Crafting Textiles: Tablet Weaving, Sprang, Lace and Other Techniques from the Bronze Age to the Early 17th Century 9781789257595, 9781789257601, 2021942856, 178925759X

Table of contents :
Contents
Acknowledgements
Introduction
List of contributors
In memoriam Peter Collingwood
Section I Tablet weaving
Chapter 1 ‘Tablet weaving is a small byway of textile production…’: Bronze and Iron Age tablet bands with stripes, meanders and triangles from the salt mines in Austria
Chapter 2 The use of weaving tablets in the production of headdresses in Egypt in the Roman and Byzantine periods: A study of a bourrelet from Antinoopolis
Chapter 3 Evidence of tablet weaving from Viking-age Dublin
Chapter 4 The so-called ‘Palermo bands’ and their technique
Section II Sprang
Chapter 5 Hairnets with gold tube beads from the Roman Rhineland and their textile technique
Chapter 6 Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt
Chapter 7 Tight-fitting clothing in antiquity and the Renaissance: Research and experimental reconstruction
Section III Braiding and lace making
Chapter 8 Braided strings and Turk’s head knots on European secular and religious textiles
Chapter 9 A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle, East Tyrol
Chapter 10 From narrow four-strand plaits to openwork bobbin-made braids and edgings
Section IV Spinning
Chapter 11 The story of twist: Handspinning as a medieval craft

Citation preview

ANCIENT TEXTILES SERIES 39

Crafting Textiles Tablet weaving, sprang, lace and other techniques from the Bronze Age to the early 17th century

Edited by

Frances Pritchard

Early Textiles Study Group Occasional Publication

Oxford & Philadelphia

Published in the United Kingdom in 2021 by OXBOW BOOKS The Old Music Hall, 106-108 Cowley Road, Oxford, OX4 1JE and in the United States by OXBOW BOOKS 1950 Lawrence Road, Havertown, PA 19083 © Oxbow Books and the individual contributors 2021 Paperback Edition: ISBN 978-1-78925-759-5 Digital Edition: ISBN 978-1-78925-760-1 (epub) A CIP record for this book is available from the British Library Library of Congress Control Number: 2021942856 All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording or by any information storage and retrieval system, without permission from the publisher in writing.

Printed in the United Kingdom by Short Run Press, Exeter For a complete list of Oxbow titles, please contact: UNITED STATES OF AMERICA UNITED KINGDOM Oxbow Books Oxbow Books Telephone (610) 853-9131, Fax (610) 853-9146 Telephone (01865) 241249 Email: [email protected] Email: [email protected] www.oxbowbooks.com www.casemateacademic.com/oxbow Oxbow Books is part of the Casemate Group

Front cover: Detail of bobbin-made braid, c. 1600, Ytterjärna church, Sweden. © Lena Dahrén Back cover: From top to bottom: Tablet-woven band from Hallstatt salt mine, c. 800−400 BC (© Natural History Museum Vienna, photo A. Rausch); Tablet-woven band from Dürrnberg salt mine, c. 350 BC (© Natural History Museum Vienna, photo A. Schumacher); Marble bust, The Cloisters Collection 66.25, Metropolitan Museum of Art (© Metropolitan Museum of Art); Modern sprang hairnet (© Anne Kwaspen)

Contents Acknowledgements����������������������������������������������������������������������������������������������������������������� v Introduction���������������������������������������������������������������������������������������������������������������������������vii List of contributors����������������������������������������������������������������������������������������������������������������ix Section I: Tablet weaving��������������������������������������������������������������������������������������������������� 1 1. ‘Tablet weaving is a small byway of textile production…’: Bronze and Iron Age tablet bands with stripes, meanders and triangles from the salt mines in Austria��������������������������������������������������������������������������������������������������������������������������� 3 Karina Grömer 2. The use of weaving tablets in the production of headdresses in Egypt in the Roman and Byzantine periods: A study of a bourrelet from Antinoopolis������������������������������������������������������������������������������������������������������������������� 23 Claire Gérentet-de-Saluneaux and Fabienne Médard 3. Evidence of tablet weaving from Viking-age Dublin����������������������������������������������� 37 Frances Pritchard 4. The so-called ‘Palermo bands’ and their technique������������������������������������������������ 53 Regula Schorta Section II: Sprang��������������������������������������������������������������������������������������������������������������� 71 5. Hairnets with gold tube beads from the Roman Rhineland and their textile technique������������������������������������������������������������������������������������������������������������������������� 73 Petra Linscheid 6. Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt�������������������������������������������������������������������������������������������������������������������������������� 85 Anne Kwaspen and Kristin South 7. Tight-fitting clothing in antiquity and the Renaissance: Research and experimental reconstruction ������������������������������������������������������������������������������������ 103 Dagmar Drinkler and Carol James

iv

Contents

Section III: Braiding and lace making������������������������������������������������������������������������129 8. Braided strings and Turk’s head knots on European secular and religious textiles��������������������������������������������������������������������������������������������������������������������������� 131 Joy Boutrup 9. A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle, East Tyrol���������������������������������������������������������������������������������������������������������� 147 Beatrix Nutz, Rachel Case and Carol James 10. From narrow four-strand plaits to openwork bobbin-made braids and edgings��������������������������������������������������������������������������������������������������������������������������� 175 Lena Dahrén Section IV: Spinning�������������������������������������������������������������������������������������������������������� 205 11. The story of twist: Handspinning as a medieval craft ����������������������������������������� 207 Katrin Kania

Acknowledgements This volume, which is a product of the Early Textiles Study Group, is the result of the input of many people, who have generously given hours of their time to it. Penelope Walton Rogers, current chairman of the Early Textiles Study Group, ably supported the publication process and due to her computer skills came to the rescue on many occasions. Michael Pollard ensured that all the images met the requirements of the publisher. Greta Bertram, Curator, Crafts Study Centre, very kindly supplied the image of Peter Collingwood during a very difficult period. Grateful thanks are due to the specialists who peer reviewed the individual papers. Grants from the Pasold Research Fund and the Abegg-Stiftung Foundation enabled a considerable number of colour images to be included in the volume making it a more worthwhile resource. Julie Gardiner and the team at Oxbow Books guided the volume through to publication. The greatest debt is owed to the authors of each chapter who have waited patiently to see their research in print.

Introduction This book is dedicated to Peter Collingwood, who was an inspiring weaver, teacher and author, and whose curiosity about textile structures led him to analyse ancient and historic textiles as well as creating new ones. An exceptional craftsman of the second half of the twentieth century, among his outstanding achievements was to invent a new type of loom and to take weaving techniques in fresh directions. His interest in ancient textiles is self-evident from his publications and often formed the starting point for his creativity. In the late 1980s he became a member of the Early Textiles Study Group and actively contributed to its mission. Not only did he lecture at one of the Group’s conferences held in Manchester in 1994 but he also invited members to join him at his home at Nayland, Essex, where he discussed items from his own collection gathered from all parts of the world with considerable enthusiasm and technical insight. His books have been mainstays of much textile research as demonstrated by the range of papers included in this volume. By writing on techniques such as sprang and ply-split braiding he brought a fresh impetus to these subjects which have been taken up creatively as well as aiding a better understanding of ancient methods. The first of his series of books on techniques was rug weaving and although the study of early rug-making attracts considerable interest and archaeological material has been explored to some extent, it has not proved possible to obtain a paper containing new research on the subject here. Sprang was his next choice of subject for investigation. This was not a technique that he particularly relished doing himself but his book provided a sound basis for analysing many aspects of the technique, which he developed alongside the noted Swiss scholar Noémi Speiser, and it has acted as a stimulus to all subsequent research in this field. The Techniques of Tablet Weaving first published by Faber and Faber in 1982 became a best seller and led the way to much experimentation and creative work as well as providing a guide to reconstruction attempts. This in turn has led to further research and has enabled the subject to be studied in much greater depth. This publication demonstrates the wealth of new research carried out on an international scale into subjects in which Collingwood took such an active interest. Some discuss recently excavated material while others investigate textiles that have been stored for many years in museum collections and offer new interpretations and insights informed by technical expertise. Early versions of eight of the chapters were presented at an Early Textiles Study Group conference held at the Wellcome Collection, London, in October 2014. To these have been added research

viii

Introduction

by Claire Gérentet-de-Saluneaux and Fabienne Médard in France, Petra Linscheid in Germany, and Anne Kwaspen and Kristin South in Egypt. It is a great sadness that Collingwood is no longer here for us to share this volume with him but his legacy in the field of early textiles continues to act as inspiration to textile scholars and craft practitioners.

