une de ces catégories.

- andouiller basilaire : première ramification à partir de la base du merrain. - andouiller de glace : ramification accessoire, parfois absente, en deuxième position sur le merrain.

- Bois de mue (n° 1 à 4) :

- andouiller central : troisième branche de la structure classique.

Ils sont au nombre de 4 dont deux droits et deux gauches. Mais ils n’appartiennent pas à la même ramure, le diamètre de la meule étant très différent :

- empaumure : ensemble des branches finales appelées épois. 26

A. Thuet: Un Atelier de Production de Peignes en Bois de Cerf de la fin de l’Antiquité tardive...

1. Bois de mue droit. D. 69 mm x 61 mm. Animal de grand âge. 2. Bois de mue droit. D. 72 mm x 59 mm. Animal de grand âge. 3. Bois de mue gauche. D. 45 mm x 35 à 40 mm. Animal plus jeune. 4. Bois de mue gauche. D. indéterminable mais animal de grand âge.

9. Bois de massacre gauche. D. 73 mm x 60 mm. Animal de grand âge. Fragment prélevé à la base du crâne. 10. Bois de massacre gauche. D. indéterminable. Prélèvement effectué au niveau du pédicule. L’usage de la scie est exclusivement employé pour découper ces fragments de ramure, que ce soit au niveau du merrain ou du pédicule. Ce dernier n’est jamais scié complètement. Souvent à mi-parcours, l’artisan exerce une pression pour arracher le pédicule à la meule. Le négatif d’une grosse esquille se forme alors sur la pièce. Le merrain est découpé de la même manière qu’un bois de mue, au niveau de son intersection avec l’andouiller basilaire (Figure 3).

Les meules sont brutes et n’ont pas été traitées. Elles sont attachées au merrain qui a été scié au niveau de l’intersection avec l’andouiller basilaire. Deux traces de scie sont visibles sur chaque exemplaire, la première pour scier en biais le merrain, la seconde pour scier en biais la base de l’andouiller basilaire (Figure 2).

Figure 3 Bois de massacre nº 6, avec les traces de découpe

Contrairement aux productions de l’atelier découvert à Drevant, dans le Cher, les pédicules n’ont pas été prélévés pour en faire des fusaïoles (Béal 1984). De même, les médaillons n’ont pas été utilisés. Leur usage était très répandu durant l’époque antique où ces éléments étaient considérés comme des amulettes porte-bonheur, symboles de force, de fécondité et de régénérescence. Pourtant, il semblerait qu’au cours de la période mérovingienne, ces usages ne soient pas tombés en désuétude... (Faider-Feytmans, Lebel 1956).

Figure 2 Différentes traces de découpes sur les meules

- Bois de massacre (n° 5 à 10) : Ils sont au nombre de 6 dont trois droits et trois gauches. Ils ne semblent pas non plus appartenir à la même ramure, bien qu’il soit délicat de se baser sur les seules observations du diamètre :

1.1.2. Quelques fragments complets et découpés (Figure 4) Seuls 6 fragments découverts dans ce lot correspondent à plusieurs parties facilement identifiables d’un bois de cerf (les autres pièces sont trop découpées pour nous permettre de les replacer sans erreur sur un bois). Ces éléments relativement complets, n’ont subi que peu de découpes et ont, soit été rejetés pour divers motifs encore difficiles à définir, soit laissés en attente pour être traités ultérieurement.

5. Bois de massacre droit. D. indéterminable. Animal jeune. Fragment prélevé à la base du crâne. 6. Bois de massacre droit. D. 60 mm x 47 mm. Animal d’âge moyen. Prélèvement effectué au niveau du pédicule (Figure 3). 7. Bois de massacre droit. D. 60 mm x environ 55 mm. Animal d’âge moyen. Prélèvement effectué au niveau du pédicule. 8. Bois de massacre gauche. D. 69 mm x 60 mm. Animal de grand âge. Prélèvement effectué au niveau du pédicule.

Les deux premiers objets sont très intéressants et correspondent à deux portions de merrain, n’ayant subi aucune découpe longitudinale en vue d’élaborer des 27

Materials of Manufacture

Le dernier élément, n°17, est très particulier et facile à replacer sur un bois de cerf. Il s’agit de l’intersection entre les épois, au niveau de l’empaumure. On observe très nettement trois départs d’andouillers très rapprochés, sans compter le merrain. Ce fragment était sans doute destiné à être exploité car l’épaisseur de bois est très développée à cet endroit (de 5 à 8 mm), qui présente une surface assez large et relativement plane. Cet usage systématique expliquerait l’absence d’autres fragments de ce type dans la série. 1.2.1. Les éléments relatifs aux intersections merrain-andouillers 53 fragments présentant des intersections ont été inventoriés (n° 18 à 70). Représentant les parties les plus massives découvertes dans ce lot, certains d’entre eux ont pû être repositionnés sur un bois de cerf. 29 sont de simples découpes, les 24 autres retracent les différents stades de la chaîne opératoire aboutissant au prélèvement d’un maximum de plaquettes. Les simples découpes se décomptent ainsi : -n° 18 à 23 : Intersection merrain-andouiller basilaire et andouiller de glace (Figure 5). Ces éléments sont facilement reconnaissables car ils présentent deux départs d’andouiller très rapprochés. 4 traces de scie sont visibles sur l’ensemble de ces fragments, deux pour le merrain et un pour chaque andouiller. A ce stade, le merrain est déjà fendu longitudinalement à l’aide d’un tranchoir.

Figure 4 Découpes 12 à 17 repositionnés sur un bois

plaquettes. Le premier, n°12, provient d’une base de merrain, découpé au niveau de l’intersection avec l’andouiller basilaire. Il atteint 35 à 45 mm de hauteur et aurait pu servir à la confection de petites plaquettes. Peut-être était-il en attente de traitement ?

-n° 24 à 26 : Intersection merrain-andouiller de glace. Ces trois pièces possèdent la particularité de ne pas être sciées au niveau de l’andouiller. Ce dernier a été brisé accidentellement durant la vie de l’animal (les cassures sont

Le second fragment, n°13, est de taille plus importante puisqu’il mesure 90 mm de hauteur. Scié aux deux extrémités, il possède un diamètre de 51 mm x 43 mm. Toute la partie spongieuse a disparu sans que l’on puisse affirmer qu’elle ait été volontairement prélevée. Par ailleurs, l’aspect externe du bois est peu marquée. Le fragment brisé en plusieurs morceaux a subi de nombreuses cassures fraîches. Il s’agit en fait des deux seuls éléments de merrain découverts dans cette série. On peut supposer qu’avant de mettre un terme à cette fabrication, l’artisan a probablement travaillé tout le stock de matière première non débitée en sa possession ; à moins que ces bois aient été entreposés dans un autre espace ou récupérés lors de l’arrêt de la production. Trois autres fragments assez facilement identifiables, n° 14,15 et 16 sont des intersections merrain-andouiller qui ont la particularité d’être complets, du moins au niveau du merrain. Celui-ci n’est pas fendu longitudinalement comme dans la plupart des éléments présentés par la suite. Mais seul un exemplaire présente un andouiller complet (n°16). Les deux autres, n° 14 et 15, sont sciés à quelques centimètres de la base de l’andouiller (entre 3 et 5 cm). Il s’agit probablement d’intersection de merrain au niveau de l’andouiller central. Le fragment n°16 est probablement situé au même endroit mais il peut s’agir aussi de la base du bois, au niveau de l’andouiller basilaire. Celui-ci possède la particularité de former un coude assez net, très marqué sur l’élément n°16.

Figure 5 Intersection merrain-andouiller nº 18 28

A. Thuet: Un Atelier de Production de Peignes en Bois de Cerf de la fin de l’Antiquité tardive...

érodées). Il s’agit sans doute de l’andouiller de glace car celuici étant «peu calcifié, il est souvent brisé lors des combats, très en avant dans le fût, parfois même à sa base» (Billamboz, 1979, p. 120). Le merrain, après avoir été scié, a été fendu longitudinalement. Les tronçons prélevés mesurent ainsi entre 30 et 56 mm. -n° 27 à 41 : Intersection merrain-andouiller (Figure 6). Ces 15 éléments ont tous subi la même découpe. Le merrain est scié au niveau des bases, ainsi que l’andouiller. Puis, le merrain est fendu longitudinalement. Des traces de tranchoir sont parfois visibles sur les bases des merrains. Les traces laissées par l’extrémité de la lame sont très fine (n°34-35), contrairement à celles produites par le passage de la scie, mesurant entre 2 et 2,5 mm d’épaisseur (35-36-40). A Drevant, les traces relevées varient de 1,5 à 3,5mm (Béal 1984, p. 9). Les tronçons de merrain ainsi obtenus sont de taille variable, entre 41 et 110 mm de hauteur. -n°42 à 44 : Intersection merrain-andouiller avec découpes différentes (Figure 7). Dans le premier cas, l’andouiller n’a pas été séparé du merrain comme à l’accoutumée. Le merrain est lui aussi intact. Seul l’andouiller a subi un prélèvement longitudinal au niveau de sa base inférieure. Ce geste a d’ailleurs entraîné sa cassure à mi-hauteur.

Figure 7 Découpes originales sur trois fragments

-n°45-46 : Intersections indéterminées. Ces deux éléments ont été fortement brûlés et sont entièrement noircis. Ils ont, semble t-il, éclaté au feu.

Le second élément est plus hypothétique : le positionnement du fragment et son orientation sur un bois de référence posent problème. Néanmoins, il semble que la partie supérieure de l’intersection ait été prélevée. Le dernier fragment est un rejet de découpe de la partie inférieure de l’andouiller associé à un fragment longitudinal de merrain. L’ensemble mesure environ 50 mm de haut.

1.2.2. Découpes sur intersections -n°47 à 50 : Intersection avec enlèvement d’une portion de surface (Figure 8). Une fois le merrain tranché en deux dans le sens des fibres, l’artisan s’applique à débiter les parties les plus planes dans lesquelles il lui sera possible de réaliser des plaquettes. Son premier geste consiste ici à trancher à partir de la base de l’andouiller la surface externe du bois en direction du merrain. -n°51 à 53 : Intersection avec enlèvement d’une surface. Seuls deux exemplaires présentant ce stade de la chaîne opératoire ont été découverts. Cette faible proportion pourrait indiquer que ce geste est assez facile à réaliser: ces deux éléments auraient été rejetés pour divers motifs. En effet, l’intégralité d’une paroi a été débitée, qui correspond, sur ces fragments, à des surfaces de 45 et 60 mm de haut x 30 et 20 mm de large. Un autre fragment probablement issu d’une empaumure, a aussi été débité sur une tranche, mesurant 50 mm de haut sur 30 à 55 mm de large. -n°54 à 63 : Débitage de forme grossièrement quadrangulaire (Figure 9). Après tranchage longitudinal du merrain, prélèvement d’une surface, l’élément issu d’une intersection est débité avec un tranchoir afin d’enlever une portion de spongiosa, la partie spongieuse située à l’intérieur du bois. Le résultat obtenu forme ainsi une sorte de cube. Les différentes hauteurs obtenues sont peu variables et s’échelonnent de 47 mm à 65 mm.

Figure 6 Découpes sur intersection merrain-andouiller 29

Materials of Manufacture

Figure 10 Intersection nº 70

échappent. Les différents fragments rejetés mesurent entre 46 mm et 105 mm de haut.

Figure 8 Intersection nº 48

-n° 69-70 : Intersections de forme triangulaire avec prélèvement d’une surface (Figure 10). Ces deux exemplaires illustrent une autre technique de débitage destinée à recueillir une petite plaque assez large. Dans un premier temps, le fragment est sectionné pour obtenir une forme triangulaire. Dans ce but, la base du merrain est sciée parallèlement à la zone courbe située entre l’andouiller et le merrain. Puis ce dernier est tranché longitudinalement le plus loin possible de l’intersection afin de former une surface relativement plane. Celle-ci est ensuite séparée du bloc par tranchage. Cet outil est ensuite utilisé pour supprimer la partie spongieuse. Au mieux, l’artisan aurait obtenu, après suppression des perlures et découpage fin, des plaquettes mesurant 45 mm de haut sur 25 mm de large. Ainsi, à partir d’une intersection merrain-andouiller, l’artisan parvenait à façonner quelques plaquettes, après tranchage longitudinal du merrain, séparation des surfaces du bois et retrait de la partie spongieuse. Il semble donc qu’aucune portion de matière première n’était délaissée. 1.3.1. Débitage sur Andouillers Autres éléments extraits d’un bois de cerf, les andouillers tiennent une place importante dans la chaîne opératoire de débitage. Quelles que soient les époques, ils sont souvent employés en tant que manches d’outils, amulettes-pendentifs ou autre objet dont l’usage nous échappe parfois. Au MoyenÂge, leur extrémité est utilisée pour la conception de petits pions de jeu, jetons, petis dés, perles...

Figure 9 Intersection nº 56

-n°64 à 68 : Amorce de plaquettes. Dernière étape avant l’obtention de plaquettes, il suffit de supprimer avec un ciseau les perlures et le reste de spongiosa. Mais aucun de ces cinq exemplaires n’est parvenu à ce stade. L’artisan a supprimé une grande partie de spongiosa (n°65) mais n’a pas poursuivi son travail pour des raisons qui nous

Observant les descriptions de Billamboz se rapportant à chaque type d’andouiller (Billamboz, 1979, p. 120), il nous 30

A. Thuet: Un Atelier de Production de Peignes en Bois de Cerf de la fin de l’Antiquité tardive...

a paru possible de définir pour certains éléments complets la position qu’ils occupaient dans la ramure. Aussi avons-nous répertoriés 10 andouillers basilaires (n°71 à 80), 1 andouiller de glace (n°81), 5 andouillers centraux (n°82 à 86) et 11 épois (n°87 à 97). Seuls 7 andouillers complets sont restés indéterminés (n° 98 à 104). Toutes les bases des andouillers et des épois sont sciées. La plupart possèdent une petite esquille. Quelques traces de râpe sont visibles sur certains exemplaires (n°72 et 81). Des traces de scie ont aussi été observées sur quelques andouillers, le plus fréquemment au niveau de la base. La largeur de la voie de la scie varie de 1,5 à 2 mm. La plus large observée au niveau d’un épois (n°95), a laissé une empreinte de 2,5 mm. Il faut noter la présence de deux petits épois sciés en plein développement (n°87 et 97), mesurant respectivement 28 mm et 30 mm. Le premier a été scié avec le bourgeon d’un second épois à peine amorcé (5 mm de haut). 13 extrémités d’andouillers ont été inventoriées (n° 105 à 117). Leur hauteur varie de 28 mm à 59 mm. 9 bases ont été sciées, 3 tranchées et 2 arrachées (peut-être est-ce une cassure involontaire?). 4 extrémités ont subi le feu. Des traces de scie sont visibles à la base de certains éléments et la largeur de la voie de la lame est comprise entre 1 mm et 1,5 mm.

Figure 11 Andouiller fendu longitudinalement

Ces éléments auraient pu être utilisés à la conception de petis pions, dés ou perles mais il ne semble pas que ce fut le cas dans cet atelier, contrairement à celui de Rheinau en Suisse où perles et amulettes sont façonnées dans les andouillers (Hedinger 2000). 1.3.2. Traitement de l’Andouiller De même que pour les intersections, les découpes sur andouiller permettent de prélever des lames de taille variable, dont la forme relativement courbe doit pourtant gêner l’artisan. De nombreux rejets de débitage ont été recensés dans cette série. Peut-être utilisait-on le feu pour redresser certaines plaques trop courbes avec plus ou moins de succès?

Figure 12 Tronçon d’andouiller nº 136

-n° 118 à 121 : Andouillers fendus longitudinalement (Figure 11). Très peu de pièces ont subi cette découpe. Il s’agit dans un cas d’un andouiller complet, scié à la base et fendu jusqu’à l’extrémité. Les autres fragments sont des extrémités d’andouillers sciés à la base, puis tranchés avec plus ou moins de réussite jusqu’au sommet. L’un de ces éléments a subi quelques aménagements de surface, au ciseau. Sa base possède d’ailleurs déjà la forme d’une plaquette.

Un seul fragment, n°136, mesurant 39 mm de haut, a subi le traitement suivant dans la chaîne opératoire : une partie de sa surface externe a été prélévée au ciseau afin de supprimer les traces des perlures. 5 facettes assez soignées sont visibles (Figure 12). 1.3.3. Tronçons d’Andouillers débités longitudinalement Cette série regroupe 34 andouillers répartis sur 3 ensembles homogènes créés en fonction de la taille des fragments conservés. L’artisan recherchait essentiellement deux longueurs de plaque, destinées au façonnage de traverses et de plaquettes d’assemblage de peignes. Ces dernières mesuraient en moyenne 45 mm de haut, tandis que les traverses mesuraient entre 80 et 120 mm de long. Ces différentes tailles apparaissent dans ces ensembles et permettent d’attribuer un usage à ces éléments.

-n° 122 à 138 : Tronçons d’andouillers Il s’agit d’une autre méthode de débitage, où l’artisan commence par découper le cylindre en fonction de la longueur de la plaque recherchée. Les éléments rejetés ou préparés mesurent entre 25 mm et 115 mm, la moyenne se situant vers 65 mm. Les deux bases sont sciées proprement et sont parallèles l’une à l’autre. Quelques petites esquilles sont visibles. Des traces de scie apparaissent sur certains fragments, leur largeur est de 2 mm. 31

Materials of Manufacture

1.4. Découpes diverses

-n° 139 à 144 : Petits gabarits Il s’agit de 6 éléments mesurant entre 36 mm et 55,5 mm. Ils ont tous été tranchés longitudinalement avec plus ou moins de réussite. L’un d’eux laisse à peine apparaître sa partie spongieuse. Deux autres possèdent une forme convexe. Ces fragments de petite taille sont des rejets de débitage inutilisables dans la conception d’un peigne.

Seules trois pièces issues de débitage n’entrent dans aucune des différentes catégories présentées ci-dessus (n°173 à 175). Deux fragments sont des rejets identiques de débitage de merrain. Il s’agit de petites rondelles issues, après découpage grossier du merrain, du redressement des deux bases du fût, afin qu’elles soient symétriques. Ces rondelles possèdent donc une base droite tandis que l’autre est fortement biseautée. Ainsi, le merrain, de forme cylindrique, aux bases parallèles permettait le façonnage de plaques de taille identique (n° 173-174).

-n° 145 à 154 : Gabarits de taille moyenne 10 éléments mesurant entre 56 mm et 88 mm ont été regroupés dans cette catégorie. Les deux bases étant sciées dans la plupart des cas, ils sont le résultat de débitage d’andouillers en tronçons. Après cette opération, l’artisan a tenté de les débiter longitudinalement. Ce travail semble délicat à accomplir puisque au moins 5 déchets sont inventoriés. Ils possèdent une forme triangulaire, avec une extrémité en pointe, qui empêche l’utilisation de ce fragment.

La dernière pièce, n°175, est de forme grossièrement quadrangulaire, obtenue à partir d’un élément indéterminé du bois de cerf. 3 des 6 faces de l’objet sont encore brutes (l’une d’elles semble polie). Deux autres sont sciées, la dernière a été obtenue par tranchage. On ignore son utilisation. Peut-être était-elle destinée au façonnage de dé ou d’une pièce de jeu?

Les 5 autres éléments sont peut-être de la matière première encore utilisable, à moins que leur section ne soit trop courbe pour en tirer une plaque suffisamment large. Néanmoins, l’un de ces éléments a subi un prélèvement par facettage de sa surface externe (Figure 13). Ces fragments étaient sans doute destinés à devenir des plaquettes d’assemblage.

2 Evolution vers l’Objet fini 2.1. Elaboration de plaques à partir des merrains

-n° 155 à 172 : Gabarits de grande taille 18 éléments mesurant entre 80 mm et 123 mm appartiennent à cette catégorie. Il est difficile d’affirmer, pour 6 d’entre eux, s’ils faisaient partie d’un tronçon d’andouiller car une seule base est visible. L’autre extrémité se temine en pointe. Cette forme est issue d’un tranchage malheureux.

Hormis deux fragments de merrain complets, aucune pièce cylindrique provenant de cette zone centrale du bois ne nous est parvenue. Seuls ont pu être inventoriés comme provenant des merrains, un nombre important de tiges plus ou moins larges et hautes, susceptibles d’être encore utilisées, associées à de nombreux rejets de façonnage. De la même manière que les andouillers, les merrains devaient être tronçonnés en fonction de la longueur de la plaque désirée. Deux grands ensembles ont été distingués, regroupant ébauches et rejets de petite taille, opposés aux pièces de grande taille.