List of contributors Joy Boutrup

Design School Kolding Kolding Denmark

Rachel Case

Independent scholar Hampden USA

Lena Dahrén

Independent scholar Stockholm Sweden

Dagmar Drinkler

Anne Kwaspen

Marie S. Curie Fellow Centre of Textile Research Copenhagen Denmark

Petra Linscheid

Research Assistant Institute for Archaeology & Anthropology University of Bonn Germany

Fabienne Médard

Independent scholar France

Textile Conservator Bayerisches National Museum Munich Germany

Beatrix Nutz

Claire Gérentet-de-Saluneaux

Frances Pritchard

Independent scholar France

Karina Grömer

Curator Prehistoric Department Natural History Museum Vienna, Austria

Carol James

Independent scholar Winnipeg Canada

Katrin Kania

Independent scholar Erlangen Germany

Independent scholar Schwaz Austria Honorary Research Fellow The University of Manchester UK

Regula Schorta

Director The Abegg-Stiftung Riggisberg Switzerland

Kristin South

BYU Egypt Excavation Project Brigham Young University Utah USA

In memoriam Peter Collingwood

Peter Collingwood (1922–2008). Image kindly provided by the Crafts Study Centre, University for the Creative Arts. IMG2021−020.

Section I Tablet weaving

Chapter 1 ‘Tablet weaving is a small byway of textile production…’: Bronze and Iron Age tablet bands with stripes, meanders and triangles from the salt mines in Austria Karina Grömer Introduction In 1982, Peter Collingwood wrote one of the fundamental books about the technique of tablet weaving. He explained all known techniques, exploiting historic and ethnographic sources. A deep insight into practical work and the creative process of tablet weaving allowed him to shed a fascinating light on this old technique, which makes it possible to produce many complex weave structures and patterns. Collingwood’s chronological list of tablet weaving before AD 1000 starts at the sixth century BC. From Central Europe he only mentions the Hohmichele find, describing it as ‘A woollen weft-face fabric found in the Hohmichele tumulus, Germany (a royal burial ground of the nearby Heuneburg hill fort), with a starting border consisting of six cords of four-strand warp twisting’ (Collingwood 1982, 12). The splendid textiles in the princely tomb in Eberdingen-Hochdorf, Germany (Banck-Burgess 1999) or those from Verucchio, Italy (von Eles 2002; Ræder Knudsen 2012), had not been published at this time. To expand the timespan here are added some fascinating new finds, originating from the salt mines at Hallstatt and Dürrnberg in Austria, dating between 1600 and 400 BC.

Hallstatt and Dürrnberg in Austria, two textile hotspots The salt mines at Hallstatt and Dürrnberg provide excellent preservation conditions for organic finds. The prehistoric underground mining activities resulted in precious layers of waste. While working in the mine, textiles as well as other organic finds were left behind after their use. In a salty environment, single-celled bacteria dry out and die (Fig. 1.1). This prevents the decomposition process of organic materials due to bacterial activity. Unlike other environments such as bogs or waterlogged contexts, the salt preserves everything without exception. Therefore, artefacts made from

4

Karina Grömer

animal as well as of plant origin are present. The textiles are preserved so well in their organic matter that they are still colourful, elastic and supple when recovered (Grömer 2016, 24−25). Hallstatt is famous for a huge cemetery with more than 1500 graves (Kern et al. 2009). The rich artefacts found in the cemetery inspired the researchers to name the Early Iron Age in Central Europe (800−400 BC) after this site ‘Hallstatt Period’. The graves enable the social hierarchy of this period to be traced back. The style and decoration of the grave goods found in the cemetery are distinctive, and artefacts made in this style are widespread in Europe. The community at Hallstatt exploited the salt mines in the area, providing them with wealth (Reschreiter 2013). Mining activities can be proved at least from the Bronze Age (1500−1200 BC). The dating of the Hallstatt salt mine is proved by hundreds of dendrochronology dates and radiocarbon dating from various artefacts,

Fig. 1.1. Salt rock from Hallstatt with enclosed textile fragment, 10.4 × 8.5 cm, Natural History Museum Vienna (Inv. No. 73.180). (© Natural History Museum Vienna, photo: A. Rausch)

1.  ‘Tablet weaving is a small byway of textile production…’

5

including textiles. The Bronze Age and Early Iron Age mining activities are archaeologically of specific importance. Among the organic finds from the salt mine are e.g. hats and shoes of leather and skin, bags for hauling the salt and wooden objects such as tool handles, wooden vessels and a spoon for cooking in the mine that still retains residues of the meal cooked. From Hallstatt, more than 700 single fragments of textiles are preserved and it is thus possible to study textile techniques from 1500−400 BC by examining them (Grömer et al. 2013). The fabrics from the salt mines are mainly made from wool, and some are of extraordinarily fine quality in tabby, basket weave and various twill variants. The Hallstatt weaver was familiar with the dyeing of thread or finished fabric in colourful shades of blue, yellow, green, red and brown. Natural shades of sheep’s wool, from white through brown to black were also exploited in their designs. This fondness for using colour is particularly effective in striped and ‘plaid’ patterns, created using coloured warp and weft threads. Colourful bands were produced with a rigid heddle or with tablet weaving, the latter are of interest here. Dürrnberg near Hallein was a large economic centre and market and, especially between the sixth and the second century BC, it was one of the leading salt-producers in southern Central Europe (Stöllner 2002; 2005). Similar to Hallstatt, the artefacts in the Dürrnberg graves represent a wealthy population. Finds of gold and imported objects of Greek and Etruscan origin reflect the wide-distance trade of salt. The outstanding survival of rich burial grounds, of settlements and production areas present a complex picture of the society over a period of more than 500 years. Unlike Hallstatt, at Dürrnberg the Iron Age settlement is known with remains of the houses and artefacts demonstrating different handcrafts employed at the settlement, including textile tools (Brand 1995). The tools have been found all over the settlement areas and this points to extensive textile production. The textiles from the various underground findspots at the salt mine in Dürrnberg are about 800 in number, with most of them dating to the Late Hallstatt and the Early La Tène Period (Stöllner 2005). In Hallstatt individual patterns and weaving techniques dominate while at Dürrnberg more standardised fabrics are known. Fine and coarse tabbies dominate in fabrics made of hemp/flax or wool; twills are known in smaller proportions, as are repp bands or tablet-woven items. If decorated, Dürrnberg textiles are striped, with the exception of one outstanding textile patterned in ‘flying shuttle’ technique. Nearly 40% of the textiles show traces of dyeing, with the dyes still impressively preserved due to the salt as bright colours in red, blue or bluish-green.

Towards the beginnings of tablet weaving in the second millennium BC From the Bronze Age findspots at Hallstatt salt mine (c. 1600−1200 BC), a distinctive textile was found. HallTex 288 consists of two light brown tabby textiles, torn into strips and tied together (Grömer et al. 2013) (Fig. 1.2). To make this fabric, Z-spun

6

Karina Grömer

single yarn with a thread diameter of 0.4−0.5 mm was chosen, the thread count is 10−12 threads per cm in warp and weft. The technical data places this textile amongst the finer Bronze Age textiles from Hallstatt. Wool measurements describe this item as high quality wool, light and unpigmented (Rast-Eicher & Bender Jørgensen 2013, table 4, sample no. 10/135, sheep cat. AB) (Fig. 1.3). What makes HallTex 288 outstanding out of all the other Bronze Age textiles from Hallstatt is the starting border. It has a warp-twined appearance, which perhaps was made in a two-hole-tablet weave technique. The starting border has blue warp-twined (for warp-twined bands and their characteristics see Collingwood 1982, 112−32) stripes on light brown background, three times alternating one dark blue and one light brown line. This is the earliest evidence for a colour-patterned, tablet-woven band in Central Europe. The Hallstatt item was made at a time when sheep with light wool evolved, which made colour patterning possible. It is known that the beginnings of tablet weaving in Europe date back to Bronze Age. Heidemarie Farke re-assessed the textile assemblage of the Middle Bronze Age Schwarza graves in Germany and identified a tablet-woven starting border (Farke 1993). Fabric No. 13c from Tumulus C1 has a twined appearance, which is contemporary with, and comparable to, HallTex 288. Both items represent one of the basic techniques of tablet weaving, the so-called ‘warp twined tablet weaving’ (Collingwood 1982, 102−111). The ‘Ramses Belt’ from Egypt, c. 1200 BC, and three linen bands dating to the twenty-second Dynasty (945–745 BC) have often been considered the oldest evidence of tablet weaving but this has been refuted by Peter Collingwood in his rigorous studies (Collingwood 1982, 10−11). Another hint for this weaving technique is the square bone tablet with holes in the corners, which was discovered in a Late Bronze Age layer (radiocarbon-dated to 1400–1075 BC) of the site Abri Mühltal I in Germany (Grote 1994, pl. 101−103). The weaving tablet was found in a context with other textile equipment, such as a polishing stone and a fragment of a spindle whorl.

Advanced techniques in the first half of the first millennium BC After the very early Bronze Age finds, which are relatively simply made, there is an increasing demand to add decoration in the Early Iron Age between 800 and 400 BC. From the Early Iron Age, at Hallstatt there are striped tablet-woven textiles made with warp-twined tablet weaving (Hundt 1960, 145−46; Grömer et al. 2013, HallTex 43 & 136). HallTex 136−1 is a tablet-woven border with a warp-stripe design, 0.7 cm wide. Horsehair is added to a wool weft thread. The pattern consists of tablet-woven warp-stripes made with tablets, each of which is threaded with a single colour: 1 tablet blue on each selvedge, at the inner part 2 tablets yellowish and 3 tablets reddish (Fig. 1.4). The Dürrnberg salt mines also yielded simple tablet-woven items, dated between the sixth and fourth century BC (Grömer & Stöllner 2011, 7.1.1). All of them are monochrome and woven with square tablets. They are made as narrow

1.  ‘Tablet weaving is a small byway of textile production…’

7

Fig. 1.2. Textile with tablet-woven starting border HallTex 288, salt mine Hallstatt, c. 1500−1200 BC, 15 × 7.5 cm, Natural History Museum Vienna (Inv. No. 120.050). (© Natural History Museum Vienna, photo: A. Rausch, graph: K. Grömer)

Fig. 1.3. Details of the starting border and fibres, Dino-Lite Digital Microscope, 30× and 250×, Natural History Museum Vienna (Inv. No. 120.050). (© Natural History Museum Vienna, microphotos: K. Grömer)

Fig. 1.4. Tablet-woven band HallTex 136 with warp-stripes, salt mine Hallstatt, c. 800−400 BC, 3.6 × 1.2  cm, Natural History Museum Vienna (Inv. No. 89.844b). (© Natural History Museum Vienna, photo: A. Rausch)

borders in simple warp twining and woven as an integral part of larger weaves. But even more complicated weaving techniques appeared at this point. How the patterning techniques in tablet weaving evolved is difficult to trace from the archaeological record; it is not yet clear which pattern type was an evolutionary basis for the other. Next to the ‘warp-twined’ tablet-weaving technique, the second basic technique is to reverse the twining direction of the cords. This can be done with all cords or

8

Karina Grömer

only in some cords (Collingwood 1982, 112−34). If single tablets are twisted in special frequencies, a ‘double faced weave with diagonal structure’ appears as known from Hallstatt where four-hole tablets threaded with two light and two dark coloured threads were used to produce patterns by being turned groupwise in different directions. This technique can be used to create intricate patterns. Two examples can be identified from Hallstatt. They have a light pattern on a darker background in different hues. HallTex 123 was made with 21 tablets, the warp consists of fine, well-twisted wool yarn, which is plied, in blue-green, brownish, olive-green and yellowish (Fig. 1.5). Dark pigmented horsehair was used as weft thread to strengthen the fabric, as it was a trimming for a sleeve. The pattern consists of yellowish meanders and cross-filled triangles. The background is blue-green and brownish, the sides of the band were made with monochrome threaded tablets in blue-green and black, which serve as a perfect frame to highlight the pattern. HallTex 186 consists of two fragments of a tablet-woven ribbon with a pattern of triangles (Fig. 1.6). It was made with 13 four-holed tablets. Plied wool was used in both warp and weft. Like the other band, the sides of the band are made with monochrome threaded tablets, in this instance dark green and dark brown. The background of the decoration zone is light olive-green and dark brown. The pattern consists of yellow

Fig. 1.5. Tablet-woven band HallTex 123 with meander and triangle design, salt mine Hallstatt, c. 800−400 BC, 1.3 cm wide, circumference 22 cm, Natural History Museum Vienna (Inv. No. 89.832). (© Natural History Museum Vienna, photo: A. Rausch)

1.  ‘Tablet weaving is a small byway of textile production…’

9

triangles filled with a twirl. This pattern reflects typical Hallstatt design principles, which can even be found on contemporary pottery (Schappelwein 1999). The fineness of the bands is extraordinary; both were made with fine plied yarns (0.2−0.4 mm diameter) used as warp and the thread count is 64 threads per cm (HallTex 123) and 52 threads per 9 mm (HallTex 186). There is also a tablet-woven band from Hallstatt patterned in a two-hole technique, HallTex 152 (Grömer et al. 2013, 483; Karisto & Grömer 2017) (Fig. 1.7).

Fig. 1.6. Tablet-woven band HallTex 186 with twirl-filled triangles, salt mine Hallstatt, c. 800−400 BC, 0.9 × 8.4 and 5.1 cm, Natural History Museum Vienna (Inv. No. 90.186). (© Natural History Museum Vienna, photo: A. Rausch)

Fig. 1.7. Tablet-woven band HallTex 152, two-hole technique, salt mine Hallstatt, c. 800−400 BC, 12.5 × 1.2 cm, Natural History Museum Vienna (Inv. No. 89.870). (© Natural History Museum Vienna, photo: A. Rausch)

10

Karina Grömer

It is a 1.2 cm wide tablet-woven border with a coloured lozenge pattern, which is sewn to the remains of a yellowish fabric. Here also the plied yarn has a diameter of only 0.3−0.5 mm and the thread count is 44 threads per cm. It was woven with 12 four-holed tablets in a two-hole technique. This technique was very popular to make complex patterns in the Early Iron Age. The Hallstatt band has a pattern-field showing yellow rhombs with a crossed filling on an olive-green and brownish-black background. The pattern-field is surrounded by one tablet with brownish-black threads on one side and a brownish-black tablet followed by two tablets with olive-green threads on the other side. To put this into context, it is worth noting that besides the bands described from Hallstatt, other archaeological sites from the Early Iron Age have provided tabletwoven bands. In addition to the techniques carried out with four-hole tablets, finds from Sasso di Furbara in Italy point to the use of triangular three-hole tablets (Mamez & Masurel 1992). The items belong to the Villanovan culture with a dating of ninth to eighth century BC. But most of the tablet weaves from other sites are made with four-hole tablets (Grömer & Stöllner 2011). Usually they are grave finds, attached to metal objects and, therefore, in most cases the original colour has unfortunately not been preserved and only the binding structure can be reconstructed from these pieces. At Apremont in France, the recognisable changes in the rotation patterns were most likely not only structural patterns with uni-coloured yarns, but perhaps also included coloured warp threads to form a colourful pattern that is lost from view today (Banck-Burgess 1999, figs 40−41). Outstanding are the woven-on tablet borders from Verucchio, Italy, dating between 725 and 650 BC (von Eles 2002; Stauffer 2012), which have structural patterns, e.g. Cloak 2 from Verucchio in Italy has a wide tablet-woven border with a triangular pattern produced by changing the turning direction of the tablets. In addition, it has stripes formed by tablets turned in the opposite direction (Ræder-Knudsen 2012, figs 11.2−11.5). From Hohmichele (late sixth/fifth century BC) there is not only the simple tablet-woven band mentioned by Collingwood that was found, but also a more complex item with double-faced weave and a diagonal structure (Banck-Burgess 1999, fig. 39). Here also the changing warp twist directions indicate that those structures are the remains of a former colour pattern of significant complexity. The most famous examples in this technique are the finds from the princely tomb of Hochdorf (late sixth century BC), which will be discussed later.