Les autres fragments sont tous issus de tronçons d’andouillers. 9 d’entre eux ont été rejetés après avoir été tranchés longitudinalement. Certains possèdent une forme courbe ou très convexe. 4 fragments ont été séparés de leur partie spongieuse. Un seul élément a subi des prélèvement de sa surface par tranchage qui n’a pas donné les résultats espérés. Ces objets étaient destinés à l’élaboration de traverses de peigne.

2.1.1. Eléments de petite taille Il s’agit de 75 pièces correspondant à de nombreux rejets de façonnage de plaquettes. Deux ensembles ont été distingués. -n°176 à 204 : Il regroupe 29 éléments mesurant 40 à 60 mm de haut sur 20 à 30 mm de large qui correspondent probablement, à titre indicatif, au 1/4 d’un tronçon de merrain débité longitudinalement. Certain d’entre eux pourraient encore servir de matière première. D’autres possèdent une forme triangulaire typique des rejets. Sur quelques exemplaires, les perlures ont été supprimées en partie, de même que la spongiosa. -n°205 à 250 : Ce groupe représente 46 éléments de même hauteur que le lot précédent mais de largeur moindre, entre 5 et 18 mm. Ce sont tous des rejets de débitage à différents stades de la chaîne opératoire. Certains ne possèdent presque plus de spongiosa, d’autres encore. Aucun n’est parvenu au stade de la suppression des perlures. Quelques pièces appartenant à ces deux ensembles ont subi le feu.

Figure 13 Tronçon d’andouiller débité longitudinalement 32

A. Thuet: Un Atelier de Production de Peignes en Bois de Cerf de la fin de l’Antiquité tardive...

2.1.2. Eléments de grande taille

2.2.1. Elaboration d’une face (Figures 14 et 15). Dessins G. Léon

Ils représentent 51 pièces sans doute destinées à l’élaboration de traverses de peignes. Elles sont réparties en deux groupes en fonction de la largeur du fragment.

-n°353 à 398 : Ces 46 déchets ont tous été travaillés sur une seule face. Leur forme est très diversifiée. Hormis 7 exemplaires de grande taille, la hauteur varie de 20 à 77 mm, la largeur de 8 à 26 mm et l’épaisseur de 2 à 10 mm. Seules quelques plaques ont été travaillées sur les deux faces. Mais il demeure souvent un peu de spongiosa, difficile à supprimer entièrement. Les pièces de grande taille mesurent de 86 à

-n°251 à 266 : Ces 16 éléments mesurent de 74 à 184 mm de haut. Cette dernière mensuration est exceptionnelle, la hauteur moyenne se situe vers 100 mm. Néanmoins, 5 éléments mesurent entre 120 et 130 mm de long. Ces fragments correspondent au 1/4 d’un tronçon de merrain, entre 20 et 30 mm de large. Certains possèdent une forme triangulaire. D’autres sont courbes. La plupart n’ont plus de spongiosa. -n° 267 à 301 : Les mêmes remarques sont valables pour cet ensemble de 35 éléments, dont la hauteur varie de 64 à 116 mm, pour une largeur moindre, de 7 à 20 mm. Ce sont tous des rebuts de débitage de plaques de traverses de peignes. Certains possèdent des traces de prélèvement de surface. Ces 126 pièces sont les seuls vestiges des portions de merrains utilisées pour le façonnage de plaques de peignes. D’après la matière première supposée (11 bois de cerf au moins, issus d’animaux âgés), ces rebuts paraissent tout à fait anecdotiques. On peut ainsi mesurer la dextérité avec laquelle l’artisan travaillait la matière, dans le cas où tous les rejets de ce travail seraient présents dans cette fosse.

Figure 14 Plaque façonnée sur une face

Avant de décrire les ultimes gestes aboutissant à la création de plaques opérationnelles, il nous reste à traiter les derniers rebuts laissés par ce travail, répartis en deux catégories. 2.1.3. Fragments de Spongiosa -n°302 à 311 : Il s’agit de dix éléments de taille variable, souvent cubiques ou allongés, séparés de la plaque à peine ébauchée au moment de l’enlèvement de la surface externe du bois. Ces trop rares fragments semblent confirmer l’hypothèse que cette collection provient d’une fosse de dépôt secondaire de débitage du bois de cerf. En effet, la découverte de ces restes spongieux aurait dû être plus importante, au vu de la quantité de plaquettes produites dans cette série. 2.1.4. Petits déchets de toute nature -n°312 à 352 : Quelques petits rebuts de débitage ont été ramassés lors de la fouille de cette structure, correspondant à 41 fragments. Il s’agit d’éléments de très petite taille, informes, issus d’éclats de débitage ou de rejets de tranchage de surface. 15 d’entre eux ont éclaté au feu. De même que pour les fragments de spongiosa, il semble que ces rejets soient peu nombreux par rapport à la quantité de plaques débitées. La même conclusion s’impose. Par ailleurs, il se peut que ce travail ne produise pas ou très peu de rejets. Seule l’archéologie expérimentale pourrait nous fournir une réponse satisfaisante. 2.2. Façonnage des Plaques L’ultime travail de débitage consiste à supprimer les différentes aspérités de la surface du bois ainsi que la partie spongieuse.

Figure 15 Plaque façonnée sur une face 33

Materials of Manufacture

137 mm sur 14 à 22 mm de large et 5 à 8 mm d’épaisseur. L’un d’eux a été travaillé sur les deux faces mais il a été rejeté, à cause de la spongiosa trop développée. 2.2.2. Deux faces polies (Figures 16 et 17). Dessins G. Léon -n°399 à 445 : Ces 47 pièces ont été travaillées sur les deux faces mais ont été rejetées pour diverses raisons. La plupart ne sont pas assez larges pour servir comme plaquettes. D’autres possèdent une forme très convexe. Certaines sont des rejets de plaquettes achevées mais de trop grande longueur. Quelques traces de spongiosa sont encore visibles par endroits. Trois rebuts de plaques ont subi le feu, de façon accidentelle ou pour assouplir la matière et faciliter le redressement de ces pièces trop courbes. Ces éléments représentent deux gabarits différents déjà observés précédemment. Les plus nombreux mesurent de 40 à 60 mm de long, le plus long 129,5 mm. Il est normal de retrouver plus de petites plaquettes que de grandes puisque l’artisan devait façonner au moins 6 à 8 plaquettes de peigne pour seulement deux traverses.

Figure 17 Plaque dont les deux faces sont travaillées

2.3. Les Objets semi-finis rejetés Une fois les plaquettes correctement formées, l’artisan les assemble entre elles et les maintient à l’aide des plaques de traverse rivetées. Puis il dresse les dents de taille différente, les plus fines pour supprimer les poux, les plus épaisses pour peigner la chevelure. Souvent, les extrémités des peignes sont décorées de motifs géométriques simples. Toutes ces étapes nécessitent un savoir-faire fondamental. Le moindre geste qui dérape, détruit, sans doute, plusieurs heures de travail. Les objets découverts dans cette fosse sont des exemples de ratés, représentant diiférents moments de l’assemblage d’un peigne. Les premières pièces sont liées aux extrémités du peigne, les autres aux traverses de l’objet.

Deux phases différentes de la chaîne opératoire sont visibles dans ce lot. La première correspond au débitage de la face interne de la pièce. La seconde est le résultat de l’assemblage des plaquettes et de leur égalisation en hauteur. C’est pourquoi de très petits déchets mesurant 20 mm de haut ont été répertoriés.

2.3.1. Description des Objets (Figure 18). Dessins G. Léon -n°446 : Plaque de forme trapézoïdale, courbe, légèrement convexe. L. base inf. 68 mm ; L. base sup. 60,5 mm ; Hteur. de 12 à 17 mm ; Ep. 3,5 mm. La base supérieure, brisée, a subi deux petites perforations juxtaposées qui sont peut-être à l’origine de la cassure (d’où rejet). De forme circulaire, elles mesurent 2 mm de diamètre et ont été forées. Cette pièce était-elle destinée à servir de plaque d’extrémité décorée de peigne? Ces dernières étaient souvent perforées de plusieurs motifs circulaires destinés à orner l’espace laissé libre entre l’extrémité du peigne et le bord de la traverse servant à maintenir les plaquettes entre elles. (Figure 18A) -n°447 : Plaque d’extrémité de peigne (?) à double denture. L. 38 mm ; l. 12 mm ; Ep. 3 mm. La pièce est brisée dans sa largeur. Des dents ont été façonnées de chaque côté de l’objet mais pas sur toute la surface, comme pour les plaques d’extrémité. De plus, elle est perforée en son centre (D. 3 mm), pour y appliquer sans doute un rivet, mais celui-ci n’a laissé aucune trace. La plaque n’a probablement jamais été assemblée. Elle a été rejetée pour une raison qui nous échappe (Figure 18B).

Figure 16 Plaque dont les deux faces sont travaillées 34

A. Thuet: Un Atelier de Production de Peignes en Bois de Cerf de la fin de l’Antiquité tardive...

B A

C

E

D

F

G

H

Figure 18 Les Objets semi-finis et complet : A Plaque de forme trapézoidale nº 446 ; B Plaque d’extrémité de peigne à double denture nº 447 ; C Plaquette d’extrémité de peigne à double denture nº 448 ; D Plaquette décorée d’extrémité de peigne à double denture nº 449 ; E Ebauche d’épingle à cheveux en os (?) nº 453 ; F Plaque de traverse brisée nº 451 ; G Deux traverses brisées au niveau des rivets, enserrant une plaquette de peigne à double denture nº 452 ; H Plaque formant traverse nº 450 35

Materials of Manufacture

-n°448 : Plaquette d’extrémité de peigne à double denture, non décorée, associée à deux fragments de traverses. L. 35 mm ; l. 20 mm ; Ep. 3 mm ; L. traverse 12 mm. l. traverse 9 mm ; Ep. traverse 2,5 mm. Avant d’être rejeté, le peigne a été scié au niveau des deux traverses comme pour séparer cette partie du reste du peigne. Quelques dents sont brisées. Or, on sait que le façonnage des dents s’opérait après assemblage des plaquettes entre elles. Le fait de casser des dents aurait pû justifier le rejet de cette pièce (Figure 18C).

-n°453 : Ebauche d’épingle à cheveux en os (?), de type Béal AXX,7 ou 8. L. act. 70 mm ; H. tête 10,5 mm ; D. tête : 8,5 mm ; D. col 3 mm ; D. max 5,5 mm.

-n°449 : Plaquette décorée d’extrémité de peigne à double denture, associée à un fragment de traverse. L. 50 mm ; l. 21 mm ; Ep. 4 mm. La plaquette est complète. L’extrémité du peigne est ornée de motifs géométriques simples, symétriques par rapport à l’axe longitudinal de l’objet. Ce décor est formé de deux échancrures dont une en V dissymétrique formant moulure. Au centre de celle-ci, un petit trou a été foré (D. 2 mm) pour rehausser le décor. Des dents de taille différente ont été façonnées de chaque côté de l’objet. Certaines sont brisées. Le rivet en fer est encore en place et maintient un fragment de traverse. Celle-ci est soulignée de trois fines rainures longitudinales. L’objet a peut-être été rejeté à cause des dents brisées (Figure 18D).

Il serait utile d’effectuer des observations de cette épingle à la loupe binoculaire pour identifier clairement le matériau dont elle est issue. Il est vrai que cela semblerait plus logique qu’elle ait été façonnée en bois de cerf puisqu’aucune découpe n’a été observée sur les restes animaux découverts dans la fosse. Le résultat de ce travail artisanal étant peu probant, la production de cet objet aurait été peu développée.

La tête est grossièrement façonnée. Elle possède une forme bulbique au sommet pointu. De nombreuses facettes la marquent. Une partie de l’os spongieux est encore visible à cet endroit. La tête est bien séparée du corps. Celui-ci est fortement renflé à 28 mm sous le col. Une partie du corps et la pointe manquent (Figure 18E).

3 Synthèse des Données 3.1. Décompte des fragments de bois de cerf 11 meules 2 tronçons de merrain 1 base d’empaumure 3 intersections complètes merrain-andouiller 29 intersections merrain-andouiller 24 découpes sur intersections 34 andouillers 13 extrémités d’andouiller 55 découpes sur andouillers 3 découpes originales 126 découpes sur merrains 10 fragments de spongiosa 41 déchets divers 46 plaques travaillées sur une face 47 plaques travaillées sur les deux faces 8 objets semi-finis

-n°450 : Plaque formant traverse, brisée au niveau d’un coup de scie à 20 mm de l’extrémité. L. 88,5 mm ; l. 12,5 mm ; Ep. 4 mm. La pièce est de forme courbe. Elle a subi cinq perforations assymétriques (D. 2 mm). Les deux premiers trous sont espacés de 52,5 mm et auraient pû servir au rivetage des plaques. C’est pour séparer cette zone de son extrémité où 3 trous ont été perforés que cette plaque a été sciée mais elle s’est brisée. Cette pièce n’est pas marquée par les stries laissées par le façonnage des dents du peigne (Figure 18H). -n°451 : Plaque de traverse brisée. L. 41 mm ; l. 11 mm ; Ep. 3 mm. La plaque, ornée de deux fines rainures longitudinales, possède encore un rivet en fer qui l’a fendillé . On ignore de quelle manière elle a été brisée (Figure 18F).

Total : 453 pièces. -n°452 : Deux traverses brisées au niveau des rivets, enserrant une plaquette de peigne à double denture. l. plaquette 19 mm ; L. max. traverse 41 mm ; Ep. peigne 10 mm. Toutes les dents sont brisées. Les traverses sont d’ailleurs marquées des stries liées au façonnage des dents. Elles sont ornées de 4 rainures longitudinales. Trois rivets en fer sont encore visibles. Ils sont séparés de 16 à 17 mm les uns des autres (Figure 18G).

3.2. Quantité de matière première utilisée Cette estimation peut être établie à partir des bases de bois et des intersections merrain-andouiller. Au nombre de 11, les premières nous apportent une estimation minimale renforcée par le nombre d’intersections merrain-andouiller (56). Sachant qu’il existe au plus trois intersections par bois, lorsque celui-ci atteint un certain âge, au moins 18 bois ont été utilisés durant l’existence de cet atelier.

2.3.2. Une ébauche de tabletterie en os? Il s’agirait du seul élément en os découvert dans cette fosse. Correspondant à une ébauche d’épingle à cheveux, cet objet nous inviterait à supposer l’existence d’un débitage parallèle à celui que nous avons étudié, portant sur des métapodes de bovidés. Or, aucun fragment de ce type n’a été découvert dans cette structure.

3.3. Approvisionnement D’après le diamètre des 11 meules découvertes, qui permet d’appréhender l’âge de l’animal, ces bois proviennent de 36

A. Thuet: Un Atelier de Production de Peignes en Bois de Cerf de la fin de l’Antiquité tardive...

bêtes assez âgées (4-6 ans) voire très âgés (6-8 ans et plus). Cela implique que ces animaux ont eu la possibilité d’atteindre un âge respectable malgrè les prélèvements effectués par les chasseurs (présence de bois de massacre). A cette époque, le cheptel était-il important? L’environnement était-il propice à son développement? L’approvisionnement dépendait néanmoins fortement des hasards de la chasse et des découvertes ponctuelles de bois tombés naturellement. Par ailleurs, on ignore tout des relations qu’entretenait l’artisan avec ses fournisseurs. De quelle manière obtenait-il ces précieux bois? Devait-il les acheter? Possédait-il des fournisseurs attitrés? Devait-il rechercher lui-même les bois tombés naturellement? En combien de temps parvenait-il à réunir plus d’une dizaine de bois de cerf?

Les plaques et plaquettes ainsi obtenues sont assemblées et maintenues avec des rivets pour former un peigne. Les plaquettes sont égalisées pour être toutes de la même hauteur (5). Ensuite, le tabletier dresse à la scie la double denture et décore les extrémités du peigne. Celui-ci est enfin prêt à la vente. 3.4.2. Comparaison avec quelques Publications Michel Petitjean n’apporte pas de précisions complémentaires à l’observation de ce procédé de fabrication de peigne (Petitjean 1995). André Gob n’a pu observer sur le site de Huy (Gob, 1986) l’ensemble de la chaîne opératoire de débitage. Mais il apporte néanmoins quelques précisions concernant les techniques de débitage et de façonnage par sciage, refendage et raclage et polissage (Gob 1986, p. 195). Ces différentes phases ont été clairement distinguées sur le site normand, hormis peut-être l’étape du polissage difficile à caractériser en elle-même, ne porduisant que des déchets volatiles.

A priori, on a l’impression que cet artisanat, très ponctuel, ne s’effectuait qu’après regroupement de la matière première suffisamment abondante pour façonner une certaine quantité d’objets susceptible de répondre à un besoin précis. Un tabletier itinérant aurait pu travailler de cette manière.

Contrairement à Saint-Clair-sur-Epte les intersections merrain-andouiller sont ici rattachées avec le merrain dans le schéma de débitage d’un bois. Par ailleurs, ces derniers sont découpés en deux grands tronçons séparés par un rejet de forme triangulaire qui n’a pas été mis en évidence à SaintClair-sur-Epte. Ces petites variantes sont peut-être liées à une gestuelle propre à chaque tabletier, basée sur l’expérience et la dextérité de chacun, à moins que ces différences soient le fait d’une évolution chronologique de la technique de débitage. Seules d’autres analyses de site pourront étayer ces hypothèses.

Mais si le fait de réunir autant de bois de cerf peut paraître difficile, pourquoi ce matériau n’est-il pas utilisé au maximum des possibilités qui s’offrent au tabletier (utilisation des médaillons et des extrémités d’andouillers)?

3.4. La chaîne opératoire, conforme aux observations déjà établies? L’ensemble des différentes phases de débitage et de façonnage menant à la fabrication d’un peigne à double denture a pû être observée à travers ces fragments.

3.5. Diffusion 3.4.1. Résumé (Figure 19) Un minimum de 18 bois a été débité sur ce site. On peut estimer qu’au moins quatre peignes peuvent être réalisé à partir d’un bois, proposant ainsi une proudction d’un nombre minimal de 72 peignes à double denture. Les travaux d’archéologie expérimentale menés en Grande-Bretagne ont montré qu’une journée était nécessaire à la réalisation d’un peigne (Crummy 2000, p. 102). L’atelier de Saint-Clair-surEpte aurait donc pu durer plusieurs mois, voire une ou deux saisons. Une fois terminés, les peignes ont été vednus aux anvirons, peut-être jusqu’à Paris même. En effet, des objets de ce type y ont été découverts (Béal, Dureuil 1996, p. 55), sans qu’aucun véritable atelier de production des peignes sur bois de cerf n’ait été découvert à l’heure actuelle. De même à Rouen, des peignes similaires ont étés mis au jour.

Après éventuelle séparation du bois sur le crâne de l’animal, le merrain est débité à la scie en tronçons de différente taille. Les andouillers sont séparés des intersections (1). Les merrains sont ensuite découpés par tronçons en fonction de la taille de la plaque désirée puis débités au tranchoir dans le sens longitudinal pour former des ébauches de plaques. Guère plus de 4 plaques peuvent être façonnées dans un merrain. Une fois rejeté les ébauches mal débitées, les lames sont séparées de la partie spongieuse, puis de la surface du bois. Elles sont ensuite polies, éventuellement redressées au feu si elles sont trop courbes. Elles sont prêtes à être assemblées (2). Les andouillers sont séparés du merrain. Certains sont sciés à leur extrémité. D’autres sont découpés en tronçons. D’autres, enfin, sont débités longitudinalement au tranchoir pour en tirer de grandes plaques. Puis l’artisan supprime la partie spongieuse et les perlures à la surface de la plaque à l’aide d’un racloir (3).

En Normandie, des peignes tardifs et mérovingiens ont été découverts en contexte de nécropole mais la plupart d’entre eux semblent être conçus, d’apres les auteurs, à partir de l’os. Il faudrait réexaminer chaque objet afin de vérifier l’origine véritable du matériau car il est souvent difficile de distinguer à l’œil nu avec certitude und objet conçu dans un os ou un bois de cerf. Seule une étude au microscope nous en certifie la provenance.