Early Celtic craftsmanship Early Celtic craftsmanship starts at the end of the Hallstatt Period with the Hochdorf finds. The tablet-woven bands from Hochdorf show various types of patterning. A very wide tablet-woven band made in a combination of double-faced weave and double-faced 3/1 broken twill using only three threads per tablet was found on a very

1.  ‘Tablet weaving is a small byway of textile production…’

11

large cauldron at the princely tomb of Hochdorf, Germany. Narrower tablet-woven bands made with four-hole tablets using only two threads per tablet and patterned with swastikas and meander motifs have also been recorded (Ræder Knudsen 1999, 56−58). In the princely tomb of Hochdorf, tablet-woven bands were also found with extra weft decoration on the surface. Here again, a tablet-woven item from the Austrian salt mines can be mentioned (Dürrnberg, No. 4470), which is outstanding concerning its technique of manufacture. This item dates to c. 350 BC due to dendrochronology dating of this part of the mine (Stöllner 2005). The border from Dürrnberg is mounted on a sleeve edge, which consists of a wool, spin-patterned tabby. The entire band was made using 29 fourhole tablets (Fig. 1.8). At the edge towards the sleeve and in the middle section there are one and four tablets each holding four blue threads. At the outer edge, the border does not have edge tablets. In the two pattern sections there are 13 and 11 tablets. For the pattern zone, backgrounds of greenish-blue and dark brown colour were fashioned to form yellow meandering motives using intricate patterning. Thus it can be seen that tablet weaving enabled people to create more complex patterns. The patterns echo contemporary motifs on pottery and bronze artefacts as well and therefore represent the Zeitgeist of a design (Grömer & Stöllner 2011, 109−11). In addition, the threads used for this band are of good quality, 0.2−0.3 mm Z- and S-plied yarns. As figured out by Lise Ræder Knudsen, the band was woven in a previously unknown technique combining pattern weaving with three colours and one empty hole in each four-hole tablet (Grömer & Stöllner 2011; Ræder Knudsen & Grömer 2012). The pattern is of a light colour and the background changes colour. The tablet-weaving technique usually cannot produce a pattern like this, but the weaver overruled the laws of tablet weaving and just pressed down the undesired colour with the shuttle. On the back of this band a lot of tiny errors can be detected (Fig. 1.8 right), indicating that the production of this pattern caused difficulties. Another detail is visible: the weft is horsehair (Fig. 1.9). This would produce a stiffer band. This

Fig. 1.8. Salt mine Dürrnberg, tablet-woven band with meander decor, c. 350 BC, 6.5 × 4.6  cm, Bergbaumuseum Bochum (No. 4470). (Photos: A. Schumacher, NHM)

12

Karina Grömer

Fig. 1.9. Salt mine Dürrnberg No. 4470, SEM pictures of details, Bergbaumuseum Bochum (No. 4470). (SEM pictures: VIAS)

phenomenon is also known from some tablet-woven and repp bands from Hallstatt (Grömer et al. 2013, HallTex 123A & 136/1). The phenomenon of tablet braids with horsehair pattern weft occurs again in Scandinavia in the Migration Period (see e.g. Nockert 1991 on the Högom find in Sweden). A similar technique to the Dürrnberg band can be identified from a tabletwoven item from Altrier in Luxembourg dated to the Early La Tène Period, but the motif differs slightly: the decoration is a red swastika (dyed with kermes) and the background is blue (Rast-Eicher & Vanden Berghe 2015, 122, fig. 10.6).

Experiments on a band from Hallstatt Analysis of textiles can reveal key data behind the finds: yarn strength, weaving method and surface quality. In experimental textile archaeology, attempts are made to understand how the items were made. A basic tenet of this method is that several skilled people should repeat the experiment in order to achieve a result that is of statistical relevance (Mårtensson 2007, 97−106). The persons involved must have experience in the relevant handcraft and the process must be completed with precise documentation. An experiment on the complex tablet-woven band HallTex 123 will now be presented in detail – with all thoughts, implications and decisions. This was undertaken within the framework of the project HallTex FWF ‘Dyeing techniques of the prehistoric textiles from the salt mine of Hallstatt – analysis, experiments and inspiration for contemporary application’ (2008–2012) which was supported by the Austrian Science Fund FWF [L 431−G02]. Together with the original find, the reconstruction of the band was included in a special exhibition, Colours of Hallstatt, at the Natural History Museum Vienna in 2012 (Fig. 1.10), as well as the tools used, the documentation, time calculations and a film made

1.  ‘Tablet weaving is a small byway of textile production…’

13

Fig. 1.10. Colours of Hallstatt exhibition, Natural History Museum Vienna 2012. (© Natural History Museum Vienna)

during the weaving process (Hofmann-de Keijzer et al. 2012). With that we ‘bridge the gap between the craft and the museum world’, as it was expressed by Peter Collingwood (1982). To recap, HallTex 123 is a tablet-woven band made with 21 tablets resulting in a meander and cross-filled triangle pattern. For the warp threads plied, hard twisted yarn was used of only 0.1−0.2  mm diameter, for the weft horsehair was employed to make it stiffer. The reconstruction started with a selection of the fibre types and dyestuffs to be used. Analysis of the fibres allowed new insights (Rast-Eicher 2013). In Hallstatt not only were the textiles investigated but also skins in order to compare them with the textiles and with still existing primitive sheep types. The Hallstatt wool is comparable with primitive sheep types such as the Vrin sheep in Switzerland. For the reconstruction different kinds of primitive sheep were selected and the wool was processed to get a result comparable to the Hallstatt finds (Rösel-Mautendorfer et al. 2012). The wool in question was from the fleece of an indigenous breed of sheep, Montafoner Steinschaf. It consists of many thick strands of kempy wool and finer fibres from the undercoat. Whilst thick strands were reduced at the early picking and carding stages, the combing and the production of well done combed tops helped to come as close as possible to the original band HallTex 123. Spinning was a challenge in order to produce the 0.1−0.2 mm yarn required with a high twist angle. From this period, in Central Europe, various types of spindle whorls made of clay are known, proving that hand spindles were used for spinning (e.g. Grömer 2016, 81−85). Using a hand spindle, the fibres were turned quite tightly, in keeping with observations of the original fabric. Machine-spun commercial yarn is spun far more loosely and is thus more susceptible to breakage. The symmetrical rotation of the spindle’s geometry affects a high moment of inertia, causing fibres from the fleece to twist into a single thread. Then two threads were

14

Karina Grömer

Fig. 1.11. HallTex 123, yarns made from a well-prepared combed top, Dino-Lite Digital microscope, c. 800−400 BC, Natural History Museum Vienna (Inv. No. 89.832). (© Natural History Museum Vienna, image: K. Grömer)

plied to get stronger yarn. The requirements for yarns found with the original textiles from Hallstatt (Fig. 1.11), also reflect what Peter Collingwood indicates the warp yarn ideally should be: strong – to withstand the hard beating; smooth – to allow the shed to clear easily; resistant to rubbing – to withstand wear in the tablets’ holes; fairly tightly plied – so that when the twist is taken from it in weaving it does not become completely unplied (Collingwood 1982, 54). Washing removes the grease, which is essential if one intends to dye the wool. Washing can be performed on the fleece, the yarn and even on the sheep before shearing. In the experiments, the plied yarn was washed at around 50°C with an infusion of soapwort. Dyestuff analysis carried out on a wide range of textiles from Hallstatt indicated which colours should be selected (Hofmann-de Keijzer et al. 2013). In this case, indigotin, coming from woad could be detected for the blue colours. However, the now yellowish-olive pattern threads just gave a neat signal for any unidentified yellow containing plant. It was, therefore, decided to make the reconstruction with

1.  ‘Tablet weaving is a small byway of textile production…’

15

Fig. 1.12. HallTex 123, pattern drawing. (© Natural History Museum Vienna, image: K. Grömer)

woad and chamomile dyed threads with this and other dyeing being undertaken by Anna Hartl, a specialist for early plant use at the University of Natural Resources and Life Science Vienna (Universität für Bodenkultur Wien). Additionally, naturally brownish-black wool was used. Two shades of colour indicate the hues the original bands might have been. Additional experiments were also undertaken on colour changes as a result of contact with cuprous salt. The ancient remnants were interred in the salt mines for millennia and may have been lying in contact with bronze picks, whose copper content could be responsible for the olive-green discoloration of numerous Hallstatt textiles. Before the actual weaving, experiments with tablets made from different materials were carried out. According to archaeological and ethnographical evidence they should be clay, leather or wood (Collingwood 1982, 27, fig. 10). Weaving with the thin, hand-spun yarn was particularly challenging. The mechanical stress of the weaving technique is considerable, above all in tablet weaving. As the tablets are turned, the threads passing through the holes of the tablets are subject to quite severe rubbing. In the experiments, the threads withstood the strain well – which is probably due to extensive preparation of the fibres and care in the spinning process. The very smooth yarn of the Hallstatt tablet-woven bands indicates many deliberate and carefully

16

Karina Grömer

executed procedures in the preparation of the wool. The result testifies to the work of specialists: the quality achieved in the bands and the enormous dedication involved make these woven bands truly precious objects. The pattern of HallTex 123 was made with 21 four-holed tablets in ‘double faced weave with diagonal structure’. The threading is with two light and two dark threads, apart from the monochrome threaded tablets at the side. The direction of insertion of the threads is vital as well as the direction of rotation. A specific pattern of turning tablets forward and backward per row has to be employed (see Fig. 1.12); in total there are more than 70 different units (1 unit = turning each single tablet forward or backward and then insert the weft thread) until the pattern sequence repeats (Grömer 2005).

Memorisation of the pattern and time consumption It is a fact that the pattern of band HallTex 123 was made in a complicated system of rotating specific tablets. The combination of different directions of rotation gives an idea of the demands on concentration and three-dimensional understanding of the prehistoric weavers. It is not known how they could remember such difficult patterns. A modern weaver has sketches to check every single step and might correct mistakes. There is nothing like a drafting system for the complex Iron Age patterns, although perhaps some abstract illustrations on contemporary pottery might be interpreted in this light. Probably in the Iron Age, people kept the information about how the weaving was done into tales or songs and handed in that way the knowledge from one generation to the next. There is interesting ethnographic evidence from Asia. Anthony Tuck mentions: Fabrication processes of complex woven patterns require the commitment to memory of a substantial amount of numerical and colour-related information. Virtually any pattern or design that is incorporated into the weave of a textile can be reduced to numeric sequences, given the structure of warp and weft. Modern observation of traditional weavers in India and Central Asia suggests that this numerical information may have first emerged in the form of memorized, rhythmic chants that allowed the weavers to both remember patterns and reproduce them as frequently as required. Moreover, the linguistic and poetic associations between weaving and singing preserved in several Indo-European languages also suggest that these chants were, at some point, sources of rhythmic or possibly metrical narration in their own right (Tuck 2006, 539).

What else was learnt about the tablet-woven bands from Hallstatt? Archaeological experiments in the manufacture of textiles provided information on the steps involved and possibly their sequence, as well as time spent. Textile production was certainly one of the more time-consuming tasks in which prehistoric people were involved. For 1 m of fine patterned tablet-woven braid HallTex 123 (not counting dyeing) the following times are estimated: one hour to sort the fleece, one hour to tease the wool, five hours to card, 80 hours to spin and

1.  ‘Tablet weaving is a small byway of textile production…’

17

ply the yarn, five hours to wash and dry, 25 hours to weave: a total of 117 hours spent (Rösel-Mautendorfer & Grömer 2012). Additionally, the tablet-woven band HallTex 123 is a good example for indicating the high standard of sheep breeding and wool preparation. Light wool was used to obtain well-dyed threads. Especially for tablet weaving it is necessary to work with well-prepared raw material. The threads of the band represent well-combed wool with fibres lying parallel, spun into even threads. The result of all this are shiny threads with a flat surface to give lustre – a quality product from the Iron Age. The selection of tablet size and material is not visible in the end product. Clay tablets, however, were more difficult to handle, because they are thicker, do not have a surface as smooth as leather or wood. Also the plied yarns threaded into the holes of clay tablets tend to become rubbed. Reconstructions of other bands were also undertaken (Grömer 2005, 88−89). And there a different expenditure of time making those objects was noticed – what is not visible if you study those artefacts through the eyes of an archaeologist trained only in typologies. For the experiments of each pattern a 1.30 m long warp was prepared. The finished ribbon is about 90 cm long, for the weaving process shrinks the length a little and the last 30 cm cannot be woven due to the tablets being positioned there. Measurements on the expenditure of time were taken at several points along the ribbons in order to obtain average time statements. The reproduction showed that the time needed for HallTex 123 is three times longer than for the tablet-woven band HallTex 186 and six times longer than for HallTex 152. Maybe in there lies a key for the understanding of the value of the items.

Conclusion Since Peter Collingwood wrote his book about tablet weaving in 1982, excavations on prehistoric sites in Central Europe have brought to light more material. It is now known from well-preserved textile finds that this weaving technique goes back to the Middle Bronze Age at least. In the Bronze Age parts of the salt mines at Hallstatt, the oldest so far known tablet woven border with colour stripes was excavated, dating between the fifteenth and thirteenth century BC. The Iron Age areas of the salt mines Hallstatt (800−400 BC) provide some tablet-woven bands created in ‘double faced weave with diagonal structure’. This technique was used to create intricate patterns, like triangles, meanders and lozenges. These patterns woven in Iron Age bands also appear on contemporary ceramic and bronze objects, especially in Eastern Hallstatt Culture. A more complex technique was used to make a recently discovered band from Dürrnberg, dated to the first half of the third century BC. The band was made using tablets with four holes, threaded and turned in a delicate way. At the pattern sections with meandering motives three holes of the tablets are threaded with dyed wool; the pattern is woven with the individual turning of groups of tablets and of single tablets.

18

Karina Grömer

Furthermore there is a technical detail that is of interest: the tablet-woven bands from Iron Age Hallstatt, as well as the patterned item from Dürrnberg, were made separately, not as starting borders. Tablet weaving was also combined with other weaving techniques, such as weaving broader bands or on a warp-weighted loom (Collingwood 1982, 12; Grömer & Stöllner 2011). The simple warp-twined items from Dürrnberg are side-borders of wool bands. Additionally, the Bronze Age tablet-woven band from Hallstatt served as a starting border – similar to other early finds such as from Schwarza. The twined appearance, done in rows, is of interest for those early tablet-woven items from Schwarza and Hallstatt. Maybe this gives a hint as to how tablet weaving evolved. Attention can be drawn to older textile techniques as reflected by Neolithic twining techniques, which were also employed to make starting borders (Götze 1908; Crumbach 2014). Examples are well known from the Swiss lakeside settlements such as Hornstaad-Hörnle, Zürich-Kleiner Haffner or Auvernier-Port (Médard 2010, fig. 54). Can it be discussed, if twill and tablet weaving have been indigenous inventions in Europe, based on cultural knowledge of this region – the twisting and twining techniques in Neolithic tradition? Also in the Near East, Neolithic and Chalcolithic finds of weft-twining are known from sites such as Nehal Hemar, Tell Halula, Çatalhöyük and the Cave of the Warrior (e.g. Shamir 2015). In the foreword of his book, Collingwood wrote: ‘Tablet weaving is a small byway of textile production’. Maybe tablet weaving evolved hand in hand with other weaving techniques – from creating starting borders. Until now there is not much empiric evidence for this thesis, albeit the Central European Bronze Age finds from Hallstatt and Schwarza look like ‘missing links’. But from the Iron Age onwards, tablet weaving left the status of a ‘byway’ and began to represent its own expression of textile art – with specific possibilities of patterning, which are far beyond other contemporary weave types. In the Iron Age, decorative and interesting patterns were made, which served as braids for clothes, belts and as decoration for wall hangings in the grave. When the textiles from Hochdorf were found in the 1980s their complicated pattern was interpreted to derive from the Mediterranean (Hundt 1985, 115). However, recent analysis of the weaving techniques (Banck-Burgess 1999, 125), the raw material and also the patterns allow us to say, that the textile art known from the bands in Central Europe is also mirroring contemporary decorative design of pottery and bronze artefacts. Direct comparisons can be traced for example between the Hallstatt band HallTex 186 and a vessel from Leobendorf and Hoste (Griebl 1996). The meander pattern of the Dürrnberg band also has its counterpart in the decorative designs on metal objects.