Enfin, les intersections merrain-andouiller sont débitées longitudinalement au tranchoir puis une surface est prélevée. La spongiosa est supprimée ainsi que les perlures (4). 37

Materials of Manufacture

Figure 19 Chaîne opératoire de débitage sur bois de cerf

CONCLUSION

réserves de matière première et de petits fragments issus d’un débitage intensif, comme la spongiosa ou des esquilles de bois. Nous serions donc en présence d’un dépôt secondaire issu d’un atelier de production. Il n’est pourtant pas exclu que la structure 50 ait été le local abritant le tabletier mais les niveaux d’occupation ont été nettoyés.

Malgré la présence de fragments de bois ayant subi un début de façonnage, de résidus de travail et d’objets finis, l’étude de la chaîne opératoire a mis en évidence l’absence de 38

A. Thuet: Un Atelier de Production de Peignes en Bois de Cerf de la fin de l’Antiquité tardive...

D’après N. Crummy, les peignes tardifs à double denture en bois de cerf ont été réalisés par des tabletiers itinérants (Crummy 2000, p. 107). L’atelier de Saint-Clair-sur-Epte pourrait effectivement s’inscrire dans ce.cadre. L’absence d’autres structures de cette époque sur ce site semble indiquer que l’artisan venu d’ailleurs s’est installé de façon très provisoire dans un endroit de passage (encore inhabité donc sans clientèle) pour y récolter la matière première et façonner ses produits puis a emprunté ces mêmes voies de communication pour écouler sa production.

de l’os préhistorique», sous la direction de H. Camps-Fabrer, Paris, 1979, p. 93-143 CRUMMY N., 2000: Bone-working in Roman Britain : a model for itinérant craftsmen ? in Polfer M. : L’artisanat romain : évolutions, continuités et ruptures (Italie et provinces occidentales). Actes du., 2è colloque d’Erpeldange. 26-28 octobre 2001. Monographie Instrumentum 20. Montagnac 2001 DESALBRES R., 1996 : « Talismans en bois de cerf : permanence à travers les âges ? » Revue Archéologique de Bordeaux. Tome LXXXVII, 1996 DIJKMAN W. ERVYNCK A., 1998 : Antler, bone, hom, ivory and teeth. The use of animal skeletal materials in roman and early medieval Maastricht. Archaeologica Mosana 1. Maastricht, 1998

L’absence de prélèvement des médaillons sur le site de Saint-Clair-sur-Epte confirme l’hypothèse présentée par R. Desalbres qui constatait un déséquilibre géographique entre la France de l’est où de nombreux médaillons sont signalés et la France de l’ouest où ces découvertes sont rares (Desalbres 1996). Peut-on imaginer que le tabletier de Saint-Clair-surEpte n’avait pas connaissance de ces pratiques, à moins que ses proches contemporains n’aient pas attribué à cet objet une telle valeur symbolique.

FAIDER-FEYTMANS G. and LEBEL P., 1956 : Talismans en bois de cerf trouvés dans les tombes mérovingiennes. Revue Archéologique de l’Est et du Centre-Est, Tome 7, p. 138 à 143. 1956 GOB A., 1986: Huy, rue des Augustins: occupation mérovingienne et vestiges du rempart du XIè s. Activités du SOS Fouilles. 4. 1986 HABERMEHL K.H., 1985 : Altersbestimmung bei Wild und Pelztieren. 1985. Hamburg et Berlin.

Trop rares sont les publications portant sur ce genre de découvertes. Pourtant elles permettent de mieux comprendre le travail de l’artisan tabletier, acteur social que l’on ne peut aborder qu’à partir de vestiges comme ceux présentés ici. Ces études posent plus de questions qu’elles n’en résolvent mais associées les unes aux autres, elles permettront de cerner plus précisément le rôle du tabletier au sein du monde artisanal et les différentes relations qu’il entretient avec la société.

HEDINGER B., 2000 : Geweihbearbeitung in spätrömischen Wachtturm von RheinauKöpferplatz. Archéologie Suisse. Tome 23. p. 104 -114. LANGLOIS J-Y., 1998: Guichainville. Eure (27). R.N. 154. « La petite dîme ». Sous la direction de J-Y. Langlois, Y-M. Adrian et L. Paez-Rezende. Rapport de fouilles. AFAN. SRA de Haute Normandie, 1998. MACGREGOR A.1985: Bone antler, ivory and horn. The technology of skeletal materials since the Roman period, Beckenham, Sydney, 1985

Bibliographie

PETITJEAN M., 1995 : Les peignes en os à l’époque mérovingienne. Evolution depuis l’Antiquité tardive. Antiquités Nationales, Torne 27, 1995

BEAL J.-CI., 1984 : Un artisanat tardif du bois de cerf à Drevant (Cher). Bulletin des Amis du Musée Saint-Vic. Saint-Amand Montrord nº 12. Décembre 1984

POULAIN-JOSIEN Th., 1976 ~ L’étude des ossements animaux et son apport à l’archéologie, Centre de Recherches sur les Techniques Gréco-Romaines, nº6, Université de Dijon, 1976

BEAL J.-C., DUREUIL J.-F., 1996: La tabletterie gallo-romaine et médiévale, une histoire d’os. Catalogue d’art et d’histoire du Musée Carnavalet, tome XI. Paris, 1996

PROST D., 1984 : Le mobilier en os gallo-romain d’EscolivesSainte-Camille; RAE, Tome 35, Fasc. 3-4, juillet-décembre 1984, p. 263 à 299.

BILLAMBOZ A., 1979 : «Les vestiges en bois de cervidés dans les gisements de l’époque holocène. Essai d’identification de la ramure et de ses différentes composantes pour l’étude technologique et l’interprétation palethnologique» dans «Industrie en os et bois de cervidés durant le Néolithique et l’âge des Métaux, première réunion du groupe de travail numéro 3 sur l’industrie

La Picardie, berceau de la France. Clovis et les derniers Romains. 1500ème anniversaire de la bataille de Soissons. 486-1986. Catalogue d’exposition. Amiens, 1986 Rouen gallo-romain, fouilles et recherches archéologiques. Catalogue collectif d’exposition. Rouen, 1982

39

Materials of Manufacture

40

I. Riddler: A Lesser Material

A LESSER MATERIAL The Working of Roe Deer Antler in England during the Anglo-Saxon Period Ian RIDDLER

INTRODUCTION

THE DEER AND THE ANTLER

Both bone and antler were worked in considerable quantities in northern Europe during the first millennium AD. Selected bones of cattle, horse, sheep and pigs formed a readily available resource suitable for object manufacture. The skeletal elements of other species were also used, if not to the same extent. Bird limbs were shaped to form whistles, flutes and needle cases, and whale bone was used on opportunist occasions for a wide variety of implements (MacGregor 1985, 44-51; Gardiner 1997; Riddler 1998).

Roe deer antler is small in comparison with the antlers of the major species of northern Europe. The constituent elements of the antler are essentially the same, each antler lying on a pedicle extending from the frontal part of the skull. The burr is accompanied by a coronet and the beam shows pronounced and obvious guttering. Unlike other antler, the outer layer is relatively thick and cortile tissue forms only a small part of the beam (Ulbricht 1978, 22 and taf 52.4-7). The beam leads to a maximum of six tines, growth of the antlers being most pronounced between the fourth and sixth years of the life of the animal. Roe deer have a life expectancy of around ten to twelve years (Putman 1988, 23 and 106; Becker 1989, 137).

The preferred skeletal material from the fourth to the twelfth century was undoubtedly red deer antler. Other species of deer were not as widely distributed across Europe and the antlers of some were less suitable as the raw materials for working. Reindeer antler, for example, was less useful for this purpose than red deer antler and, as the most northerly of the cervids, its distribution was effectively confined to Scandinavia (Ulbricht 1978, 20-1; MacGregor 1985, 16, 35 and fig 27e). Elk antler, in contrast, provided wide, palmate surfaces, allowing larger and more complex objects to be produced (see Gostenčnik, this volume). Its distribution was restricted, however, to Scandinavia and central Europe (Ulbricht 1978, 21-2; Rülcker 1986).

The small size of the antler and the pronounced guttering of the beam are distinctive features that allow it to be readily identified. Objects that are little modified from the antler can easily be recognised. Roe deer antler was generally too small to be used for composite objects and for the period under study it is likely that its use was heavily restricted, and that most of the objects for which it was selected as the raw material can be identified. The range is tiny, in comparison with that for red deer antler, but judicious use of the material can be seen on a number of occasions.

Roe deer were widely distributed across northern Europe from early prehistory onwards, sharing that distinction with red deer. Nowadays they are found in almost every country of western Europe. They are much smaller than these other cervids, however, and their antlers extend to just 1.5 – 3% of the overall weight of the animal, substantially less than the percentage for red deer (Ulbricht 1978, 22). Unlike red deer, their antlers develop over the winter and are shed at its onset, in October or November.

Any broad study of this type is unfortunately hampered by the lack of detailed information provided in some publications. As noted by Becker, in the absence of a basic description of the material, comparisons between sites can be very difficult, if not impossible (Becker 1993, 135). Postcranial bones of roe deer have been found in small numbers in a large number of Roman and post-Roman settlements. In many cases, however, their occurrence is briefly noted but is not discussed, and the relative quantities of post-cranial bone and antler are not specified. Bourdillon (this volume) has shown also how a consideration of the range of skeletal elements of deer can highlight their relationship with antler working. Without the appropriate level of detail, this cannot easily be done. In the absence of any descriptions of antlers, for example, it is not possible to ascertain whether they had been worked, and the extent to which they had been modified. Comparison with other worked material from the same site can also place roe deer antler within a broader context of animal utilisation, but this has not always been achieved. All of this is balanced to some extent at least by the fact that

The small size of these antlers effectively meant that they were of comparatively little use in object manufacture. Yet roe deer antler appears in small but consistent quantities in settlements in England and on the Continent from the Roman period to the twelfth century. A brief review of post-Roman examples from Anglo-Saxon contexts, seen alongside those from elsewhere in northern Europe and contrasted with the evidence for the working of red deer antler, may shed some light on the reasons behind this occasional but persistent preference. 41

Materials of Manufacture

most of the post-Roman roe deer antler from England described here has been examined by the author. Details of post-cranial roe deer have also been obtained, wherever possible.

antler, which is conspicuously low in number. The minimum number of slaughtered deer may, however, have been fairly low. Interestingly, the same can be said for red deer at this site, for which just a few fragments of antler, as well as four calvaria (with most of the antler removed) occurred, in contrast to a reasonable quantity of post-cranial bone. For red deer, Harman noted that ‘much of the antler used came from carcasses’, a clear contrast with the situation at Burgh Castle, where most of the antler was naturally-shed (Harman 1993, 231 and table 40).

Ulbricht and Becker have noted that the study of worked antler and bone is essentially a collaborative effort between the archaeozoologist and the object specialist (Ulbricht 1980; Becker 1993). A comparison of antlers against post-cranial bone can indicate the value of deer within a provisioning regime. For post-Roman red deer, antler customarily exceeds post-cranial bone in number and much of the former is naturally-shed, suggesting that it was obtained from the countryside specifically to be worked. Can the same be said of roe deer ?

Within these Roman contexts, the quantity of roe deer antler is not significantly greater than the number of post-cranial bones of the animal. Whilst red deer may have been valued as much for their antlers as their meat, the antlers of roe deer were not viewed as a viable resource for craftsmen. Roe deer antler was used for object manufacture during the Roman period, albeit on a small scale. At Colchester, a naturally-shed roe deer antler came from a first century context. It had been laterally perforated just above the burr, but was not otherwise worked (Crummy 1983, 173 and fig 211.4757). Its function is unclear, although it may possibly have served as an amulet. As a naturally-shed antler, it is unlikely to have overtly symbolised the hunting of deer. Roman historical sources rarely mention roe deer, although they often accompany Diana in art of the period, as also do fallow deer (Schmid 1965, 60; Deschler-Erb 1998, 87). Both Roman and post-Roman writers viewed deer antlers as symbols of immortality, because of their renewal each year. They were also indicators of power and fertility, each new set of antlers exceeding its predecessor in size, until the animal reached old age (Meaney 1981, 188).

THE LATE ROMAN BACKGROUND King has noted that cervid bones are more commonly seen in late Roman assemblages in Britain than in those of the early Roman period. The general tendency at that time was for red deer to exceed roe in number within faunal assemblages. His interpretation of this situation lay with the assumption that ‘the size and meat weight of red deer, and the strength of its antlers, made it the first choice of the hunt’ (King 1978, 216). Davies has indicated that both red and roe deer formed a part of the Romano-British military diet (Davies 1971, 128 and table 1). Several sites from East Anglia provide a useful background for the utilisation of roe deer during the late Roman period. Grant refers to eight fragments of roe deer antler from late Roman contexts at Burgh Castle in Norfolk. They were thought to be of little significance in relation to the six postcranial bones of the same animal, which indicated that roe deer were hunted nearby. All six bones may have come from the same animal (Grant 1983, 109 and table 2). Red deer antler was also worked at the shore fort, probably in the late Roman period. The red deer antler includes 30 burrs, eighteen of which are naturally-shed (Grant 1983, 110; Riddler, Trzaska-Nartowski and Hatton, forthcoming). A contrast was drawn between the red deer antler, which was not accompanied by many post-cranial bones, and roe deer, for which the antler only exceeded the post-cranial bone in number by a small amount (Grant 1983, 110).

A roe deer antler from Stone Castle Quarry, Greenhithe provides a different perspective on the use of the raw material. The antler is naturally-shed and the beam has been laterally perforated above the burr. The tines have been removed, leaving an object described as a ‘bodkin’ (Detsicas 1966, 189 and fig 21.2). Although of Roman date, this object anticipates one of the principal uses of roe deer antlers in the post-Roman period. At Augst the fragments of roe deer antler included a skullattached antler with knife marks, as well as a small, indistinct fragment (Deschler-Erb 1998, catalogue nºs 5530 and 5563). Roe deer had been used for several objects, amongst which was another amulet, in this case produced from the distal end of a roe deer metacarpus (Deschler-Erb 1998, 171 and taf 40.3987).

For Caister-on-Sea, to the south of Burgh Castle, Harman noted that roe deer had been found in unusually high numbers within the Roman contexts (Harman 1993, 225). 52 fragments came from Roman contexts, five from post-Roman contexts, and 21 from the plough-soil (Harman 1993, tab 40). Most of the bones were post-cranial although a skull-attached antler was recovered, as well as one example that had been naturally-shed (Harman 1993, 231-2). Here again, the presence of roe deer appeared to relate to the hunting of wild animals in the locality but the animal was thought to be of little significance as a food source, despite the relatively high number of bones recovered. The quantity of post-cranial roe deer bone at Caister-on-Sea easily exceeds the amount of

EARLY AND MIDDLE SAXON ROE DEER ANTLER Roe deer antler has been recorded in small numbers from several Anglo-Saxon settlement sites. One of the earliest postRoman examples is likely to be an antler from St. Mary Cray in Kent (Hart 1984, 214). This came from the excavation of a sunken-featured building thought to be of fifth century date 42

I. Riddler: A Lesser Material

(Hart 1984, 194). A naturally-shed antler was accompanied by a tibia from a mature roe deer. The antler is briefly mentioned, with no indication of whether it had been worked.

from roe deer antler, are known from Frisia and from Hamwic, Middle Saxon Southampton (Roes 1963, pl XXXIX.9; Riddler 1986, fig 1; 1993, 119). 33 fragments of roe deer were identified altogether from Anglo-Saxon features at West Stow (Crabtree 1991, table 1). It was emphasised, however, that hunting played a very small role in the overall subsistence strategy for the Anglo-Saxon site, much as at Caister-on-Sea (Crabtree 1991, 106). Interestingly, the greater quantity of red deer antler (17 fragments) was matched by larger amounts of post-cranial bone (58 fragments). This might suggest that antler working was a minor activity, practised on occasion from antlers gathered from hunted animals. This, however, was not the case. Although the quantities of antler working waste from contemporary sites like Mühlberg and Saint-Clairsur-Epte are absent at West Stow (Teichert 1983; Thuet, this volume), the number of bone and antler objects from the site is considerable and it reflects a well-developed and localised craft. In general terms, the quantities of antler waste from early Anglo-Saxon sites are very small, and the number of objects is usually much greater. Antler waste is, however, recorded from most sites of this date (Riddler 1996).

At Barton Court Farm, Oxfordshire, the beam of a roe deer antler had been shaped by knife to a point (Figure 1 A). The object came from a late Roman enclosure ditch but was thought to be possibly of early Anglo-Saxon date (de Hoog 1984, microfiche 5:A11). Comparison with other objects, described below, indicates that this was a reasonable suggestion. A single, skull-attached antler came from a well at Pennyland (Riddler 1993, 119 and fig 62.74). One of its tines had been facetted by knife, but it was otherwise unworked. The poor condition of the deer bones from the site meant that it was not possible to identify them securely to species; post-cranial roe deer may have been present (Williams 1993, 152). Slightly greater and much better preserved quantities are known from West Stow in Suffolk, where five fragments of roe deer antler showed traces of modification by knife or draw-knife (Crabtree 1985, 96; 1991, 26 and table 10). Here there is also some evidence for the utilisation of the antler. One example has been modified to form an antler stamp (West 1985, fig 61.13; Riddler 1986, 18). Similar stamps, also made

Selected bones of roe deer were used in some cases for quite specific purposes. A collection of almost forty astragali were recovered from a cremation urn at Caistor-by-Norwich (Myres and Green 1973, 98-9, 114-7, pl XIXb and fig 5).

A

B

Figure 1 A: Roe Deer Antler Implement from Barton Court Farm, Oxfordshire (after de Hoog 1984); B: Roe Deer Antler Stamp from Hamwic. Both at 1:1 scale. 43

Materials of Manufacture

Most were the astragali of sheep but at least two were from roe deer and one of them included a runic inscription, the meaning of which is unclear, although it could conceivably include the term raha, ra, for roebuck (Page in Myres and Green 1973, 116-7). The set of astragali from Caister-byNorwich can be compared with groups from cremations on the Continent, and particularly with a cremation burial from Sörup, where at least 45 astragali were found in association with three rectangular dice (Raddatz 1981, 47 and taf 48; Knol 1987; Kuhl 1984; Lemkuhl 1985). Astragali of roe deer would have been interchangeable with those of sheep or goats and difficult to distinguish from them; the inscription may therefore have helped in that respect.

Barton Court Farm, Dublin and Ipswich (de Hoog 1984, fig 93.9; Riddler, Trzaska-Nartowski and Hatton forthcoming). A fragmentary example was recovered from the Buttermarket at Ipswich (Figure 2 A). They may have been used as punches on softer, organic materials like leather, although they lack percussion marks and their sharpened points may simply have served to incise those materials. The remaining roe deer antler from Ipswich shows few signs of working. Two burrs remain attached to part of a skull from a ninth or tenth century context at the Buttermarket. The other pieces are fragmentary and only a small portion survives of each antler. They consist mainly of segments of burr and pedicle, with parts of the accompanying beam. The beam has usually fractured and the tines are absent. Significantly, there is no sign with these fragments of the use of the saw in their dismemberment. Saws were used by antler and bone workers to separate sections of the raw material and were amongst their most important tools (Ulbricht 1978, 26; Riddler forthcoming A). Indeed, aside from the implement from Lower Brook Street, there are few indications of the working of the roe deer antler from Ipswich at all, although it was often found in association with worked bone and antler waste. Several deposits at the Buttermarket and Greyfriars Road can be identified as the locations of antler and bone working, on the basis of the quantity of waste materials found there (Riddler, Trzaska-Nartowski and Hatton, forthcoming). There are associations of roe deer antler, red deer antler and goat horn cores at Bridge Street and of red and roe deer antler within an assemblage at the Buttermarket (Crabtree 1996, 68; Riddler, Trzaska-Nartowski and Hatton forthcoming). Quantities of post-cranial roe deer from Ipswich are relatively small and amount to just twelve pieces (Crabtree 1996, table 1). 27 post-cranial fragments of red deer, in contrast, can be compared with over 2,300 pieces of red deer antler. 69.4% of the red deer antler is naturally-shed (Riddler, TrzaskaNartowski and Hatton forthcoming).