Acknowledgements This research was undertaken in strong cooperation between the Natural History Museum Vienna, the Museum Hallein and Bergbaumuseum Bochum (Thomas Stöllner). Parts of the experiments have been carried out within the framework

1.  ‘Tablet weaving is a small byway of textile production…’

19

of the project HallTex FWF ‘Dyeing techniques of the prehistoric textiles from the salt mine of Hallstatt – analysis, experiments and inspiration for contemporary application’ (2008–2012) which was supported by the Austrian Science Fund FWF [L 431−G02] (Regina Hofmann-de Keijzer and Anna Hartl). My warmest thanks are for the anonymous reviewer who helped to make the text more clear and who added precious information.

Bibliography

Banck-Burgess, J. 1999. Hochdorf IV. Die Textilfunde aus dem späthallstattzeitlichen Fürstengrab von Eberdingen-Hochdorf (Kreis Ludwigburg). Forschungen und Berichte zur Vor- und Frühgeschichte in Baden-Württemberg 70 (Stuttgart, Theiss Verlag). Brand, C. 1995. Zur eisenzeitlichen Besiedlun des Dürrnberges bei Hallein. Internationale Archäologie 19 (Espelkamp, Verlag Marie Leidorf). Collingwood, P. 1982. The Techniques of Tablet Weaving (London, Faber and Faber). Crumbach, S. 2014. Mit dem Webstuhl in die Vorzeit! Textilforschung und Rekonstruktion textiler Techniken in der ersten Hälfte des 20. Jahrhunderts mit Ausblick auf die Folgen am Beispiel Brettchenweben. In Experimentelle Archäologie in Europa 13 (Bilanz), 194−203. Farke, H. 1993. Textile Reste von Dietzhausen und Schwarza. In R. Feustel, Zur bronzezeitlichen Hügelgräberkultur in Südthüringen, Alt-Thüringen 27, 53–123. Götze, A. 1908. Brettchenweberei im Altertum. Zeitschrift für Ethnologie 40, 482−500. Griebl, M. 1996. Siedlungsobjekte der Hallstattkultur aus Horn, N.Ö. In E. Jerem & A. Lippert (eds), Die Osthallstattkultur. Akten des Internationalen Symposiums Sopron 10−14. Mai 1994 (Budapest, Archaeolingua Publishers), 95–114. Grömer, K. 2005. Tablet woven ribbons from the prehistoric Salt-mines at Hallstatt, Austria. In P. Bichler, K. Grömer, A. Kern & R. Hofmann-de Keijzer (eds), ‘Hallstatt Textiles’ Technical Analysis, Scientific Investigation and Experiment on Iron Age Textiles, British Archaeological Reports International Series 1351 (Oxford, Archaeopress), 81–90. Grömer, K. 2016. The Prehistoric Art of Textile Making – The development of craft traditions and clothing in Central Europe, Veröffentlichungen der Prähistorischen Abteilung 5 (Vienna, Verlag des Naturhistorischen Museums Wien), chapter A2. Grömer, K., Kern, A., Reschreiter, H. & Rösel-Mautendorfer, H. 2013. Textiles from Hallstatt. Weaving Culture in Bronze and Iron Age Salt Mines. Textilien aus Hallstatt. Gewebte Kultur aus dem bronze-und eisenzeitlichen Salzbergwerk. Archaeolingua 29 (Budapest, Archaeolingua Publishers). Grömer, K. & Stöllner, Th. 2011. Ein abgerissener Ärmel aus dem Salzbergwerk Dürrnberg. Neue Erkenntnisse zur Brettchenwebtechnik in der Eisenzeit in Mitteleuropa. Jahrbuch RömischGermanisches Zentralmuseum Mainz 56 (Mainz 2011), 105–157. Grote, K. 1994. Die Abris im südlichen Leinebergland bei Göttingen. Veröffentlichungen der urgeschichtlichen Sammlung des Landesmuseums zu Hannover 43 (Hildesheim, August Lax Verlagsbuchhandlung). Hartl, A., van Bommel, M.R., Joosten, I., Hofmann-de Keijzer, R., Grömer, K., Rösel-Mautendorfer, H. & Reschreiter, H. 2015. Reproducing colourful woven bands from the Iron Age salt mine of Hallstatt in Austria. An interdisciplinary approach to acquire knowledge of prehistoric dyeing technology. Journal of Archaeological Science: Reports 2, 569−95. Hofmann-de Keijzer, R., Kern, A & Putz-Plecko, B. (eds), 2012. Colours of Hallstatt. Textiles Connecting Science and Art (Vienna, Verlag Naturhistorisches Museum Wien). Hofmann-de Keijzer, R., van Bommel, M.R., Joosten, I., Hartl, A., Proaño Gaibor, A.N., Heiss, G., Kralofsky, R., Erlach, R. & de Groot, S. 2013. The colours and dyeing techniques of prehistoric