The quantity of roe deer antler from Middle Saxon sites is slightly greater than that seen previously in the early AngloSaxon period. Totals from rural sites accord with the earlier pattern; the difference lies in the amount seen within some of the pre-Viking trading and production centres. At Brandon, a roe deer antler had one of its tines removed and the other sharpened to a point. The guttering of the beam has been retained but most of the traces of the burr, including the coronet, have been removed by knife. It is similar to the example from Pennyland (Riddler 1993, 119 and fig 62.74) although the tine is curved. There is a resemblance, also, to the pointed implement from Barton Court Farm, which is also curved, although the remnant of a second tine is retained there (Figure 1 A). The quantities of post-cranial roe deer from Brandon and from the contemporary site of Wicken Bonhunt, provide two of the largest assemblages from AngloSaxon England, with 108 and 143 fragments respectively (Crabtree 1996, table 1). The quantities of roe deer from some of the Middle Saxon production and trading centres are themselves disappointing. A single piece of post-cranial roe deer came from excavations at Jubilee Hall in London (Cowie, Whytehead and Blackmore 1988, 154 and table 21). Another fragment came from the National Gallery site, further to the west, but here also, no roe deer antler was recovered (Whytehead, Cowie and Blackmore 1989, 164 and table 19). The Royal Opera House excavations produced three fragments of roe deer, but no distinction between post-cranial bone and antler was made in the report (Malcolm, Bowsher and Cowie 2003, 319, table 72). Slightly higher totals are recorded for the Middle Saxon phases of the site at Fishergate in York (O’Connor 1991, 259 and table 69).

The situation is similar at Hamwic. Amongst the material identified to date are ten fragments of roe deer antler, stemming from six sites, all of which are of eighth to ninth century date. They include three sections of tine, three fragments of tine and adjoining elements of beam, and one naturally-shed burr. As at Ipswich, most of these pieces came from features that contained copious amounts of worked antler and bone waste. This is the case, in particular, for the Six Dials sites, located within the northern part of Hamwic, where several antler and bone workshops can be identified (Riddler 2001; Andrews 1997, 227-30; Bourdillon, this volume). Here too, roe deer antler was adopted for a specific purpose, in this case in the production of stamps for use on leather and ceramics. One example (Figure 1 B) has come from Six Dials, alongside eight made from red deer antler (Riddler 1986, fig 1; Riddler forthcoming B). The Hamwic stamp has been produced from a tine sawn from a mature roe deer antler. The point of the tine has been sawn away laterally and a cruciform pattern has been incised there. A similar roe deer antler stamp has come from West Stow (West 1985, fig 61.13). This example utilises the beam as a handle, with its truncated end used as a stamp, alongside the rounded end of a tine.

Ipswich and Hamwic provide a little more substance. Thirteen fragments of roe deer antler have come from four separate sites at Ipswich (Riddler, Trzaska-Nartowski and Hatton, forthcoming). The earliest piece is an antler from a Middle Saxon context at Lower Brook Street, which has been sawn across the pedicle and its tines have been trimmed away to produce a pointed implement. It served either as a punch or as a tool designed to produce incisions on softer materials, like leather or ceramics. The modified antlers from Barton Court Farm and Brandon may have been used for a similar purpose; they differ mainly in the curve of the pointed end. Implements with straight shafts which taper to rounded points were also produced from red deer antler and are known from 44

I. Riddler: A Lesser Material

B

A

Figure 2 A: Roe Deer Antler Implement from Ipswich Buttermarket; B: Roe Deer Antler Implement from Westbury, Buckinghamshire (after Ivens, Busby and Shepherd 1995). Both at 1:2 scale.

roe deer, derived from a minimum of seventeen individuals (Ulbricht 1978, 107; Reichstein 1991, 85-6). The quantity of antler and post-cranial bone from Haithabu has to be seen in the context of the extensive excavations at the settlement and the overall abundance of skeletal materials there, and roe deer antler was no more common at Haithabu than at any other contemporary site. The assemblage includes eight skullattached antlers and eleven that are naturally-shed (Reichstein 1991, 89). Ulbricht established that roe deer antler was used at Haithabu for one specific type of needle. Two examples of this needle type were identified, both of which have broad, rectangular perforations and lightly-curved shafts. A third example has subsequently been discovered and comes from excavations in the harbour area of the settlement (Ulbricht 1994, 673). In addition, an unfinished roe deer antler needle was also recovered, which had been partially cut to shape, utilising the burr area as a stock (Ulbricht 1978, 82 and taf 52.5-7). Ulbricht noted that similar objects from Staraja Ladoga were interpreted as objects for binding and weaving (Ulbricht 1978, 82).

Summary details of post-cranial roe deer from Hamwic have been provided above by Bourdillon (tables 1, 2, 7 and 8). Extensive surveys of over 300,000 fragments of animal bone provided very little post-cranial roe deer or antler at all. Eight fragments came from Melbourne Street, including one calvarium with antler still attached, and a few fragments were recovered from other sites (Bourdillon and Coy 1980, 114 and table 17.1; Bourdillon, this volume). Red deer antler easily outnumbers post-cranial material of that species from Hamwic. 81.2% of the red deer antler is naturally-shed (Riddler, Trzaska-Nartowski and Hatton forthcoming). The post-cranial red deer shows a clear selection of skeletal elements that are associated with the antlers, or with the working of bone. In contrast, the small quantity of postcranial roe deer bone is more widely distributed over the body and presents no obvious association with antler or bone working.

ENGLAND AND THE CONTINENT IN THE EIGHTH TO TWELFTH CENTURIES

The Haithabu evidence is important for several reasons. The roe deer antler can be linked to the production of a specific object type and the means of working can be identified on the basis of the unfinished piece. The antler occurred in two concentrations, defining at least two episodes of deposition (Ulbricht 1978, 107 and kartierungen 17). The quantity of roe deer antler is substantially less than the amount of postcranial bone. This led Reichstein to suggest that it was obtained as a by-product of the hunting of the animal (Reichstein 1991, 88), echoing the situation outlined for the

The level of utilisation of roe deer antler and the limited range of objects produced from the material both recall the situation with contemporary continental sites. Nineteenth century excavations at Dorestad produced no bone or antler from roe deer at all, and the twentieth-century excavations provided only a single roe deer metatarsus (Prummel 1983, 34, 228, 246 and table 2). At Haithabu, 37 fragments of roe deer antler were discovered, as well as 142 fragments of post-cranial 45

Materials of Manufacture

Roman period in southern England. As a by-product, it nonetheless had a well-defined use. Following these hunting episodes, some of the antler at least was passed on to bone workers.

of the beam has been smoothed, and a forked object utilising two tines has been created (Biddle 1990, fig 114.996). This implement came from a context of the early thirteenth century, but it could conceivably be of a slightly earlier date.

To the south-east of Haithabu lies Starigard/Oldenburg, one of the principal Slavic fortified settlements of the ninth to twelfth centuries (Gabriel and Kempke 1988). A single roe deer antler came from a Phase 5 deposit there, of c. 10751225 (Prummel 1993, 132). It is a skull-attached example, shaped to a point in the manner of a number of the objects described above. Similar objects are known also from Mecklenburg and Neu Niehöhr but the largest collection of its type, consisting of 29 roe deer antler objects, came from a contemporary settlement at Berlin Spandau (Becker 1989, 136-8 and taf 42). 74 roe deer antlers of seventh to twelfthcentury date were recovered from excavations there, but over half of these were not worked at all, reflecting in greater measure the situation at Ipswich. The general reluctance to use a saw in separating the material also recalls the Ipswich assemblage (Becker 1989, 138). The awl-like objects utilised the beam and, in some cases, one or more of the tines, which were worked to round-sectioned points. The beam was smoothed and several of the objects were polished from use (Becker 1989, 138). The function of these multifarious objects is not clear but Becker felt that they were not simple awls for leather or dense textile and that they may have been used in some manner in the manufacture of cordage (Becker 1989, 139).

A modified roe deer tine from Saddler Street at Durham is substantially similar in form, and perhaps also function, to the series from Barton Court Farm, Brandon and Ipswich. It came from a medieval context. The burr and one of the tines have been removed, leaving a tine and section of beam. A small quantity of post-cranial roe deer from the site had been butchered (Carver 1979, 50). A medieval roe deer antler from Westbury (Figure 2 B) represents a variation on this form, however, with one tine broken away and the other facetted to a point. The burr area, however, has been neatly rounded with the coronet removed and it seems that this was the principal working surface for the object (Riddler 1995, 3923 and fig 184). A single antler from a twelfth-century context at Townwall Street, Dover (Figure 3) is skull-attached and little modified, although the beam has been facetted by knife, indicating an intention to transform the antler into an object, possibly another pointed implement. It recalls several of the Berlin Spandau implements (Becker 1989, taf 42). On the basis of the review above, quantities of roe deer antler appear to increase steadily over time within Anglo-Saxon England. Equally, however, their relative representation within worked antler and bone assemblages is fairly constant across a number of sites. On the Continent, the increased quantity seen at Schleswig, as against Haithabu, is worthy of note but roe deer antler is still a minimal presence at both sites. They are prominent at Berlin Spandau but half of the roe deer antlers there were not used at all, and comparatively little use was made of an abundant resource. A small range of implements that were produced from roe deer antler can be identified. They include stamps and pointed, awl-like implements with straight or curved terminals. At Haithabu, roe deer antler was used in the manufacture of one type of needle.

The contemporary settlement of Mikulčice has also produced several implements of roe deer antler (Kavánová 1995, 231 and abbn XLV.2, XLVII.10-11 and taf 32.6, 11 and 12). They include an unfinished pendant with a lateral hole bored close to the burr, and another example of an awl-like object, for which the coronet of the burr has been retained. Unused roe deer antlers were also recovered (Kavánová 1995, 247 and taf 35). Closer to Haithabu, its successor settlement at Schleswig has also produced a small quantity of roe deer antler (Ulbricht 1984, 16, taf 7.1 and 53.1-4 and 8). 32 fragments were recovered, a number of which had also been shaped to pointed implements (Ulbricht 1984, tafn 53.1 and 92.7). These objects are matched by the Slavic examples, as well as by several from Late Saxon and Anglo-Norman deposits in England.

In each case, roe deer antler was used to manufacture objects that were also produced in red deer antler. There is no category of objects for which roe deer antler was unequivocally the preferred material. Bourdillon and Coy suggested that ‘the main reasons for capturing this species would be for skin, meat and sport’ (Bourdillon and Coy 1980, 114). This was clearly the case in the Anglo-Saxon period, with the added proviso that antler obtained as a by-product of these activities could, on occasion, be utilised by resourceful antler workers. The occurrence of roe deer antler alongside red deer antler waste within several assemblages from both Hamwic and Ipswich implies that it was worked by craftsmen, and it was not a commodity available to all. It may have been passed to bone and antler workers in the same way as domestic animal bone, given that most roe deer antlers appear to come from slaughtered animals.

Two roe deer tines were encountered amidst the 31 fragments of antler from St. Peters Street, Northampton (Oakley 1979, 308). They are described as having ‘polished tips’, indicating some wear through handling or use. Small quantities of postcranial roe deer were also recovered (Williams 1979, table 39). A section of tine and beam from Beverley also has traces of wear, as well as some cross-hatching, on one of the surviving tines (Foreman 1991, 194 and fig 130.1155). It came from a tenth or early eleventh-century context and roe deer was present in reasonable numbers across almost all of the phases of the site, in contrast to red deer, which was almost completely absent (Scott in Armstrong, Tomlinson and Evans 1991, 219, 221 and table 39B). It can be compared with a roe deer antler from Winchester, for which the main section

Roe deer antler was not generally favoured and at some sites, like Haithabu, its utilisation may have been opportunistic and 46

I. Riddler: A Lesser Material

Bibliography ANDREWS, P., 1997, Excavations at Hamwic: Volume 2. Excavations at Six Dials, CBA Research Report 109, London ARMSTRONG, P., TOMLINSON D. and EVANS, D. H., 1991, Excavations at Lurk Lane, Beverley, 1979-82, Sheffield Excavation Reports 1, Sheffield BECKER, C., 1989, Die Geweihfunde vom Spandauer Burgwall, in A. von Müller and K. von Müler-Muči, Ausgrabungen, Funde und naturwissenschaftliche Untersuchungen auf dem Burgwall in Berlin-Spandau, Berlin, 101-274 BECKER, C., 1993, Zur Aufdeckung von Kausalitäten zwischen Ernährungsgepflogenheiten und Knochenverarbeitung, in H. Friesinger, F. Daim, E. Kanelutti and O. Cichocki, Bio Archäologie +. Frühgeschichtsforschung, Archaeologia Austriaca Monograph 2, Vienna, 133-57 BIDDLE, M., 1990, Object and Economy in Medieval Winchester, Winchester Studies 7ii, Oxford BOURDILLON, J. and COY, J., 1980, The Animal Bones, in P. Holdsworth, editor, Excavations at Melbourne Street, Southampton, 1971-1976, CBA Research Report 33, London, 79-118 CARVER, M. O. H., 1979, Three Saxo-Norman Tenements in Durham City, Medieval Archaeology 23, 1-80 COWIE, R., WHYTEHEAD, R. L. and BLACKMORE, L., 1988, Two Middle Saxon Occupation Sites: Excavations at Jubilee Hall and 21-22 Maiden Lane, Transactions of the London and Middlesex Archaeological Society 39, 47-164 CRABTREE, P., 1985, The Faunal Remains, In S. West, West Stow. The Anglo-Saxon Village, East Anglian Archaeological Reports 24, Gressenhall, 85-96 CRABTREE, P., 1991, West Stow, Suffolk: Early Anglo-Saxon Animal Husbandry, East Anglian Archaeological Reports 47, Gressenhall CRABTREE, P., 1996, Production and Consumption in an early Complex Society: Animal Use in Middle Saxon East Anglia, World Archaeology 28, 58-75 CRUMMY, N., l983, The Roman Small Finds from Excavations in Colchester 1971-9, Colchester Archaeological Reports 2, Colchester DAVIES, R. W., 1971, The Roman Military Diet, Britannia 2, 12242 DESCHLER-ERB, S., 1998, Römische Beinartefakte aus Augusta Raurica. Rohmaterial, Technologie, Typologie und Chronologie, Forschungen in Augst 27, Augst DETSICAS, A. P., 1966, An Iron Age and Romano-British Site at Stone Castle Quarry, Greenhithe, Archaeologia Cantiana 81, 13690 FOREMAN, M., 1991, The Bone and Antler, in P. Armstrong, D. Tomlinson and D. H. Evans, editors, Excavations at Lurk Lane, Beverley, 1979-82, Sheffield Excavation Reports 1, Sheffield, 183-96 GABRIEL, I. and KEMPKE, T., 1988, Zur Abfolge der Befestigungen in Starigard/Oldenburg, Bericht der RömischGermanischen Kommission 69, 48-54

Figure 3 Roe Deer Antler from Dover, Townwall Street (drawn by Beverley Leader). 1:1 scale.

GARDINER, M., 1997, The Exploitation of Sea-mammals in medieval England: Bones and their Social Context, Archaeological Journal 154, 173-95

based on its occasional availability. However, where red deer antler was also available, there was little call for roe and the range of implements produced from the raw material is small, if not entirely insignificant. As we learn more about the utilisation of animal resources, the role of different forms of antler and the mechanisms of their supply promise to be further clarified.

GRANT, A., 1983, The Animal Bones, In S. Johnson, Burgh Castle, Excavations by Charles Green 1958-61, East Anglian Archaeology Report 20, Gressenhall, 108-10 HARMAN, M., 1993, The Animal Bones, In M. J. Darling and D. Gurney, Caister-on-Sea. Excavations by Charles Green, 195155, East Anglian Archaeology Report 60, Gressenhall, 223-38 47

Materials of Manufacture HART, F. A., 1984, Excavation of a Saxon Grubenhaus and Roman Ditch at Kent Road, St. Mary Cray, Archaeologia Cantiana 101, 187-217

RIDDLER, I. D., 1995, Bone, Antler and Ivory, in R. Ivens, P. Busby and N. Shepherd, Tattenhoe and Westbury. Two Deserted Medieval Settlements in Milton Keynes, Buckinghamshire Archaeological Society Monograph Series 8, Aylesbury 1995, 318-9 and 390-3

de HOOG, V., 1984, The Finds, in D. Miles (editor), Archaeology at Barton Court Farm, Abingdon, Oxon, CBA Research Report 50, London, microfiche 5

RIDDLER, I. D., 1996, The Antler Waste, in R. J. Williams, P. J. Hart and A. T. L. Williams, Wavendon Gate. A Late Iron Age and Roman Settlement in Milton Keynes, Buckinghamshire Archaeological Society Monograph Series 10, Aylesbury, 130-6

KAVÁNOVÁ, B., 1995, Knochen- und Geweihindustrie in Mikulčice, in F. Daim and L. Poláček, Studien zum Burgwall von Mikulčice, Band 1, Spisy Archeologického Ústavu av čr Brno 2, Brno, 113-378

RIDDLER, I. D., 1998, Worked Whale Vertebrae, Archaeologia Cantiana 118, 205-15

KING, A., 1978, A Comparative Survey of Bone Assemblages from Roman Sites in Britain, Institute of Archaeology Bulletin 15, 20732

RIDDLER, I. D., 2001, The Spatial Organisation of Bone-Working at Hamwic, in D. Hill and B. Cowie, Wics: the Early Medieval Trading Centres of Northern Europe, Sheffield, 61-6

KNOL, E., 1987, Knucklebones in Urns: Playful Grave-goods in Early Medieval Friesland, Helinium 27, 280-8

RIDDLER, I. D., forthcoming A, The Technology of the Working of Bone, Antler, Ivory and Horn, in J. Bayley, editor, Archaeotechnology English Heritage Archaeological Reports, London

KUHL, I., 1984, Animal remains in cremations from the Bronze Age to the Viking Period in Schleswig-Holstein, in C. Grigson and J. Clutton-Brock, editors, Animals and Archaeology 4: Husbandry in Europe, British Archaeological Reports, International Series 227, Oxford

RIDDLER, I. D., forthcoming B, An Antler Stamp from Melbourne Street, Southampton RIDDLER, TRZASKA-NARTOWSKI and HATTON, forthcoming, An Early Medieval Craft. Antler and Boneworking from Ipswich Excavations 1974-1994, London

LEHMKUHL, U., 1985, Archaozoologische Betrachtung uber Astragalusfunde aus Brandgrabern der romischen Kaiserzeit in Mecklenburg und ihr Brauchtum, Jahrbuch für Bodendenkmalpflege in Mecklenburg 32, 241-51

ROES, A., 1963, Bone and Antler Objects from the Frisian Terp Mounds, Haarlem

MACGREGOR, A., 1985, Bone, Antler, Ivory and Horn. The Technology of Skeletal Materials since the Roman Period, London

RÜLCKER, J., 1986, Das Elchwild: Naturgeschichte, Ökologie, Hege und Jagd des europäischen Elches, Hamburg

MALCOLM, G., BOWSHER, D. and COWIE, R., 2003, Middle Saxon London. Excavations at the Royal Opera House 1989-99, MoLAS Monograph 15, London

SCHMID, E., 1965, Damhirsche im römischen Augst, Ur-Schweiz 29, 53-63 TEICHERT, M., 1983, Geweihreste aus der germanischen Siedlung bei Muhlberg, Kr. Gotha, Zeitschrift für Archaologie 17, 115-22

MEANEY, A. L., 1981, Anglo-Saxon Amulets and Curing Stones, British Archaeological Reports, British Series 96, Oxford

ULBRICHT, I., 1978, Die Geweihverarbeitung in Haithabu, Die Ausgrabungen in Haithabu Band 7, Neumünster

MYRES, J. N. L. and GREEN, B., 1973, The Anglo-Saxon Cemeteries of Caistor-by-Norwich and Markshall, Norfolk, London O’CONNOR, T. P., 1991, Bones from 46-54 Fishergate, The Archaeology of York. The Animal Bones 15/4, York

ULBRICHT, I., 1980, Knochen- und Geweihfunde aus AltSchleswig. Aussagen zu Wirtschaft, Handel und Gesellschaft, Lübecker Schriften zur Archäologie und Kulturgeschichte 4, 21114

OAKLEY, G. E., 1979, The Worked Bone, in J. Williams, St. Peter’s Street, Northampton. Excavations 1973-1976, Northampton Archaeological Monograph 2, Northampton, 308-17

ULBRICHT, I., 1984, Die Verarbeitung von Knochen, Geweih und Horn im mittelalterlichen Schleswig, Ausgrabungen in Schleswig. Berichte und Studien 3, Neumünster

PRUMMEL, W., 1983, Early Medieval Dorestad. An Archaeozoological Study, Excavations at Dorestad 2, Nederlandse Oudheden 11, Amersfoort

ULBRICHT, I., 1994, Beobachtungen an wikingerzeitlichen Gebrauchsnadeln aus Geweih aus dem Hafen von Haithabu, Marburger Studien zur Vor- und Frühgeschichte 16, 671-4

PRUMMEL, W., 1993, Starigard/Oldenburg. Hauptburg der Slawen in Wagrien. IV Die Tierknochenfunde, Neumünster

WEST, S. E., 1985, West Stow. The Anglo-Saxon Village, East Anglian Archaeology 24, Gressenhall

PUTMAN, R., 1988, The Natural History of Deer, London

WHYTEHEAD, R. L., COWIE, R. and BLACKMORE, L., 1989, Excavations at the Peabody site, Chandos Place, and the National Gallery, Transactions of the London and Middlesex Archaeological Society 40, 35-176

RADDATZ, K., 1981, Sörup I. Ein Gräberfeld der Eisenzeit in Angeln, Offa-Bücher 46, Neumünster REICHSTEIN, H., 1991, Die wildelebenden Säugetiere von Haithabu (Ausgrabungen 1966-1969 und 1979-1980), Berichte über die Ausgrabungen in Haithabu 30, Neumünster

WILLIAMS, J. H., 1979, St. Peter’s Street, Northampton. Excavations 1973-1976, Northampton Archaeological Monograph 2, Northampton

RIDDLER, I. D., 1986, Pottery Stamps - a Middle Saxon Viewpoint, Medieval Ceramics 10, 17-22

WILLIAMS, R. J., 1993, Pennyland and Hartigans, Buckinghamshire Archaeological Society Monograph 4, Aylesbury

RIDDLER, I. D., 1993, Saxon Worked Bone, in R. J. Williams, editor, Pennyland and Hartigans, Buckinghamshire Archaeological Society Monograph Series 4, Aylesbury, 107-119

48

J. Bourdillon: Bias from Bone Working at Middle Saxon Hamwic, Southampton

BIAS FROM BONE WORKING AT MIDDLE SAXON HAMWIC, SOUTHAMPTON Jennifer BOURDILLON

Studies of faunal remains from Hamwic, Middle Saxon Southampton, were carried out over a number of years at the Faunal Remains Unit, the University of Southampton, by Jennifer Bourdillon and Jennie Coy. Some of the results of these studies have been published whilst others remain, as yet, in archive (Bourdillon and Coy 1980; Bourdillon 1980; 1983; 1988; 1994; Coy 1981; 1982). The site of Hamwic remains central to any study of the Middle Saxon trading and production centres. It was founded in the later seventh or early eighth century and extended to over 50ha during the course of 150 years of occupation (Morton 1992, 26-9; Andrews 1997, 248-56; Morton 2001). Precious metals were worked there and objects of silver and copper alloy (and possibly gold) were produced, although specific centres of manufacture have yet to be recognised (Hinton 1996, 97-8; Andrews 1997, 219-21). The smithing of iron was carried out, if not the smelting of the raw material. Textile manufacture was prominent and trade with the Continental may have formed one reason for the location and scale of the settlement (Morton 1992, 56-7; Andrews 1997, 238-9; Hodges 1981).

representation, distribution over the body, measurements and fusion, tables on which the report’s broad interpretations were based. As noted above, concentrations of offcuts from Hamwic bone working had been found in the late 1960’s by Addyman and Hill, who used this evidence in a useful preliminary study of particular working techniques (Addyman and Hill 1969, 757). Faunal questions as such had not been examined at this time, but it was clearly going to be important to investigate any selection of material for bone working as a check on likely bias among the remaining bones, and Driver (1984) studied the bones from the next site which gave a good number of offcuts (Site 14) with this question to the forefront of his mind. The key question that lay behind these investigations was the extent to which bone and antler workers selected their raw material from the sample available. If deliberate selection existed, could it be recognised, analysed and quantified, and how did it affect the faunal remains assemblage ? What kind of bias was present and could it be seen in the choice of species, body part, age and metric analysis ? With its background of faunal studies and investigations into bone and antler working, and with a close collaboration between Southampton Museums Archaeology Section and the Faunal Remains Unit, Hamwic offered an ideal location for such a study.

The analysis of animal bones from Melbourne Street was instrumental in developing methods of study for large groups of material, and for comparing Hamwic results with those from Continental Europe. Subsequent studies dealt with questions of recovery, searched for groupings within pit assemblages and examined contextual variability (Bourdillon 1994, 121).

The Variability Study initiated with the Six Dials excavations included one pit chosen for its known concentration of worked offcuts (Site 31, F1005), and a further concentration was found in a feature selected on other grounds (the late well, Site 31, F2048), the latest - and the largest - bone working assemblage excavated from that site. This analysis was carried out in close association with the bone working study being made by Ian Riddler; he also made a new and fuller study of the worked material first published by Addyman and Hill, and of that from Site 14, and in the tables which support the present discussion the data on the worked material from these sites are largely derived from Riddler’s work.

Hamwic is known not only for its animal remains, but also for the working of both bone and antler. Early reviews of the settlement noted the presence of bone and antler waste, and related this to the finished products, including combs, spindle whorls, pins and pin-beaters (Addyman and Hill 1969, 75-7; Holdsworth 1976, 45-8). It was only with the work of Jonathan Driver, however, that the scale of bone working first became evident (Driver 1984; Riddler 1992; 2001). Subsequent studies have begun to outline the nature of the waste materials and objects from the settlement (Riddler 1986; 1992; 1993; 1997; 2001).

F 2048 at Site 31 was a circular pit some 2.0m in diameter and 1.3m deep. Its original function was not clear, but the two primary layers contained few bones. It may originally have been used for the disposal of cess (Andrews 1997, 228)1. Most of the animal bone was found in the general rubbish

The Melbourne Street assemblage produced no concentrations of bone working, though one or two pieces of worked material were found as background noise in many features: there was a total of 49 fragments of antler, nearly all of them sawn, plus some 20 worked offcuts of cattle bone and a single worked offcut from horse. These worked fragments were included in the normal tables for species

1

49

Note that in Andrews 1997 the pit is referred to by the number 5238 (Andrews 1997, 228 and fig 37).

Materials of Manufacture

layers. One of these contexts also contained considerable quantities of oyster and mussel shells. The main rubbish layers were overlain by other contexts which accumulated most likely as a form of infilling as the main layers sank into the ground, and these layers were also rich in bone. Hamwic pits frequently show the greatest concentration of bone working offcuts in the infilling layers at the top, but F 2048 was unusual in that there were a great many sawn offcuts in the rubbish layers as well. A total of 3,020 fragments of unworked animal bone were identified from normal recovery, as well as 2,724 worked offcuts, 1,790 of which are antler (Andrews 1997, 228).

Tables 1 to 6 show the considerable variety in the main worked concentrations under discussion, particularly from the pit F 2048, details of which are provided in Tables 1 to 4, and other contexts within Six Dials, against which comparisons are made in Tables 5 and 6. The results are very marked by the fragment count, and although the weights go some way towards moderating the first startling contrasts some important differences still remain. Significant quantities of cattle, sheep and horse bones occur as worked offcuts, particularly from F 2048, and elements of post-cranial red deer were also worked. Red deer antler is also noticeable for its presence in that assemblage. Where this worked bone and antler waste occurs, there appear to be differences in faunal representation, and these are investigated here.

This feature provided the greatest concentration of bone working debris from all of the Six Dials excavations. There is the added interest that it has been phased as quite late in the Six Dials sequence – probably soon after AD 850, which would make it similar in date to the well F2014 on Site 30, one of the main features which were taken to be late for purposes of comparison in the Six Dials Variability Study (Andrews 1997, 249-50).

These differences are not simply topographical: the assemblages in Tables 5 and 6 are all from the Six Dials’ excavations, which like Site 32 were in the north of the town, whereas Site 14 was at its southern edge. But similar disparities could be seen also within the Site 14 assemblage (Driver 1984; Morton 1992, 150-2 and fig 59).

Since the writing of the Melbourne Street report it has become clear that at Hamwic the pattern of deposition of worked offcuts is not like that of domestic bone rubbish. The unworked bones are generally homogeneous throughout the town, but for the worked material there is a general background of sporadic offcuts and also some particular worked concentrations that not only stand out against the background for their general abundance but also differ sharply between themselves in style and selection and technique (Riddler 2001). There may have been changes over time, for at Six Dials there would seem to have been a considerable increase in the emphasis on working in the later phases of occupation, and even within the features worked offcuts were strongly concentrated in the higher levels. Sieved recovery has been useful in producing some small fragile shavings of antler that have been of value in the assessment of working techniques, but otherwise it has added little that is new to the main faunal data in this field and for purposes of valid comparison the tables supporting the present discussion show material from normal recovery only.

Table 1

Within the phases of activity, too, important differences remain between early and late contexts. One main difference lies in the use of antler. The antler was consistent in itself in that it came overwhelmingly from red deer - no fallow deer bones of any sort have been found at Hamwic, and though a few bones of roe deer have been found as unworked rubbish, there are no post-cranial offcuts from this species and the use even of its antler was very rare. The use of roe deer at Hamwic and elsewhere in Anglo-Saxon England is described elsewhere in this volume (see Riddler, above). It is the use of antler relative to other bones that provides the first great contrast in the worked assemblages: F 2048 in particular has a noted concentration, while on Site 14 the dearth of antler contrasts with the abundance of both cattle and horse (Tables 1 and 7). There are various possible interpretations for these differences in the relative proportions of worked bone and antler, some of which have been explored recently (Riddler 2001). In this connection it is worth looking both relatively and absolutely at the other bones of deer (Tables 8 and 9). Table

Hamwic, Six Dials. SOU31, Feature 2048. Identified Fragments from Normal Recovery

a) excluding worked offcuts: Cow

Sheep

Goat

Pig

Horse

Dog

Top

288

287

2

76

11

Rubbish

937

1058

4

175

28

2

6

251

39

2

Bottom Total:

25

16

1250

1361

Cat 1

Fowl

Goose Red Deer Roe Deer Wild Bird Fish Total:

9

2

1

4

2

683

46

12

21

1

3

9

2296

55

14

21

2

7

11

3020

41 1

b) worked offcuts: Cow

Sheep

Horse

Top

55

1

4

Rubbish

718

17

54

Bottom

1

Total:

774

18

Red Deer 7

58

7

50

Antler

Total:

26

86

1602

2398

159

160

1787

2644

J. Bourdillon: Bias from Bone Working at Middle Saxon Hamwic, Southampton Table 2

Hamwic, Six Dials. SOU31, Feature 2048. Weights of Identified Mammal Fragments

a) excluding worked offcuts: Cow

Sheep

Goat

Pig

Horse

Dog

Top

4105

1150

30

450

165

5

Rubbish

14615

4890

175

2240

330

Bottom

720

240

19440

6280

205

Cow

Sheep

Horse

Top

840

20

135

Rubbish

8855

90 110

Total:

Cat

Red Deer

25

Roe Deer

Total:

10

5915

10

22585

300

960 2690

495

5

25

Post-Cranial

Antler

Total:

410

1405

890

70

7230

17135

15

45

1025

70

7655

18585

300

20

29460

b) worked offcuts: Red Deer:

Bottom

30

Total:

9725

Table 3

Hamwic, Six Dials. SOU31, Feature 2048. Unidentified Material Fragments: Not Worked

Weight (g):

Mean Fragment Weight (g)

Worked

Not Worked

Worked

Not Worked

Worked

Top

561

30

1005

100

1,8

3,3

Rubbish

1814

154

2555

735

1,4

4,8

184

3585

Bottom

2

Total:

Table 4

25

2377

12,5 835

1,5

4,5

Hamwic, Six Dials. Unidentified Material, excluding Worked Offcuts

SOU31, F2048 Identified

Unidentified:

Fragments

no.

% fragments

% weight

Mean fragment weight

Top

683

561

45,1

14,5

1,8

Rubbish

2296

1814

44,1

10,6

1,4

Bottom

41

2

4,7

2,5

12,5

3020

2377

44,1

11,2

1,5

Whole study

19182

9124

32,3

7,9

2,6

All pits

8046

3901

32,7

9,2

2,7

Pit F1005

852

961

53

16,5

1,3

Late contexts

3256

1619

33,3

8,1

2,9

Total: Six Dials Study:

Table 5

Hamwic, Six Dials. Mean Fragment Weights of Cattle, Sheep, Pig and Horse

SOU31, F2048 Cattle

Sheep

Pig

Horse

14,1

4

5,9

15

Rubbish

15,6

4,6

6,5

11,8

Bottom

27,7

15

All Contexts

15,5

4,5

6,3

12,7

Whole study

21,2

6,6

12,6

58,7

All pits

18,9

6

11,3

43,3

10

5,8

7,3

(7.9)

20,1

8

12,7

63,0

Top

Six Dials Study:

Pit F1005 Late contexts

51

Materials of Manufacture Table 6

Hamwic, Six Dials. SOU31, Feature 2048. Mean Fragment Weights of Worked Bone Fragments (g/fragment) SOU31, F2048 Red Deer: Cattle

Sheep

Horse

Top

15,3

(20.0)

33,8

Rubbish

12,3

5,3

16,5

10

Bottom

(30.0) 6,1

17,7

10

All Contexts

12,6

Post-Cranial

Six Dials Study:

Table 7

Pit F1005 offcuts:

15,0

2,5

F2014 offcuts:

26,2

15,0

75,0

SOU14. Chapel Road. Worked Bone and Antler Offcuts, by Fragment Count Principal Features:

Remainder: Overall Total

19

24

26

27

28

30

Metacarpus

33

123

1

72

15

77

3

324

Metatarsus

37

164

7

77

16

59

10

370

Cattle

Metapodial

2

4

Radius

15

29

2

27

2 5

31

2

111

Tibia

17

61

3

35

6

51

5

178

Radius or Tibia

2

1

10

5

6

Rib

1

8

1

3

5

1

19

Mandible

4

35

2

20

21

3

85

Scapula

3

7

2

23

3

60

18

4

Femur

2

4

2

Humerus

5

3

3

Ulna

8 1

12

1

Other

1

1

1

Horse Metacarpus

7

14

1

2

Metatarsus

3

5

1

Metapodial

1

2

1

15

2

3

18

2

Radius Tibia Ulna

1

Mandible

2

1

2 3

27 1

13 4

1

6 11

24 3

37 1

Carpal

3

5

1

1

Sheep Metapodial

2

2

Radius

1

Post-Cranial Red Deer Red Deer Antler

5

1

1

1

1

24

12

12

Roe Deer Antler Total:

3 5

1 131

523

19

285

8 shows the relative representation of wild species by fragment count from F 2048 and from the remainder of the Six Dials study. Table 9 contrasts quantities of post-cranial bone from F 2048 and Site 14. Bones of deer are never frequent, either as offcuts or unworked, but it is F 2048 that

53

315

58 1

36

1362

again shows their greatest concentration. There is clear selection, with the offcuts coming from the straighter limb bones - metapodia, radius and tibia - or from the skull at the removal of the antler, and most of the unworked deer bones could well have reached this feature in immediate association 52

J. Bourdillon: Bias from Bone Working at Middle Saxon Hamwic, Southampton Table 8

Hamwic, Six Dials. SOU31, Feature 2048. Relative Representation of Wild Species by Fragment Count (excluding Worked Offcuts)

SOU31, F2048 Identified fragments Top Layers

683

Rubbish

2296

Bottom

41

Post Cranial Red Deer Post Cranial Roe Deer n % n %

Wild Bird

Fish

n

%

n

%

1

0,1

4

0,6

2

0,3

21

0,9

1

0,0

3

0,1

9

0,4

3020

21

0,7

2

0,07

7

0,2

11

0,4

Whole study

19182

12

0,06

9

0,05

14

0,07

19

0,1

All pits

8046

2

0,02

3

0,04

4

0,05

Pit F1005

852

1

0,1

Late contexts

3256

4

0,1

All Contexts

Six Dials Study:

Table 9

4

0,1

4

0,1

1

0,1

3

0,1

Hamwic, Six Dials. SOU31, Feature 2048 and SOU14. Worked Offcuts. Distribution over the Body

SOU31, Feature 2048 Cow

Sheep

n

%

1

0,1

Scapula

1

0,1

Radius

12

1,6

Ulna

3

0,4

Mandible

Horse

n

%

1

5,6

Carpal

Post Cranial Red Deer

n

%

n

%

9

15,5

1

14,3

1

14,3

2

3,4

1

1,7

Tibia

16

2,1

5

27,7

15

25,9

2

28,5

Metacarpus

172

22,2

2

11,1

8

13,8

2

28,5

1

14,3

Metatarsus

199

25,7

3

16,7

10

17,3

Metapodial

349

45,1

3

16,7

13

22,4

Rib

21

2,7

4

22,2

SOU 14. All Contexts Cow

Sheep

n

% 7,2

Scapula

60

5,1

Radius

111

9,4

1

0,1

1

0,9

Tibia

178

15,1

37

33,0

Metacarpus

324

27,5

27

24,1

Metatarsus

370

31,4

13

11,6

Metapodial

10

0,8

4

3,6

Rib

19

1,6

Ulna

%

Post Cranial Red Deer

85

Mandible

n

Horse

1

33,3

Carpal

Femur

8

0,7

Humerus

12

1,0

2

66,7

n

%

5

4,5

24

21,4

1

0,9

Pedicle Total

1178

3

112

either with the antler or with those bones which provided the offcuts. The distribution over the body for roe deer, which comes rather more from the trunk, points to a contrast with the bones of red deer, which were not brought in as venison.

n

%

1

33,3

2

66,7

3

The results from the whole Variability Study confirm the general rarity of post-cranial deer (Table 8) and one can now say that the scarcity of post-cranial deer as found in the Melbourne Street assemblage does not seem to have been any diversion of such bones in good numbers into the working 53

Materials of Manufacture

concentrations. It may be of interest that the only bone working feature that gives any reasonable representation for post-cranial deer is the one which is also the highest in antler; and for the picture of Hamwic as a whole it may be worth remembering that this feature comes from the latest phase.

consistent was the preference for these bones. Only Site 14, with its many worked fragments of horse, breaks from this pattern and adds five worked fragments of mandible to the usual range (Tables 7 and 9). Horse radii, tibia and metapodial bones have also been found unworked and not linked with any offcuts, and it is not easy to make the appropriate correction for the bias produced for this selection for working (Table 11). It may be said, however, that the trunk bones of horse are gravely under-represented in the excavated material as a whole. It is the lower limbs that are normally found, and sometimes the heads, and this seems to be the pattern whether or not the material is worked. The Variability Study looked at further context types than the pits whose assemblages had so dominated earlier results, but though horse was found to be somewhat more common in the ditch (with a concentration from a single individual at the bottom of feature 2014) there is no real abundance anywhere. It would seem that the trunks of horses were generally disposed of well away from the town. At the same time, the appropriate lower limb bones were taken in to some of the centres of working - but they were not always so taken, and not to all such centres. They are very rare in other worked bone assemblages within northern Hamwic (Table 12). Since it was only parts of these horses that were selected, the lower limbs from the radius or tibia downwards, these could have been brought in to town from some distance. There is no need from the bone working evidence to postulate any large population of horses living in or near the town.

For the worked bones of sheep, too, it is F 2048 that shows more interesting abundance; but set against the vast numbers of sheep bones in the Hamwic record as a whole, their occasional use for working need give no concern about any possible bias in the main record. Study of the objects and waste indicates that they were used principally as the handles for handled combs, and they had little other use. Table 9 shows clear selection in those bones of the body which were worked in sheep, and the bone most favoured was the tibia (albeit only slightly above the remainder); but of these tibia offcuts some are proximal, others are distal and this further dilutes the threat of any serious bias in general population parameters such as those of size and age. Thus for post-cranial red and roe deer, and for sheep, there is no appreciable influence on the faunal remains record from their sporadic use in bone working. The numbers involved are relatively small and no appreciable bias can be seen. The use of post-cranial red deer may have been opportunist and relates perhaps to the occasional presence of carcasses or parts of carcasses in the town, as a by-product of the requirement for antler. On the other hand, there could indeed be a bias in the horse bones found at Hamwic. It was argued for Melbourne Street that since horses at Hamwic were eaten only rarely, if at all, they were not likely to be fairly represented in the food waste (Bourdillon and Coy 1980, 104-5). The finding of so many worked horse offcuts from Site 14 led Driver (1984) quite rightly to conclude that horses were more important at Hamwic than earlier figures had shown. Some questions, however, still remain.