20

Karina Grömer

textiles from the salt mines of Hallstatt. In K. Grömer, A. Kern, H. Reschreiter & H. RöselMautendorfer (eds), Textiles from Hallstatt. Woven culture from Bronze and Iron Age salt mines. Archaeolingua 29 (Budapest, Archaeolingua Publishers), 135–62. Hundt, H.-J. 1960. Vorgeschichtliche Gewebe aus dem Hallstätter Salzberg. Jahrbuch des RömischGermanischen Zentralmuseums 7 (Mainz), 145−46. Hundt, H.-J. 1985. Die Textilien im Grab von Hochdorf – hervorragende Zeugnisse frühen Kunsthandwerks. Katalog ‘Der Keltenfürst von Hochdorf’, Ausstellung in Köln 31.1. − 31.3.1986 (Stuttgart). Karisto, M. & Grömer, K. 2017. Different solutions for a simple design: new experiments on the tablet weave HallTex 152 from the salt mine Hallstatt. In EXAR Bilanz 2017, Experimentelle Archäologie in Europa, Jahrbuch 2017, Heft 16, 60−69. Kern, A., Kowarik, K., Rausch, A.W. & Reschreiter, H. 2009. Kingdom of Salt. Veröffentlichungen der Prähistorischen Abteilung (VPA) 3 (Vienna, Verlag des Naturhistorischen Museums Wien). Mamez, L. & Masurel, H. 1992. Étude complémentaire des vestiges textiles trouvés dans l’embarcation de la nécropole du Caolino à Sasso di Furbara, Origini 16, 295−310. Mårtensson, L. 2007. Investigating the function of Mediterranean Bronze Age. Textile tools using wool and flax fibres. Experimentelle Archäologie in Europa 6 (Bilanz), 97−106. Médard, F. 2010. L’art du tissage au Néolithique. IVe-IIIe millénaires avant J.-C. en Suisse (Paris, CNRS Èditions). Nockert, M. 1991. The Högom Find and Other Migration Period Textiles and Costumes in Scandinavia (Umeå, University of Umeå, Department of Archaeology). Ræder Knudsen, L. 1999. Technical description of fragments of the broad tablet wiven band found in the big cauldron of Eberdingen-Hochdorf. In J. Banck-Burgess, Hochdorf IV. Die Textilfunde aus dem späthallstattzeitlichen Fürstengrab von Eberdingen-Hochdorf (Kreis Ludwigburg). Forschungen und Berichte zur Vor- und Frühgeschichte in Baden-Württemberg 70 (Stuttgart, Theiss Verlag), 56−58. Ræder Knudsen, L. 2012. Case study: the tablet-woven borders of Verucchio. In M. Gleba & U. Mannering (eds), Textiles and Textile Production in Europe from Prehistory to AD 400. Ancient Textiles Series Vol. 11 (Oxford, Oxbow Books), 254–66. Ræder Knudsen, L. & Grömer, K. 2012. Discovery of a new tablet weaving technique from the Iron Age. Archaeological Textiles Review 54, 92–97. Rast-Eicher, A. 2013. The fibre quality of skins and textiles from the Hallstatt salt mines. In K. Grömer, A. Kern, H. Reschreiter & H. Rösel-Mautendorfer (eds), Textiles from Hallstatt. Weaving Culture in Bronze and Iron Age Salt Mines. Archaeolingua 29 (Budapest, Archaeolingua Publishers), 135–62. Rast-Eicher, A. & Bender Jørgensen, L. 2013. Sheep wool in Bronze Age and Iron Age Europe. Journal of Archaeological Science 40, 1224–41. Rast-Eicher, A. & Dietrich, A. 2015. Neolithische und bronzezeitliche Gewebe und Geflechte. Die Funde aus den Seeufersiedlungen im Kanton Zürich. Monographien de Kantonsarchäologie Zürich 46 (Zürich und Egg, Baudirektion Kanton Zürich). Rast-Eicher, A. & Vanden Berghe, I. 2015. Altrier (LUX): A fresh look at the textiles. In K. Grömer & F. Pritchard (eds), NESAT XII. The North European Symposium of Archaeological Textiles, 21st – 24th May 2014 in Hallstatt, Austria. Archaeolingua Main Series 33 (Budapest, Archaeolingua Publishers), 117−23. Reschreiter, H., 2013. Das Salzbergwerk Hallstatt – The prehistoric salt-mines at Hallstatt. In K. Grömer, A. Kern, H. Reschreiter & H. Rösel-Mautendorfer (eds), Textiles from Hallstatt. Weaving Culture in Bronze and Iron Age Salt Mines. Textilien aus Hallstatt. Gewebte Kultur aus dem bronze- und eisenzeitlichen Salzbergwerk. Archaeolingua 29 (Budapest, Archaeolingua Publishers), 12−32. Rösel-Mautendorfer, H. & Grömer, K. 2012. Spannende Experimente. In Broschüre zur Ausstellung, Hallstattfarben. Textile Verbindungen zwischen Forschung und Kunst. Ausstellung im Naturhistorischen Museum Wien 1. Februar 2012 bis 6. Jänner 2013 (Vienna, Verlag des Naturhistorischen Museums), 22−23.

1.  ‘Tablet weaving is a small byway of textile production…’

21

Rösel-Mautendorfer, H., Grömer, K. & Kania, K. 2012. Farbige Bänder aus dem prähistorischen Bergwerk von Hallstatt. Experimente zur Herstellung von Repliken, Schwerpunkt Faseraufbereitung und Spinnen. In Experimentelle Archäologie in Europa (Bilanz), 190−201. Shamir, O. 2015. Textiles from the Chalcolithic Period, Early and Middle Bronze Age in the Southern Levant – the continuation of splicing. Archaeological Textiles Review 57, 12−25. Schappelwein, C. 1999. Vom Dreieck zum Mäander. Untersuchungen zum Motivschatz der Kalenderbergkultur und angrenzender Regionen. Universitätsforschungen zum Prähistorischen Archäologie 61 (Vienna). Stauffer, A. 2012. Case study: the textiles from Verucchio, Italy. In M. Gleba & U. Mannering (eds), Textiles and Textile Production in Europe from Prehistory to AD 400. Ancient Textiles Series Vol. 11 (Oxford, Oxbow Books), 241−53. Stöllner, Th. 2002. Der prähistorische Salzbergbau am Dürrnberg bei Hallein II. Dürrnberg-Forschungen Bd 3 (Rahden/Westfalen, Verlag Marie Leidorf). Stöllner, Th. 2005. More than old rags – textiles from the Iron Age salt-mine at the Dürrnberg. In P. Bichler, K. Grömer, A. Kern & R. Hofmann-de Keijzer (eds), Hallstatt Textiles – Technical Analysis, Scientific Investigation and Experiments on Iron Age Textiles. British Archaeological Reports S1351 (Oxford, Archaeopress), 161–74. Tuck, A. 2006. Singing the rug: patterned textiles and the origins of Indo-European metrical poetry. American Journal of Archaeology 110, 539–50. Von Eles, P. (ed.) 2002. Guerriero e sacerdote. Autorità e nell`età der ferro a Verucchio. La Tomba der Trono (Florence, Edizioni all insegna dell Giglio).

Chapter 2 The use of weaving tablets in the production of headdresses in Egypt in the Roman and Byzantine periods: A study of a bourrelet from Antinoopolis Claire Gérentet-de-Saluneaux and Fabienne Médard

Introduction The study presented here has been undertaken on behalf of the Louvre within the framework of a collections review devoted to archaeological objects, material evidence and human remains from the cemeteries of Antinoopolis (Lintz & Coudert 2013). Following a multi-disciplinary study, the textile analysis focus was especially on a selection of unpublished headdresses, which had been collected for their novelty. The tubular borders of the headdresses (referred to here as bourrelets) particularly caught attention on account of their number and their aesthetic and technical diversity. One headdress, which in common with some others was made with weaving tablets, was singled out for detailed study. Conserved at the Musée Guimet, Lyon, the headdress is entirely made using this technique and it proved possible to discover the process from start to finish. Alongside analysis of the archaeological artefact, experimental work led to a realisation that the object was highly complex despite its apparent simplicity. Bourrelets It is not difficult to define the bourrelets: the literature provides more or less precise descriptions calling them ‘chenilles’ (literally ‘caterpillars’), ‘plaited textiles’, ‘corded textiles’ and ‘neck-bands’. Albert Gayet, the site director, described them as follows: ‘L’encoloure, assez curieuse, est formée par un gros bourrelet de laine. C’est tantôt une sorte d’énorme chenille de 10 à 12 cm de diamètre, tantôt une simple bande de tissue natté, terminé à ses deux extrémités par des franges, repliées en deux et cousues, après avoir, au préalable, éte rembourrée de morceaux d’éponge, de laine ou de crin’ (Gayet 1898, 11). More recently Florence Calament described them: ‘Le bourrelet de ‘tête’ ou de ‘tour de cou’ était réalisé

24

Claire Gérentet-de-Saluneaux and Fabienne Médard

en chenille de laine ou en tissue natté (formant des motifs de chevrons) ou côtelé, le plus souvent bicoloure (jaune et rouge ou rouge et vert)’ (Calament 2005, 286−87). A lack of technical knowledge of the objects and the difficulty of interpreting how they were worn explain the imprecise nature of the language. Their position on the bodies preserved in museums is, in fact, not always consistent with how they were found. Removal, transport, display, rearrangement and the disappearance of items of clothing have confused the interpretation of their function. Fortunately, numerous discoveries and pictorial evidence show that the bourrelets were intended to enhance women’s faces by edging cloaks, mantles, veils or headdresses. A funerary stele and a sculpture preserved in the Louvre (inv. nos. AF 11892 & E 12429) bear witness to this. Fig. 2.1. Albert Gayet showing the funeral This is further confirmed by some examples still clothes of Thaïs at the Musée Guimet (Paris, 1898). The early twentieth- attached to the remains of veils for instance one century reconstruction of the byzantine in the Louvre that is still attached to a fragment costume indicates the style of dress with of linen that may have belonged to the mantle the placement of the bourrelet on the of the so-called ‘Pagan Amazon of Antinoopolis’ (inv. no. AF 13312; Bénazeth 2011, 20), and also head (Santrot 2001) the one presented here. Finally, insofar as they are sometimes also associated with hairnets, some bourrelets are closer to a hair accessory than a clothing ornament (the Louvre, inv. nos. E 29486−29487, E 29482, E 32646 & E 31970?; Durand & Saragoza 2002, 127; Bénazeth 2011, 16) (Fig. 2.1). The bourrelets are complex objects made using various techniques, including looped pile, braiding and weaving, which allow for the production of numerous effects. The example studied here is particularly unusual because it was made using a technique that is usually associated with other types of products, namely weaving with tablets.