At Melbourne Street all evidence of horse had been of animals that were fully mature (Bourdillon and Coy 1980, 89, 105 and fig 17.2). The picture has now slightly changed in that two fragments of young horse have been found among the worked offcuts from two other sites - a sawn distal metacarpus from Site 14, not only unfused but also quite small, and from F 2048 a distal tibia which though of good size was nevertheless unfused. No other evidence of immature horse has yet been found from Hamwic.

In many ways horse bones may be more suitable for working than are those of cattle: their radius and tibia have large flat surfaces and the structure of the main horse metapodia renders them inherently the more solid. Tables 10 and 11 show how

Horses, however, were of small importance in bone working when set against the predominance of cattle, and it has been the worked offcuts from cattle that have given the most

Table 10 Hamwic, Six Dials. SOU31, Feature 2048. Relative Representation of the less common domestic species, by fragment count SOU31, F2048 Horse

Goat

n

%

n

%

Top Layers

683

11

1,6

2

0,3

Rubbish

2296

28

1,2

4

0,2

2

0,1

3020

39

1,3

6

0,2

2

0,07

1

Whole study

19182

97

0,5

121

0,6

18

0,1

All pits

8046

6

0,07

14

0,2

3

0,04

Pit F1005

852

7

0,8

13

1,5

Late contexts

3256

52

1,6

32

1,0

Bottom All Contexts

n

Fowl

Cat

Dog

Identified fragments

%

Goose

n

%

n

%

n

%

1

0,1

9

1,3

2

0,3

46

2,0

12

0,5

0,03

55

1,8

21

0,7

14

0,07

185

1,0

118

0,6

6

0,07

90

1,1

83

1,0

1

0,1

13

1,5

1

0,1

6

0,2

21

0,6

41

Six Dials Study:

2

54

0,1

J. Bourdillon: Bias from Bone Working at Middle Saxon Hamwic, Southampton Table 11 Hamwic, Six Dials. SOU31, Feature 2048. Horse: Distribution over the Body unworked n % Skull

4

10,3

Tooth Fragment

1

2,6

offcuts n

%

Long Bones: Radius

9

Ulna

2

Tibia

1

7,7

2

44,8

15

Feet/Ankles: Carpal

1

Second Metacarpal

4

Fourth Metacarpal

2

Metacarpus

8

Metatarsus

2

Metapodial

2

Fourth Metatarsal

2

Tarsal

4

First Phalanx

Vertebra (cervical)

43,6

10

55,2

13

1 1

2,6

Scapula

3

7,7

Rib

10

25,5

Total:

39

58

Melbourne Street cattle metapodia had been taken to other areas for working. If these bones were lacking from the whole Melbourne Street record, some adjustments would be needed not only to the estimates of ageing but also to distribution over the body for cattle and to the relative frequency of the species.

frequent and subtle occasions of bias. Tables 7 and 9 show that for cattle, as for so much else, there were great differences in selection between the various offcut concentrations: Site 14, for example, makes strong use of the mandibles and in this it is almost unique. Taken all in all, however, it may be said that the use of the cattle metacarpus and metatarsus was very strong indeed, extending to over 85% of the worked cattle bones in each of the principal assemblages, with the exception of Site 14. If selection for bone working has led to serious sample bias it is therefore with the main metapodial bones that the earlier interpretations will have been the most at risk. In Table 12 offcut concentrations are provided for several other northern Hamwic assemblages and these emphasise this situation. Cattle metapodia are predominant and there is little use of other worked bone.

There is also the question of a likely selection for size. Driver (1984) found such selection in the worked cattle bones from Site 14, but with a wider range of material and with the realisation of the many differences between the various assemblages of offcuts it is right to look at the question again, and in particular the good assemblage of Site 31 F 2048 was the occasion for a metrical study in some detail. F 2048 gave few measurements of unworked cattle bones, and as a very late assemblage in the Hamwic sequence there was the chance that unusual sizes were the results not of selection for working but of broad animal changes over time. For this reason the tables below (Tables 13 to 19), though centred on the F2048 assemblage, draw also on a far wider range of material. The data for F 2048 are located near the centre of the page: the comparisons which are placed above are for groups which either span the main years of settlement at Hamwic or are phased earlier than F 2048, whilst below F 2048 in each table are placed first the late contexts extracted from the Variability Study, and then the data for Site 169, F 10675, which has been phased as probably the latest feature on Anglo-Saxon Six Dials. Data for worked and unworked material are presented first separately and then jointly, since in looking for possible selection one must look for the degree of variation within the sample as a whole. Data are presented

Some ageing discrepancies were noted in the Melbourne Street report, but these mainly concerned the presence or absence of quite young animals whose presence was indicated more strongly by the mandibles than by the evidence from epiphyseal fusion (Bourdillon and Coy 1980, 86-8). Such animals would be those with porous bones and no indication of these has been found in the bone working assemblages. There were, however, also some discrepancies in the fusion record within the broader age groups. At Melbourne Street the distal tibia showed a markedly higher proportion of older animals. Sawn and therefore worked material shows a consistently higher ratio of fusion than unworked material, and since also there was a far greater use of cattle metapodia than tibia in working, it seems probable that the lack of offcut concentrations in the Melbourne Street material, and the strong presence of these elsewhere, meant that some of the 55

Materials of Manufacture Table 12 Northern Hamwic. A Selection of Worked Bone Assemblages SOU24, F4006 Cow n

Sheep

Horse

n

Mandible Scapula Radius Ulna Carpal Tibia Metacarpus Metatarsus Metapodial Rib Total:

1 39 36 22

14

101

15

Percentage of Metapodia:

96,0

n

Post Cranial Red Deer n

Total:

3

1

1 0

1

117

SOU26, F8007 Cow n Mandible Scapula Radius Ulna Carpal Tibia Metacarpus Metatarsus Metapodial Rib Total:

2 45 41 10 1 104

Percentage of Metapodia:

92,3

Sheep

Horse

n

n

Post Cranial Red Deer n

5

2

0

2

0

106

SOU32, F 307 Cow n Mandible Scapula Radius Ulna Carpal Tibia Metacarpus Metatarsus Metapodial Rib Total:

3 61 33 1 1 103

Percentage of Metapodia:

92,2

Sheep

Horse

n

n

1 3

Post Cranial Red Deer n

2

1 1 1 1

3

1

108

here for those cattle bones that were the most commonly worked, i.e. the metapodia, with dimensions for both the proximal and distal metacarpus and metatarsus (Tables 15, 16, 18 and 19). In addition, radius and tibia are also considered (Tables 14 and 17). Figures 1 - 4 relate to the same material but are based directly on the individual measurements.

for the unworked material, but the coefficient of variation is higher. For Site 32 the measurements of worked radii are higher than those for the unworked, but this is in an unworked sample of only three bones, and for F 2048 the two radius offcuts include one that is very small indeed, so small that it had to be checked very carefully for deer.

The different bones of the body do not all present the same picture, but for each one the results are consistent between the tables and the histograms. For worked cattle radii (Table 14) there was no apparent selection for size. For the Variability Study not only is the offcut mean a little lower than the mean

Samples for the metacarpus (Tables 15 and 16) are generally much larger and one may say with greater confidence that both in F 2048 and in the other assemblages here examined no selection for size has been established. For the tibia on the other hand (Table 17) there may have been a certain 56

J. Bourdillon: Bias from Bone Working at Middle Saxon Hamwic, Southampton Table 13 Some Comparisons of Size Factors (The relative size of animals, seen against the Melbourne Street assemblage) Cow not worked Cow Offcuts All Cow Sheep Pig

Six Dials Main Study n:

100.6% 408

103.5% 139

101.3% 547

98.7% 561

99.5% 24

n:

3

102.3% 64

102.3% 67

98.0% 20

3

n:

100.9% 322

106.7% 108

102.4% 430

99.6% 373

98.3% 110

n:

98.8% 31

104.7% 177

103.9% 205

98.2% 78

97.8% 14

n:

97.8% 70

104.7% 61

101.0% 131

97.4% 91

98.9% 20

n:

98.6% 57

99.3% 5

98.7% 62

99.1% 80

94.3% 22

SOU 31 F1005 SOU 32 all features SOU31 F2048 Six Dials Study, Late Contexts SOU169, F10675

Table 14 Hamwic Cow Radius Proximal Width

x

Range

n

S

CV

73,9

60.0-96.1

116

7.0

9.5

73,8 72,9 73,7

62.2-94.1 64.7-82.6 62.2-94.1

34 5 39

7.9 8.0 7.9

10.8 11.0 10.7

73,2

61.3-86.0

14

7.5

10.2

73.2

61.3-86.0

14

7.5

10.2

Not Worked Offcuts All

(70.4)

70.4

1

Not Worked Offcuts All

71.6 (74.0) 71.9

65.7-90.3 69.9-81.7 65.7-90.3

21 3 24

5.6

7.8

5.6

7.8

Not Worked Offcuts All

69.5 (63.2) 67.9

67.8-71.3 54.1-72.4 54.1-72.4

6 2 8

1.4

2.1

5.3

8.2

Not Worked Offcuts All

73.4 (71.3) 72.9

65.4-82.1 64.7-82.6 64.7-82.6

9 3 12

6.0

8.2

6.7

9.2

Not Worked Offcuts All

70.2 (72.4) 70.5

59.2-84.4 72.4 59.2-84.4

6 1 7

8.3

11.8

7.6

10.8

Melbourne Street All Six Dials Variability Study Not worked Offcuts All

Six Dials Supplementary Study Not Worked Offcuts All

SOU 31, F1005

SOU 32, all features

SOU 31, F2048

Six Dials Study, Late Contexts

SOU 169, F10675

selection for size in the material of the Variability Study and perhaps in F 2048, but not in the small sample from Site 32.

the front and back legs lies in the respective breadth of their bones, in that the radius and metacarpus are generally wider and flatter and might well provide a more suitable working surface without a need to have a preference as to size, whereas the narrower shafts of tibia and metatarsus might have made some selection more desirable. Possibly, though, what one is seeing is the absence of many of the smaller (and most likely female) bones. The common sexing pattern for the Hamwic cattle was found at Melbourne Street be a ratio of roughly 2:1 for female to male, and the metatarsus

It is the cattle metatarsus that shows the greatest differences between worked and unworked bones, and for this bone of the body selection for size is quite clear and can be seen with both the proximal and the distal widths (Tables 18 and 19). In itself, this leads to further questions. Why is there apparent selection for the cattle metatarsus, but not for the metacarpus ? The first thought is that the difference between 57

Materials of Manufacture Table 15 Hamwic Cow Metacarpus Proximal Width

x

Range

n

S

CV

53,5

40.9-64.2

33

5,8

10,8

53,9 53,9 53,9

45.8-66.8 45.8-62.8 45.8-66.8

22 46 68

6,1 5,1 5,4

11,4 9,5 10

Not Worked Offcuts All

(53.5) (50.3) (52.7)

47.3-61.7 50.3

3 1 4

Not Worked Offcuts All

52,7 52,7

45.8-61.9 45.8-61.9

31 31

4,7 4,7

9,0 9,0

58,7

45.7-65.6 TO ADD TO ADD

7

6,3

10,7

54,2 54,2

45.2-65.9 45.2-65.9

65 65

4,8 4,8

8,9 8,9

Not Worked Offcuts All

(48.8) 57,3 55,4

47.4-50.7 49.3-62.8 47.4-62.8

3 10 13

5,2 5,9

9,1 10,7

Not Worked Offcuts All

(56.2)

47.9-60.6

4

(56.2)

47.9-60.6

4

Melbourne Street All Six Dials Variability Study Not worked Offcuts All

Six Dials Supplementary Study

SOU 31, F1005

SOU 32, all features Not Worked Offcuts All

SOU 31, F2048 Not Worked Offcuts All

Six Dials Study, Late Contexts

SOU 169, F10675

histograms (Figures 1 and 2) certainly do not bear this out. Perhaps the apparent selection for size has arisen not because the larger bones were deliberately taken for working and the smaller ones rejected, but because the larger ones more often left measurable offcuts whilst many smaller ones were worked so extensively that they are lost altogether when it comes to data on size.

One thing that is clear is that in considering the changing sizes of the animals like is best compared with like, so that worked and unworked material are separately compared, female with female where possible, and male with male. Where Size Factors are concerned (Table 13) it is to be remembered that the Melbourne Street material on which they were based had a minimal representation of worked offcuts, and where material comes from bone working assemblages a first supposition would be that it might well be somewhat larger. If this is so then the Size Factors in Table 13 need not rule out the wider interpretation offered in the Variability Study, that there was a slight overall decline in cattle sizes in the later years of the town.

Another possibility lies in the use to which the bones were put. It was noted above that the principal bones of sheep to be utilised for working were the tibia and metapodia. Both bone types were adopted as the handles for handled combs, a common comb type produced extensively at Hamwic during the eighth and ninth centuries. The other favoured bone type for the handles was the cattle metatarsus, but the cattle metacarpus was not used for this purpose. Selection of metatarsals for bones suitable for comb handles is inherently likely, and this may have affected the archaeological record. However, if the cattle metatarsus was preferentially selected in this way, there is no obvious accompanying reason to explain why selection should also be apparent with the cattle tibia as well. Was it simply that it came to the bone worker alongside the metatarsus, as a part of the back leg, or were there other reasons behind the selection ?

And it must be remembered that selection was not total: even for those bones of the body where clear anomalies are shown, some small bones still survive as measurable offcuts and some large bones were never used at all. We do not seem to be looking at any pattern of clear and repetitive organisation where animals were reared for their special suitability in providing raw material for craftsmen, but rather at one where craftsmen made good use of a wide range of what was available but where their preferences and ways of working have left certain patterns in the record. 58

J. Bourdillon: Bias from Bone Working at Middle Saxon Hamwic, Southampton Table 16 Hamwic Cow Metacarpus Distal Width

Melbourne Street All

x

Range

n

S

CV

55,9

48.6-67.1

49

5,8

10,4

57 56,2 56,4

50.0-70.3 49.1-69.8 49.1-70.3

12 40 52

7,1 5,5 5,8

12,5 9,8 10,3

Six Dials Variability Study Not worked Offcuts All

Six Dials Supplementary Study Not Worked Offcuts All

(56.6)

50.1-67.4

3

(56.6)

50.1-67.4

3

Not Worked Offcuts All

57,9 57,9

51.1-69.8 51.1-69.8

15 15

5,4 5,4

9,3 9,3

Not Worked Offcuts All

58,5 58,7 58,6

45.5-71.8 48.8-68.6 45.5-71.8

14 65 79

8,1 5,1 5,7

13,8 8,7 9,7

Not Worked Offcuts All

56,7 56,7

48.0-69.1 48.0-69.1

49 49

5,3 5,3

9,4 9,4

Not Worked Offcuts All

(50.4) 55,2 55

50.4 49.1-67.5 49.1-67.5

1 19 20

5,5 5,4

9,9 9,9

Not Worked Offcuts All

(56.0) (53.3) (55.3)

49.8-64.4 53.3 49.8-64.4

3 1 4

SOU 31, F1005

SOU 32, all features

SOU 31, F2048

Six Dials Study, Late Contexts

SOU 169, F10675

12

Worked

Unworked

10

Number

8

6

4

2

0 42-3

44-5

46-7

48-9

50-1

52-3

54-5

56-7

58-9

Breadth (mm)

Figure 1 Cow Metatarsus Distal Breadth: Six Dials Studies 59

60-1

62-3

64-5

Materials of Manufacture Table 17 Hamwic Cow Tibia Distal Width

x

Range

n

S

CV

56,8

49.1-67.9

111

4,7

8,3

57,1 61,3 57,7

48.3-66.7 55.0-68.4 48.3-68.4

35 6 41

4,7 5 4,9

8,2 8,2 8,5

56.2

49.7-66.1

6

5,6

9,9

56.2

49.7-66.1

6

5,6

9,9

Not Worked Offcuts All

(68.4)

68.4

1

Not Worked Offcuts All

57,7 56,4 57,5

50.3-68.1 52.0-61.6 50.3-68.1

41 5 46

4,7 4 4,6

8,1 7,0 7,9

Not Worked Offcuts All

54,9 (57.9) 55,8

52.4-61.1 48.5-66.2 48.5-66.2

7 3 10

3

5,6

5,1

9,1

Not Worked Offcuts All

56.6 59,8 57,6

50.1-65.5 55.0-64.8 50.1-65.5

11 5 16

4,4 4,1 4,4

7,7 6,8 7,7

Not Worked Offcuts All

56.6 (57.2) 56.7

52.3-62.6 57.2 52.3-62.6

9 1 10

3,7

6,5

3,5

6,1

Melbourne Street All Six Dials Variability Study Not worked Offcuts All

Six Dials Supplementary Study Not Worked Offcuts All

SOU 31, F1005

SOU 32, all features

SOU 31, F2048

Six Dials Study, Late Contexts

SOU 169, F10675

12

Worked

Unworked

10

Number

8

6

4

2

0 42-3

44-5

46-7

48-9

50-1

52-3

54-5

56-7

58-9

Breadth (mm)

Figure 2 Cow Metatarsus Distal Breadth: Site 32, all contexts 60

60-1

62-3

64-5

J. Bourdillon: Bias from Bone Working at Middle Saxon Hamwic, Southampton Table 18 Hamwic Cow Metatarsus Proximal Width x Melbourne Street All

Range

n

S

CV

43,5

38.2-48.6

26

2,5

5,8

44,4 48,3 46,1

38.8-52.3 40.9-54.1 38.8-54.1

27 21 48

4,3 3,1 4,3

9,7 6,5 9,2

Six Dials Variability Study Not worked Offcuts All

Six Dials Supplementary Study Not Worked Offcuts All

(47.9)

41.0-51.5

3

(47.9)

41.0-51.5

3

Not Worked Offcuts All

48,4 48,4

40.9-52.1 40.9-52.1

10 10

3,3 3,3

6,8 6,8

46,1

42.1-51.2 TO ADD

6

6,8

3,2

47,1 47,1

38.6-54.5 38.6-54.5

33 33

3,5 3,5

7,5 7,5

41.6 48,2 45,5

39.2-47.8 42.4-54.1 38.8-54.1

7 10 17

3,0 3,3 4,5

7,2 6,8 10,0

SOU 31, F1005

SOU 32, all features Not Worked Offcuts All

SOU 31, F2048 Not Worked Offcuts All

Six Dials Study, Late Contexts Not Worked Offcuts All

SOU 169, F10675 Not Worked Offcuts All

12

Worked

Unworked

10

Number

8

6

4

2

0 42-3

44-5

46-7

48-9

50-1

52-3

54-5

56-7

58-9

60-1

62-3

64-5

Breadth

Figure 3 Cow Metacarpus Distal Breadth: Six Dials Studies 61

66-7

68-9

70-1

71-2

Materials of Manufacture Table 19 Hamwic Cow Metatarsus Distal Width x

Melbourne Street All

Range

n

S

CV

50,4

44.6-62.3

43

4,0

8,0

51,4 54,1 53,2

45.5-61.1 40.9-59.6 40.5-61.1

13 24 37

4,6 4,4 4,6

8,9 8,2 8,7

Six Dials Variability Study Not worked Offcuts All

Six Dials Supplementary Study Not Worked Offcuts All

(46.2)

46.2

1

(46.2)

46.2

1

Not Worked Offcuts All

51,2 51,2

45.5-55.9 45.5-55.9

5 5

4,9 4,9

9,5 9,5

Not Worked Offcuts All

51,7 56,4 55,3

45.0-61.8 49.7-63.6 45.0-63.6

11 29 40

5,1 2,8 4,0

9,8 5,0 7,2

Not Worked Offcuts All

56,0 56,0

48.2-62.1 48.2-62.1

22 22

3,7 3,7

6,6 6,6

Not Worked Offcuts All

55,0 55,0

46.0-59.6 46.0-59.6

13 13

4,4 4,4

8,0 8,0

Not Worked Offcuts All

(49.9) (49.9)

45.6-53.2 45.6-53.2

2 2

SOU 31, F1005

SOU 32, all features

SOU 31, F2048

Six Dials Study, Late Contexts

SOU 169, F10675

16

Worked

Unworked

14

12

Number

10

8

6

4

2

0 42-3

44-5

46-7

48-9

50-1

52-3

54-5

56-7

58-9

60-1

62-3

64-5

Breadth (mm)

Figure 4 Cow Metacarpus Distal Breadth: Site 32, all contexts 62

66-7

68-9

70-1

71-2

J. Bourdillon: Bias from Bone Working at Middle Saxon Hamwic, Southampton

14

Worked Unworked

12

10

Number

8

6

4

2

0 42-3

44-5

46-7

48-9

50-1

52-3

54-5

56-7

58-9

60-1

62-3

64-5

66-7

68-9

70-1

71-2

Breadth (mm)

Figure 5 Cow Tibia Distal Breadth: Six Dials Studies

12

Worked Unworked 10

Number

8

6

4

2

0 42-3

44-5

46-7

48-9

50-1

52-3

54-5

56-7

58-9

60-1

62-3

64-5

Breadth (mm)

Figure 6 Cow Tibia Distal Breadth: Site 32, all contexts 63

66-7

68-9

70-1

71-2

Materials of Manufacture

12

Worked Unworked 10

Number

8

6

4

2

0 42-3

44-5

46-7

48-9

50-1

52-3

54-5

56-7

58-9

60-1

62-3

64-5

66-7

68-9

70-1

71-2

Breadth (mm)

Figure 7 Cow Tibia Distal Breadth: Six Dials late contexts

Bibliography

DRIVER, J., 1984, Zooarchaeological Analysis of Raw Material Selection by a Saxon Artisan, Journal of Field Archaeology 11, 397-403

ADDYMAN, P. V. and HILL, D. H., 1969, Saxon Southampton: a review of the evidence. Part II: Industry, Trade and Everyday life, Proceedings of the Hampshire Field Club and Archaeology Society 26, 6l-96

HINTON, D. A., 1996, The Gold, Silver and other Non-Ferrous Alloy Objects from Hamwic, and the Non-Ferrous Metalworking Evidence, Southampton Finds Volume Two, Stroud HODGES, R., 1981, The Hamwic Pottery: the Local and Imported Wares from 30 years’ Excavations at Middle Saxon Southampton and their European Context, CBA Research Report 37, London

ANDREWS, P., 1997, Excavations at Hamwic. Volume 2: Excavations at Six Dials, CBA Research Report 109, London BOURDILLON, J., 1980, Animal Husbandry in the Southampton area as suggested by the excavated bones, Proceedings of the Hampshire Field Club and Archaeological Society 36, 181-91

HOLDSWORTH, P. E., 1976, Saxon Southampton: A New Review, Medieval Archaeology 20, 26-61 MORTON, A. D., 1992, Excavations at Hamwic. Volume 1: excavations 1946-83, excluding Six Dials and Melbourne Street, CBA Research Report 84, London

BOURDILLON, J., 1983, The animal bone of Hamwih – some comparisons, Archaeozoology I, Sczeczin, 515-23 BOURDILLON, J., 1988, Countryside and Town: the Animal Resources of Saxon Southampton, in D. Hooke, editor, AngloSaxon Settlements, Oxford, 177-95

MORTON, A., D., 2001, Southampton, in D. Hill and R. Cowie, editors, Wics. The Early Medieval Trading Centres of Northern Europe, Sheffield Archaeological Monographs 14, Sheffield, 8992

BOURDILLON, J., 1994, The Animal Provisioning of Saxon Southampton, in J. Rackham, editor, Environment and Economy in Anglo-Saxon England, CBA Research Report 89, London, 1206

RIDDLER, I. D., 1986, Pottery Stamps - a Middle Saxon Viewpoint, Medieval Ceramics 10, 17-22 RIDDLER, I. D., 1992, Boneworking and pre-Viking Trading Centres, in Art and Symbolism. Medieval Europe 1992, Preprinted Papers 7, York, 149-156

BOURDILLON, J. and COY, J., 1980, The Animal Bones, in P. Holdsworth, editor, Excavations at Melbourne Street, Southampton, 1971-1976, CBA Research Report 33, London, 79118

RIDDLER, I. D., 1993, Worked Bone, in M. F. Garner, Middle Saxon Evidence at Cook Street, Southampton (SOU254), Proceedings of the Hampshire Field Club and Archaeology Society 49, 113-6

COY, J., 1981, Animal Husbandry and Faunal Exploitation in Hampshire, in S. J. Shennan and R. T. Schadla-Hall, editors, The Archaeology of Hampshire, Hampshire Field Club and Archaeology Society Monograph 1, Winchester, 95-103

RIDDLER, I. D., 1997, Combs with Perforated Handles, Archaeologia Cantiana 117, 189-98

COY, J., 1982, The Role of Wild Fauna in Urban Economies in Wessex, in H. K. Kenward and A. Hall, editors, Environmental Archaeology in an Urban Context, CBA Research Report 43, London, 107-16

RIDDLER, I. D., 2001, The Spatial Organisation of Bone and Antler Working in Trading Centres, in D. Hill and R. Cowie, editors, Wics. The Early Medieval Trading Centres of Northern Europe, Sheffield Archaeological Reports 14, Sheffield, 61-6 64

I. Riddler and N. Trzaska-Nartowski: Late Saxon Worked Antler Waste from Holy Rood, Southampton

LATE SAXON WORKED ANTLER WASTE FROM HOLY ROOD, SOUTHAMPTON (SOU106) Ian RIDDLER and Nicola TRZASKA-NARTOWSKI

INTRODUCTION

been adopted on Anglo-Saxon assemblages, but others have not. The principal Anglo-Saxon assemblages to have been published, to date, are those from York (Rogers 1993, 124564; MacGregor, Mainman and Rogers 1999, 1903-22). Accordingly, various analytical techniques have been applied to the Holy Rood material, even though the assemblage is quite small, as a precursor to studies that explore antler and bone working in Anglo-Saxon England in greater detail (Riddler, Trzaska-Nartowski and Hatton forthcoming; Riddler forthcoming A and B).

The Anglo-Saxon archaeology of Southampton can be separated into two principal components. The Middle Saxon settlement at Hamwic has been extensively published, although elements of its archaeology remain to be analysed. The late Saxon settlement, in contrast, has been excavated but is not, as yet, comprehensively published. It sits, to some extent, in the shadow of Winchester, much of whose late Saxon archaeology has still to appear in print. In any consideration of these sites, the activities of craftsmen are of particular importance. Episodes in the production of objects can be glimpsed from a study of the waste materials and even small assemblages can be rewarding, in this respect. This contribution considers an assemblage of 77 fragments of antler, retrieved from excavations within late Saxon Southampton at Holy Rood (SOU106), and it is concerned with methods of study as well as with results. The study of antler waste was pioneered in Poland on assemblages from Wolin and Gdansk (Cnotliwy 1956; 1958; 1970; 1973). Techniques utilised there were further developed by Ingrid Ulbricht in her study of antler waste from Haithabu, and of waste of both bone and antler from Schleswig (Ulbricht 1978; 1984). Ulbricht’s work, in particular, has proved to be influential on subsequent studies of antler waste, alongside the analyses of Ambrosiani, Becker and Teichert, as well as Thuet (Ambrosiani 1981, 98-127; Becker 1989; Teichert 1983; Thuet, this volume). Some of those techniques have

THE QUANTITY OF WASTE 77 pieces of red deer antler, weighing 1,719g, were retrieved from seventeen separate contexts at Holy Rood. All of the antler can be identified as deriving from red deer. The various elements of the assemblage are quantified in Table 1. They include eight burrs, twenty-two tines, sixteen fragments stemming from junctions of the tine and beam and two segments of cortile tissue. Five of the eight burrs (1-5), or 62.5% of the sample, are naturally-shed. This is the customary situation for early medieval waste antler. With the exception of one assemblage from Schleswig, where the ratio of naturally-shed to skullattached burrs approximated to 1:1, and an assemblage from Chester where the ratio was 9:7 (Ulbricht 1984, 67; Ward

Table 1 Antler Waste from Holy Rood, Southampton Feature

Tines

19 20 52 54 55 58 59 61 79 83 84 85 86 94 95 96 97 Total:

1 2 7 2 1

Burrs

Crowns

Beam or Tine Sections

1

Beam/Tine Junctions

Cortile Tissue

Unfinished Objects

5

1 1 1 1

1 2 2 1

1 1 1

1 3

1 4

2 2

1 2

22

8

3

1 1 5 3

11 2 1 24

15 65

2

3

2

3

Total 1 2 13 2 1 1 2 3 3 1 1 2 5 1 25 13 1 77

Materials of Manufacture

1994, 92-3), as well as the assemblage from Saint-Clair-surEpte described above by Annick Thuet, naturally-shed burrs are usually more common than those still attached to the skull of the animal, and occur in a ratio of 2:1 or more (MacGregor 1985, 35; Riddler forthcoming A). The number of burrs from Holy Rood exceeds the number of crowns, a familiar situation with such assemblages, which can be seen also at Ipswich and York, amongst other sites (Rogers 1993, 1250-1; MacGregor, Mainman and Rogers 1999, tables 160-1; Riddler, Trzaska-Nartowski and Hatton forthcoming).

suggesting that a number of the tines were further worked. There is some evidence for this practice within the assemblage, as noted below. Relative totals of antler can also be assessed in very broad terms by weight. Two modern comparable red deer antlers, forming a pair shed by a single animal, with six tines on each, weigh 1.21 and 1.23kg. The antler from Holy Rood weighs 1,719g, providing an average of 215g per antler, or around 18% of a comparable modern antler. Due allowance has to be made for the weight loss incurred by antler which has been buried for over a millennium, but only a quarter to a third of each antler was recovered, to judge from weight alone, and more than half is missing, when the surviving elements of the antler are considered. Much of the missing material will have been used in object production and it was not recovered by excavation. To look at the assemblage in another way, around two-thirds of each antler was used in object production, and the remaining one third was discarded.

The surviving pieces allow estimates to be made of the minimum number of antlers originally present, as well as indicating what may now be missing from the assemblage. The eight surviving burrs are accompanied by three crowns and twenty-two tines. Mature red deer can easily obtain six to eight tines per antler and MacGregor has suggested that an average of six tines could be found on each antler used for working (MacGregor 1985, 14). With eight burrs, therefore, 48 tines would be expected, as well as eight crowns. Less than half of the anticipated quantity of tines and crowns were actually recovered, however. Elsewhere, it has been noted that the number of burrs corresponds more readily with antlers including just four tines, on average, generally because several tines have often been removed from each antler to be used as implements (Riddler forthcoming A). At Holy Rood there are roughly three times the number of tines to burrs,

The surviving fragments can be located on a diagram to show their distribution across the antler (Figure 1). They consist largely of tine ends, burrs, crowns and sections of beam or tine, together with their junction points. The missing elements are the cortile tissue, almost all of which is absent, as well as most of the sections of the antler beam. This is a common situation, which is repeated with the Saint-Clair-sur-Epte assemblage above.

METRICAL ANALYSIS Seven of the eight burrs (1-4 and 6-8) can be measured, using the system first outlined by Müller-Using for Wolin and subsequently applied to material from a number of early medieval sites, including Haithabu, Hamwic, Burgh Castle and Ipswich (Müller-Using 1953; Reichstein 1969; 1991, 769; Ulbricht 1978, diagramme 77; Riddler, Trzaska-Nartowski and Hatton forthcoming). Under this system the

Figure 2. Measurement Points for Antler Burrs (after Müller-Using 1953)

Figure 1 Distribution of Waste from Feature 96, located on a Red Deer Antler 66

I. Riddler and N. Trzaska-Nartowski: Late Saxon Worked Antler Waste from Holy Rood, Southampton

circumference of each burr is recorded, where possible, and four measurements are taken across the burr and its coronet (Figure 2). These measurements are presented as graphs in Figures 3 to 7.

includes over 400 burrs. The Holy Rood burrs appear to be relatively large, when viewed against that sample. Two have circumferences over 200mm and the third, at 167mm, is still above the mean circumference figure for Hamwic, of 163.7mm.

The circumferences of four burrs (1-3 and 7) could be measured and they have been plotted against the Hamwic data in Figure 3. The site of Hamwic was located to the east of the late Saxon settlement at Southampton and flourished as a production and trading centre from the late seventh to the mid to late ninth century (Morton 1992; Garner 1993, 1997 and 2001; Andrews 1997). Copious quantities of both bone and antler waste have come from Hamwic, principally from three areas of the settlement (Riddler 1992; 2001 and forthcoming A and B). The antler waste from Hamwic

This tendency is confirmed if the individual diameter measurements for the late Saxon burrs are also considered. In the first diameter measurement (Figure 4) the Holy Rood burrs are evenly dispersed across the sample. In the second, however, they cluster towards the larger end of the scale (Figure 5). This is the case also for the two width measurements (Figures 6 and 7). In general terms, therefore, they form a group of relatively large burrs, when compared with the Hamwic standard.

10

Number

Hamwic

Holy Rood

5

0 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 Circumference (mm)

Figure 3. Antler Burr Circumferences from Hamwic and Holy Rood, Southampton

10

Number

Hamwic 1

Holy Rood 1

5

0 40 42 44 46 48 50 52 54 56 58 60 62 64

66 68 70 72 74 76 78 80 82 84 86 88 90 92

Diameter (mm)

Figure 4. First Diameter Measurement, Holy Rood and Hamwic 67

Materials of Manufacture 10

Number

Hamwic 2

Holy Rood 2

5

0

30

32

34

36

38

40

42

44

46

48

50

52

54

56

58

60

62

64

66

70

68

72

74

Diameter (mm)

Figure 5. Second Diameter Measurement, Holy Rood and Hamwic

Hamwic3

Holy Rood 3

Number

15

10

5

0 1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

31

33

35

37

39

41

Diameter (mm)

Figure 6. Third Diameter Measurement, Holy Rood and Hamwic

Any conclusions to be drawn from these measurements must necessarily be tentative, given the small size of the sample. It can nonetheless be suggested as a working hypothesis that the deer whose antlers were available for working within late Saxon Southampton may have been a little larger than those used on average by the Middle Saxon bone and antler workers of nearby Hamwic. In Figure 8 the Holy Rood burrs are compared for their circumference measurements against those from late Saxon deposits at West Quay in Southampton (Riddler forthcoming C), and the tendency for the late Saxon

burrs to be larger than those from the Middle Saxon settlement is emphasised once again.

THE NATURE OF THE WASTE MATERIAL Features 52, 95 and 96 stand out for the quantity of their antler, above the small amounts retrieved from the remaining features. The site records suggest that all three features were 68

I. Riddler and N. Trzaska-Nartowski: Late Saxon Worked Antler Waste from Holy Rood, Southampton 25 Hamwic 4

Holy Rood 4

20

Number

15

10

5

0 20

22

24

26

28

30

32

34

36

38

40

42

44

46

48

50

52

54

56

58

60

Diameter (mm)

Figure 7. Fourth Diameter Measurement, Holy Rood and Hamwic

10

Number

Hamwic

Holy Rood

West Quay

5

0 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 Circumference (mm)

Figure 8. Antler Burr Circumferences for Hamwic, Holy Rood and West Quay, Southampton

rubbish pits, and that feature 52 may have been later in date than the other two pits, although this is not certain. The contents of each of these features are set out in Table 2 below. The material from these features can be grouped into three processual categories (as well as a miscellaneous group), following the system adopted by Christophersen (1980, 15964), which distinguishes between

1 2 3 4

unused raw material secondary waste material semi-manufactures miscellaneous material

The first column of Table 2 is a concordance of these categories with this assemblage. There is no antler from this 69

Materials of Manufacture Table 2

Antler Waste from Features 52, 95 and 96

Category

Waste Type

1 1 1 2 2 2 3 Total:

Burrs Tines Crowns Beam or Tine Sections Beam/Tine Junctions Cortile Tissue Semi-Manufactures

Feature: 95 2 1 1 11 5 2 3 25

52 7 1 5

13

assemblage that would correspond with Christophersen’s category 4, miscellaneous material.

96 2 4 2 2 3

13

Others 4 10 10 2

26

The three sections of crowns (9-11) were recovered from features 95 and 96. The tine ends have been removed from each one by sawing or, in one case, by fracturing them with the aid of a knife. Each crown was then removed from the beam by lateral sawing just below the tines. Where the beam was fairly broad it has been sawn from several directions, using the same method as with the burrs.

Category 1 waste material consists of burrs, crowns and tines. These are the main elements present within the waste assemblage and they represent material that was discarded at an early stage of working, when the antler was being dismembered. The burrs have been discussed above in terms of their size. Four of the eight burrs have been sawn across the beam above the brow tine and in each case that tine has also been sawn away. The beam has been sawn from two or more directions and the cortile tissue at the centre was then snapped away. The skull-attached burr from feature 96 (7) has been treated in the same way. In one case most of the coronet around the burr has been removed by knife.

Tine ends (12-33) customarily form the most common element of antler waste. They were generally discarded but were occasionally adapted for use as implements, usually with comparatively little modification. Two modified tines (32-3) were identified within the assemblage and they are shown in Figure 10. Both are facetted on two faces, but they differ in other respects.

In contrast, a burr from feature 95W (6) has been heavily modified (Figure 9 A). The beam and the brow tine have been removed (the latter by knife, rather than by saw) and the coronet of the burr has been cut away. One side of the burr above the coronet shows a series of more than twenty lateral incisions, all of which were made with a knife. They are neatly confined to one side of the burr, below the brow tine; the reason for their presence is not clear.

The tine from feature 96 (32) is a short, wedge-like implement, which has been broken in antiquity (Figure 10 C). It retains a series of knife marks. Schoknecht and Ulbricht noted the presence of antler wedges amidst the antler waste from Menzlin and Haithabu (Schoknecht 1978, 68 and taf 32.25.5-11; Ulbricht 1978, 50 and tafn 27 and 45). They have since been recorded from numerous sites in England, mostly in contexts of ninth to eleventh century date (Riddler, Trzaska-

A

B

Figure 9. A: Modified Burr from Feature 95W; B: Section of beam with incised scoring mark, from Feature 52 70

I. Riddler and N. Trzaska-Nartowski: Late Saxon Worked Antler Waste from Holy Rood, Southampton

Nartowski and Hatton forthcoming). They were particularly common at Coppergate in York, where two distinct types were identified (MacGregor, Mainman and Rogers 1999, 1997-8 and 2043-5). The first type, which is seen here, is facetted to a point on one or several faces. The second type also includes a notch cut into one face, and these implements may not have been used for the same purpose, although Ulbricht noted that those from Haithabu often included traces of hammering on the broad face (Ulbricht 1978, 83 and taf 45.6. 7 and 15; MacGregor 1985, 57; MacGregor, Mainman and Rogers 1999, 1998; Halpin 2000, 174).

they are widely distributed across northern and central Europe (Ulbricht 1978, 83-4 and taf 46; MacGregor, Mainman and Rogers 1999, 1998-9; Kavánová 1995, 218). Two forms were identified at Haithabu, and a third type has been added from Mikulčice (Ulbricht 1978, 83-4; Kavánová 1995, 220-2). A small, straight variant was the most common at Haithabu, if not elsewhere, and it often included a lateral perforation at the hollowed end. These may have served as points for antler spears, although other functions are also possible (Ulbricht 1978, taf 46.1-16; Kavánová 1995, 220). Longer, curved tine ends, of the type seen here, were less common at Haithabu, but present in some numbers at both York and Mikulčice. The hollowing of the blunt end is seldom very deep with the Anglo-Saxon examples of this type and it is not sufficient for the tang of an implement, or for a wooden shaft (MacGregor, Mainman and Rogers 1999, 1998). This suggests that some of these hollowed tines were residues from antler working, rather than implements in their own right. Equally, however, the example from Holy Rood has been carefully tapered on two faces to produce a rounded end, and it would have functioned very well as an implement used to score surfaces with the narrow, facetted end. It can be compared, in this respect, with a similar implement from a contemporary deposit at the Buttermarket, Ipswich (Figure 10 B).

A wedge from Haithabu was found embedded in a length of antler and it had been used to split the beam into sections (Ulbricht 1978, 50 and taf 27; MacGregor 1985, fig 34). The Holy Rood wedge was probably used for the same purpose. Ambrosiani’s experimental work showed that it was not necessary to sharpen them to a fine point, when using them in this way (Ambrosiani 1981, 112). The second implement (33) is longer and includes part of the curve of the end of the tine (Figure 10 A). It has also been sharpened on two faces, and it is partially hollowed. Similar implements are again known from Haithabu and York, and

C

B

A Figure 10. Antler Tine Implements. A: Holy Rood; B: Ipswich, Buttermarket; C: Fragmentary Wedge from Holy Rood; 71

Materials of Manufacture

The waste material of category 2 consists of 41 fragments of antler. Most of this material is unexceptional and consists of sections of discarded surface material or cortile tissue. The latter is poorly represented, however, with just two small fragments (34-5). The largest piece of antler is represented by a hollowed section of beam (61) from feature 96BC. Several other fragments have also had the cortile tissue removed, and are partially hollowed. A junction of tine and beam from feature 52 (65) includes a vertical scoring mark cut with a knife (Figure 9 B). This reflects the technique described by Ulbricht, under which sections of beam were scored on the outer surface and separated into quadrants with the aid of wedges (Ulbricht 1978, 27-30; Ambrosiani 1981, 112-3).

Table 3 Relative Quantities of Antler, by Category %

Unused raw material

33

42.85

Waste material

41

53.25

3

3.9

Semi-manufactures

THE UNFINISHED OBJECTS The antler waste was distributed across the site, but there were concentrations in features 52, 95 and 96, as noted above (Tables 1 and 2). Sections of antler from features 95, 96 and 97 provide an indication of the products manufactured on the site. Feature 95 produced a blank for a tooth segment (75) and a fragmentary connecting plate for a comb (Figure 11 A and C). Both have been cut to shape using a saw and a draw-knife, but neither has progressed any further. The connecting plate (76) shows no signs of any rivet holes and

In Table 3 the relative totals of waste material are grouped by number according to the system of Christophersen (1980, table 1 and fig 3). The totals recall those obtained for a deposit from Lund, in terms of the ranking of categories. In both cases, waste material exceeds unused raw material, which itself exceeds the quantity of unfinished objects present.

A

Number

B

C

D

Figure 11. Unfinished Antler Objects from Holy Rood: A Unfinished Tooth Segment blank (75); B Antler strip with decoration at one end (77); C Unfinished comb connecting plate (76); D undecorated broad antler strip (48) 72

I. Riddler and N. Trzaska-Nartowski: Late Saxon Worked Antler Waste from Holy Rood, Southampton

it had not, therefore, reached the stage where it would be fastened to another connecting plate, shaped and decorated. The tooth segment also shows no indication of riveting and it does not belong to an assembled comb.

noted above, the tines were removed by sawing cleanly, usually in a single direction. Remnant saw traces indicate that blades of 1.2 – 1.3mm were used to separate the burr and the beam, whilst thinner blades of 0.8 – 1.0mm were utilised on the tines and in the subsequent dismemberment of the material. These measurements lie well within the ranges established for antler from various sites (Table 4).

Combs were not the only product, however. Three further pieces of antler (47, 48 and 77) provide an indication of the production of another type of object. Each has been split from the beam or a tine and trimmed with a draw-knife to provide long, rectangular section of antler, which is lightly curved in section (Figure 11 D). The end of one piece (77) retains vestiges of decoration, undertaken with a saw blade 0.3mm in width. A single ring-and-dot motif lies between paired crossing diagonal lines (Figure 11 C). These sections of antler are too broad and slender in section to have been used with combs. They have not been riveted and they show no signs of saw marks from the cutting of comb teeth. The smoothing of segments of both beam and crown tine that is visible on offcuts from features 52 and 96 reflects an interest in providing long, broad and thin plates of antler. It is possible that they were cut and shaped to form the sides of comb cases. Comb cases, however, are quite rare, in comparison with the number of combs of the period (MacGregor 1985, 96-8; Riddler, Trzaska-Nartowski and Hatton forthcoming). It is more likely that they stem from casket mounts, which were produced in various materials, including antler, during the Anglo-Saxon period (MacGregor 1985, 197-200; 2000, 1602; MacGregor, Mainman and Rogers 1999, 1954-9). Although most caskets include bone mounts, antler and whalebone were also used as raw materials in their manufacture.

Antler wedges were used to split sections of beam and tine into flat segments with – in one case at least (Figure 9 B) – saw marks used to define the lines of separation. A knife was used to smooth some areas of antler, and to incise surfaces, most notably on the burr from context 95W (Figure 9 A). A draw-knife was almost certainly used to smooth the antler surfaces. The partial hollowing of sections of antler is quite common within the assemblage and, as at Middle Saxon London, this is probably a reflection of the use of an iron tool to hold the antler whilst it was being worked (Malcolm, Bowsher and Cowie 2003, 172). In most cases the antler is lightly indented at the centre, but several fragments have been completely hollowed.

A BROADER CONTEXT As noted above, copious quantities of red deer antler and worked bone have come from eighth and ninth century contexts at nearby Hamwic. Small quantities of roe deer antler have also been recovered there (see Riddler above). Late Saxon worked antler and bone material from Southampton is, as yet, much more sparse. MacGregor has published a brooch mould from late Saxon Southampton, which can be compared with examples from Haithabu and Ipswich (MacGregor 1980; Ulbricht 1978, 756 and taf 36; Newman 1993; West 1998, 68 and fig 97.7; Riddler, Trzaska-Nartowski and Hatton forthcoming). In addition, a small quantity of antler waste, amounting to just a few pieces, has come from excavations at Lower High Street and there is also a small assemblage of late Saxon date from West Quay.

THE TOOLS OF THE ANTLER WORKER The antler assemblage provides useful indirect evidence for the tools used by the craftsman. The two skull-attached burrs have been separated from the skull with the aid of an axe. In one case the fragment retains part of the skull and in the other the pedicle has been hacked away below the burr. With all of the burrs, the beam and brow tine were removed by sawing, usually from several directions. The same method can be seen with a few of the crown fragments and it occurs wherever there was a broad expanse of antler to be sawn. In most cases the antler was sawn cleanly, but vestiges of cortile tissue remain in a few instances. They indicate that the antler was sawn from several directions and then snapped away, leaving a remnant stub at the centre. With one exception,

The dearth of antler waste from late Saxon Southampton stands in apparent contrast to the quantities from Hamwic and from some Continental sites, notably Wolin and Haithabu. The late Saxon period has rightly been portrayed as ‘perhaps the golden age of the bone and, more particularly, the antler worker’ (MacGregor 1985, 48) but the quantities of waste

Table 4 Saw Blade Widths from Various Sites Site

Range (mm)

Reference

Haithabu

0.1 – 2.6

Ulbricht 1978, 36

Ipswich

0.3 – 3.7

Riddler, Trzaska-Nartowski and Hatton forthcoming

London

1.0 – 3.0

Malcolm, Bowsher and Cowie 2003, 172

Gdansk

0.6 – 2.0

Hilczerówna 1961, 62

Southampton

0.8 – 1.3

Hamwic

0.4 – 2.4

Riddler and Trzaska-Nartowski forthcoming 73

Materials of Manufacture

from most late Saxon sites are, in fact, disappointingly low. Less than a dozen pieces of antler have come from extensive excavations at Thetford (Rogerson and Dallas 1984, 190 and 191-2; Dallas 1993, 188; Riddler forthcoming C). Post-cranial bones of red, roe and fallow deer were recovered from excavations at Winchester, but they were accompanied by only 24 pieces of antler (Biddle 1990, 255). Around 50 fragments of antler came from late Saxon Worcester (Carver 1980, 174 and 192-9). Double that amount of antler came from contexts mainly of late Saxon date at St. Peters Street at Northampton (Oakley 1979, 308 and fig 136; Harman and Shaw in Williams, Shaw and Denham 1985, 75). Two concentrations of antler could be identified, one of which was related to an excavated structure. 99 fragments of antler are recorded from late Saxon contexts within one site at Chester (Ward 1994, 92-3).

antler from most sites, and larger quantities from a few locations (Riddler forthcoming D). Contemporary deposits at York may also illustrate a similar situation. The quantity of antler from Coppergate, for example, can be contrasted with the few fragments retrieved from 22 Piccadilly (MacGregor, Mainman and Rogers 1999, tables 160 and 163). At both Hamwic and London deposits of bone and antler shavings are a good indication of the location of workshops (Riddler 1993, 115; Malcolm, Bowsher and Cowie 2003, 173). In the absence of a sampling system at Holy Rood, however, no shavings were recovered. The antler waste from Holy Rood lies close to the location of a late Saxon antler worker and represents part of a production episode which was concerned both with comb manufacture and with the production of broad, decorated antler strips, probably intended for caskets. From the quantity of waste recovered, this appears to have been a small scale enterprise, utilising just eight antlers; but it is difficult to tell whether all of the waste of this episode was retrieved, and greater quantities may lie nearby, in the vicinity of this deposit. No overall site plan survives of the excavations but the main features lay to one side and close to the edge of the site.

Greater quantities can, however, be seen at several sites. Around 300 fragments of antler came from Coppergate at York and nineteenth-century excavations nearby also revealed copious amounts of waste material (MacGregor, Mainman and Rogers 1999, 1872-3 and 1903-12). A little over 2,000 fragments of antler have come from Ipswich, the majority coming from contexts of ninth century or later date. A number of workshops can be identified within the town (Riddler, Trzaska-Nartowski and Hatton forthcoming). Further afield, at least two large assemblages of antler waste have come from excavations in Dublin (O’Riordain 1971, pl VIIb; National Museum 1973, 15-16 and pl 8).

CONCLUSIONS The assemblage of 77 fragments of antler from Holy Rood is a small one, but it contains several points of interest. Minor evidence for the working of red deer antler - and no other type of skeletal material - is present, and the central deposit of this assemblage may have lain a little to one side of the features identified here. Material from various stages of working procedures is present, including eight burrs, fragments of three crowns and three unfinished objects. The burrs appear to be a little larger than those encountered from Middle Saxon Hamwic, and the deposit itself differs a little from the bone and antler waste characteristic of that settlement. The absence of bone, as distinct from antler, would itself suggest that the deposit was late Saxon, and that the production episode occurred prior to the eleventh century, when bone began to be reinstated as the principal working medium.

The spatial distribution of bone and antler waste from the pre-Viking trading and production centres of eighth to ninthcentury date has been considered elsewhere (Riddler 2001; Riddler, Trzaska-Nartowski and Hatton forthcoming). Similar conditions may apply also to late Saxon settlements. The small quantities of antler waste recovered from many sites of late Saxon date may reflect the fact that the craft was carried out by a few specialists who produced a reasonably wide range of implements (dominated, however, by combs) from specific properties in a number of production ‘episodes’ (Riddler forthcoming A). On some occasions excavations have been centred on these properties and abundant quantities of waste have been recovered from the structures and their yards. The remaining waste, of the type and quantity seen here, may represent the vestiges of material deposited close to the working area, but slightly peripheral to it. If the specific properties of antler and boneworkers are not excavated, little evidence for the craft is forthcoming.

Bibliography AMBROSIANI, K., 1981, Viking Age Combs, Comb Making and Comb Makers in the Light of Finds from Birka and Ribe, Stockholm Studies in Archaeology 2, Stockholm

Nearby Hamwic provides a good illustration of this point. Bone, antler and whalebone waste from the Melbourne Street sites amounted to just a few fragments and very little came from excavations further to the south at Cook Street (Bourdillon and Coy 1980, 113-4; Garner 1993, 113-5; 2001, 187). Nearby, however, considerable quantities of worked bone were found at Chapel Street and much greater amounts have subsequently been recovered from some of the properties within the Six Dials excavations (Driver 1984; Riddler 2001; Morton 1992, 182-4; Andrews 1997, 227-31; see also Bourdillon above). Middle Saxon London also appears to follow a similar pattern, with small deposits of

ANDREWS, P., 1997, Excavations at Hamwic. Volume 2: Excavations at Six Dials, CBA Research Report 109, London BECKER, C., 1989, Die Geweihfunde vom Spandauer Burgwall, in A. von Müller and K. von Müler-Muči, Ausgrabungen, Funde und naturwissenschaftliche Untersuchungen auf dem Burgwall in Berlin-Spandau, Berlin, 101-274 BIDDLE, M., 1990, Object and Economy in Medieval Winchester, Winchester Studies 7ii, Oxford BOURDILLON, J. and COY, J., 1980, The Animal Bones, in P. Holdsworth, editor, Excavations at Melbourne Street, 74

I. Riddler and N. Trzaska-Nartowski: Late Saxon Worked Antler Waste from Holy Rood, Southampton Southampton, 1971-1976, CBA Research Report 33, London, 79118

O’RIORDAIN, B., 1971, Excavations at High Street and Winetavern Street, Dublin, Medieval Archaeology 15, 73-85

CARVER, M. O. H., 1980, Medieval Worcester, An Archaeological Framework, Transactions of the Worcestershire Archaeological Society 7

OAKLEY, G. E., 1979, The Worked Bone, in J. Williams, St. Peter’s Street, Northampton. Excavations 1973-1976, Northampton Archaeological Monograph 2, Northampton, 308-17

CHRISTOPHERSEN, A., 1980B, Raw material, Resources and Production Capacity in early Medieval Comb Manufacture in Lund, Meddelanden fran Lunds Universitets Historiska Museum New Series 3, 150-65

PUTMAN, R., 1988, The Natural History of Deer, London

CNOTLIWY, E., 1956, Z badań nad rzemiosłem, zajmującym się obróbką roga I kościna Pomorzu Zachodnim we wczesnym średniowieczu, Materiały Zachodnio-Pomorskie 2, 151-79

REICHSTEIN, H., 1991, Die wildelebenden Säugetiere von Haithabu (Ausgrabungen 1966-1969 und 1979-1980), Berichte über die Ausgrabungen in Haithabu 30, Neumünster

CNOTLIWY, E., 1958, Wczesnośredniowieczne przedmióty z rogu I kości z Wolina, ze stanowiska 4, Materiały ZachodnioPomorskie 4, 155-240

RIDDLER, I. D., 1992, Boneworking and pre-Viking Trading Centres, in Art and Symbolism. Medieval Europe 1992, Preprinted Papers 7, York, 149-156

CNOTLIWY, E., 1970, Pracownie grzebiennicze na Srebrnym Wzgorzu w Wolinie, Materiały Zachodnio-Pomorskie 16, 20987

RIDDLER, I. D., 1993, Worked Bone, in M. F. Garner, Middle Saxon Evidence at Cook Street, Southampton (SOU 254), Proceedings of the Hampshire Field Club and Archaeology Society 49, 113-6

REICHSTEIN, H., 1969, Untersuchungen an Geweihresten des Rothirsches, Berichte über die Ausgrabungen in Haithabu 2, Neumünster, 57-71

CNOTLIWY, E., 1973, Rzemioslo rogownicze na Pomorzu wczesnosredniowiecznym, Gdansk

RIDDLER, I. D., 2001, The Spatial Organisation of Bone-working at Hamwic, in D. Hill and B. Cowie, editors, Wics: the Early Medieval Trading Centres of Northern Europe, Sheffield, 61-6

DALLAS, C., 1993, Excavations in Thetford by B. K. Davison between 1964 and 1970, East Anglian Archaeology 62, Gressenhall

RIDDLER, I. D., forthcoming A, The Technology of the Working of Bone, Antler, Ivory and Horn, in J. Bayley, editor, Archaeotechnology, English Heritage Archaeological Reports, London

DRIVER, J. C., 1984, Zooarchaeological Analysis of Raw-Material Selection by a Saxon Artisan, Journal of Field Archaeology 2, 397-403 GARNER, M. F., 1993, Middle Saxon Evidence at Cook Street, Southampton (SOU 254), Proceedings of the Hampshire Field Club and Archaeological Society 49, 77-127

RIDDLER, I. D., forthcoming B, Antler and Boneworking in Northern Europe, AD 300 – 1200, London

GARNER, M. F., 1997, Further Middle Saxon Evidence at Cook Street, Southampton (SOU 567), Proceedings of the Hampshire Field Club and Archaeological Society 52, 77-87

RIDDLER, I. D., forthcoming C, Objects and Waste of Bone and Antler, in H. Mills, Excavations at Mill Lane, Thetford, East Anglian Archaeology, Gressenhall

GARNER, M. F., 2001, A Middle Saxon Cemetery at Cook Street, Southampton (SOU 823), Proceedings of the Hampshire Field Club and Archaeological Society 56, 170-91

RIDDLER, I. D., forthcoming D, The Small Finds, in G. Brown, Excavations at Exeter Street, London, PCA Monograph Series, London

HALPIN, A., 2000, The Port of Medieval Dublin, Dublin

RIDDLER, I. D., TRZASKA-NARTOWSKI, N. I. A. and HATTON, S., forthcoming, An Early Medieval Craft. Objects and Waste of Bone, Antler and Ivory from Ipswich Excavations, 1974-1994, East Anglian Archaeology, Gressenhall

HRUBÝ, V., 1957, Slovanské kostěné, předměty a jejich výroba na moravě, Památky Archeologické 48, 118-217 KAVÁNOVÁ, B., 1995, Knochen- und Geweihindustrie in Mikulčice, in F. Daim and L. Poláček, Studien zum Burgwall von Mikulčice. Band 1, Spisy Archeologického Ústavu av Čr Brno 2, Brno, 113-378

RIDDLER, I. D. and TRZASKA-NARTOWSKI, N. I. A., forthcoming, Middle Saxon Bone and Antler Working at Hamwic, Southampton Finds Volume 4, London

MACGREGOR, A., 1985, Bone, Antler, Ivory and Horn. The Technology of Skeletal Materials since the Roman Period, London

ROGERS, N. S. H., 1993, Anglian and other Finds from Fishergate, The Archaeology of York 17/9, London

MACGREGOR, A., 2000, Objects of Bone, Antler and Ivory, in P. Ellis, Ludgershall Castle. Excavations by Peter Addyman 19641972, Wiltshire Archaeological and Natural History Society Monograph 2, Salisbury, 160-8

ROGERSON, A., and Dallas, C., 1984, Excavations in Thetford, 1948-59 and 1973-80, East Anglian Archaeology 22, Gressenhall SHOKNECHT, U., 1978, Menzlin. Ein frühgeschichtlicher Handelsplatz an der Peene, Beiträge zur Ur- und Frühgeschichte der Bezirke Rostock, Schwerin und Neubrandenburg 10, Berlin

MACGREGOR, A., MAINMAN, A. and ROGERS, N. S. H., 1999, Bone, Antler, Ivory and Horn from Anglo-Scandinavian and Medieval York, The Archaeology of York 17/12, York

TEICHERT, M., 1983, Geweihreste aus der germanischen Siedlung bei Muhlberg, Kr. Gotha, Zeitschrift für Archaologie 17, 115-22 ULBRICHT, I., 1978, Die Geweihverarbeitung in Haithabu, Die Ausgrabungen in Haithabu Band 7, Neumünster

MALCOLM, G., BOWSHER, D. and COWIE, R., 2003, Middle Saxon London. Excavations at the Royal Opera House 1989-99, MoLAS Monograph 15, London

ULBRICHT, I., 1984, Die Verarbeitung von Knochen, Geweih und Horn im mittelalterlichen Schleswig, Ausgrabungen in Schleswig. Berichte und Studien 3, Neumünster

MORTON, A. D., 1992, Excavations at Hamwic, Volume 1: Excavations 1946-83, excluding Six Dials and Melbourne Street, CBA Research Report 84, London MULLER-USING, D., 1953, Über die frühmittelalterliche Geweihreste von Wollin, Saugtierkundliche Mitteilungen 1, 64-7

WARD, S., 1994, Excavations at Chester. Saxon Occupation with the Roman Fortress: Sites Excavated 1971-1981, Archaeological Service Excavation and Survey Reports 7, Chester

National Museum of Ireland, 1973, Viking and Medieval Dublin. National Museum Excavations, 1962-1973, Dublin

WEST, S. E., 1998, A Corpus of Anglo-Saxon Material from Suffolk, East Anglian Archaeology 84, Gressenhall

NEWMAN, J., 1993, Three Antler Moulds from Ipswich, Finds Research Group Datasheet 17, Norwich

WILLIAMS, J. H., SHAW, M. and DENHAM, V., 1985, Middle Saxon Palaces at Northampton, Northampton 75

Materials of Manufacture

76

Materials of Manufacture

76