Archaeological research

Description The bourrelet was studied by the authors in 2012 at the Musée Guimet in Lyon before the collection was transferred to the Musée des Confluences. It was on the head of a mummy inventoried under the number 90002430 but otherwise known as ‘The Diounesast’. ‘The Diounesast’, or the royal favourite of Antinous, was discovered in 1906 during the eleventh season of Gayet’s excavations of the cemeteries of

2.  The use of weaving tablets in the production of headdresses in Egypt

25

Fig. 2.2. Female mummy, inv. 2430. Musée Guimet, Lyon. (Photo: F. Médard)

Antinoopolis. It was donated to the Musée Guimet in Lyon in 1907. ‘The woman had been buried in a very poorly preserved wooden coffin of which only a few fragments were brought back to France… The mummy was dressed in a long tunic reaching to just below the knees… Numerous traces of gilding were noted on all visible parts of the body, on the face, which appears to have been completely gilded, as well as on the legs and feet… An enormous bourrelet (…) is positioned today above the head without actually touching it. It is impossible to know whether this dress accessory really belonged to the trousseau which accompanied the deceased in death or if it came from another tomb’ (Lintz & Coudert 2013, 204) (Fig. 2.2). Slightly turned to the right, the deceased’s face is framed by the bi-coloured bourrelet, the two ends of which currently lie at shoulder level. The item is worked in alternating natural and red wool arranged in a checked pattern on eight columns and 48 rows, except for the ends, which are made in wool of uniformly natural colour. It measures 86 cm in length, 17 cm in circumference and 6 cm in diameter at the thickest point. Bourrelets such as this are sometimes padded, or sometimes made in such a way that the weave itself gives the object its volume (Bénazeth 2011, 16). This exemplar belongs to the second category, its volume and firmness are due to the thickness of the wool worked in looped pile (Fig. 2.3). Technical analysis Detailed examination of the inside of the bourrelet enabled the determination of technical data and the identification of the interplay of warp and weft threads by using tablets as weaving tools. Two different warps work simultaneously: the first in linen, called the ground warp, supports the second, which is made of thick strands of natural and red wool, which can be called the pile warp because the strands forming the regular loops are anchored in the ground. Made of two colours, the pile warp includes warp floats that are visible on the back of the bourrelet at the places where the looped pile appears on the front. The pile loops, 3 cm in height, are made from

26

Claire Gérentet-de-Saluneaux and Fabienne Médard

Fig. 2.3. Mummy, inv. 2430, Musée Guimet, Lyon. Detail of the bourrelet on the head. (Photos: F. Médard)

wool strands 3 cm in diameter. The work progresses vertically using simultaneously eight natural wool and eight red wool strands (Fig. 2.4). From a strictly technical point of view, the bourrelet is constructed on a twined base involving the use of eight tablets operating the warp threads, which are S2Z plied in natural linen (i.e. the ground warp) (Seiler-Baldinger 1991, 85). Each tablet is threaded with two threads in diagonally opposite holes making a total of 32 warp threads (since the warp threads are double). The tablets, which are all threaded in a S-direction, are given a half-turn at each step and a weft thread is inserted after every three turns creating the checked pattern formed by the alternation of the natural and red wool (Fig. 2.5). The reverse also reveals a fragment of an undyed textile, probably a vestige of the veil to which the bourrelet was sewn. The fabric, probably linen, is in tabby weave with a dominant warp or weft made using single S-spun yarns, 0.1−0.2  mm in diameter. The thread count is 33/18 threads per cm (Fig. 2.6). When removed from the ‘loom’, the long rectangle of piled textile is shaped into the bourrelet by folding it in half lengthways in order to join the side edges with a seam. The seam was stitched using an undyed linen thread, S2Z plied, loosely twisted and 1 mm in diameter (Fig. 2.7).

2.  The use of weaving tablets in the production of headdresses in Egypt

27

Fig. 2.4. Mummy, inv. 2430, Musée Guimet, Lyon. Detail of the wool strands forming the bourrelet. (Photo: F. Médard)

Fig. 2.5. Diagram of the bourrelet structure. (Drawing: F. Médard)

28

Claire Gérentet-de-Saluneaux and Fabienne Médard

Fig. 2.6. Mummy, inv. 2430, Musée Guimet, Lyon. Interior of the bourrelet. In its hollow is a fragment of cloth corresponding to the textile to which it was stitched. (Photo: F. Médard)

Fig. 2.7. Mummy, inv. 2430, Musée Guimet, Lyon. Left: sewing thread closing the seam. Right: the end of the bourrelet showing the finishing off of the warp threads. (Photos: F. Médard)

To complete the analysis, two samples were taken in order to establish the presence or absence of colorants. The first sample was taken from an area where the yellowish wool could have owed its colour to the use of a dyestuff or to stains caused by aging. The second sample was taken from an area where the red colour was evident. The analyses, undertaken by Witold Nowik at the Laboratory for Historic Monuments (Champs-sur-Marne, France), revealed natural undyed wool for the light yarn and wool dyed with madder for the red.

2.  The use of weaving tablets in the production of headdresses in Egypt

29

Comparisons It is impossible to claim to be exhaustive in terms of comparative material. Numerous bourrelets are reported from the excavations at Antinoopolis but there is no complete inventory. Examples similar to the one described here, made using looped wool pile incorporated during tablet weaving with a linen warp, exist in the collections of the Louvre (inv. E 31970 and E 32654; Bénazeth 2011, 20). There is at least one apparently similar example in the Musée de Tissus, Lyon (inv. no. 47556), and another in the Musée des Beaux Arts, Grenoble, belonging to the person known as ‘The Prophetess’. On the latter example, the tablet-weaving technique used to bind in the wool loops is clearly visible, even more so than on the headdress at Lyon. At least one other is held at the Musée Royal d’Art et d’Histoire, Brussels (inv. no. E 1045; Van Raemdonck et al. 2011, 228−29; Linscheid 2011, 304, pl. 58). It can be reasonably assumed that bourrelets chenillés mentioned in the literature also correspond to looped wool specimens (Durand & Saragoza 2002, 127; Calament 2005, 397, 399; Bénazeth 2011, 14, 20−32), although the term on its own is not sufficient to infer a weaving technique similar to that identified here.

Experimental reconstruction Due to the originality and technical novelties of the bourrelet, a tablet-weaving experiment, not previously explored, was undertaken. The intention was to fabricate it as similarly as possible to the original, taking into account the technique rather than the restoration. The aim was to test all the aspects of the item and to understand the difficulties and subtleties of making it. Preparation The search for an adequate raw material The first step was to obtain raw materials as close as possible to the original, especially wool corresponding to that used for the long strands. It was prepared without twist that was probably regulated by means of a perforated disc, called a diz. A preliminary warning is necessary: suitable wool is not available commercially. Therefore, Dominique Le Jean, a felt-worker, was consulted. After examining the mummy at the Musée Guimet, she determined that rather than the combed merino wool used in the first attempts which was too slippery and springy, the most appropriate contemporary material would be a slightly coarser wool derived from the hybrid breeds classified as ‘Préalpes’. Bought from the breeder, it does not contain kemp fibres and corresponds to a medium-grade wool with diameters of 28.3 microns and a fibre length of 6 cm. Consequently a beautiful combed quantity of ecru wool was provided by the Association pour la Promotion du Pastoralisme dans Alpes-Maritimes (APPAM). The thick strands are regulated by pulling the wool through the orifice of a diz, placed next to a comb. The raw wool is processed until the desired width and thickness

30

Claire Gérentet-de-Saluneaux and Fabienne Médard

have been obtained. The result is a long coiled strand around 13 cm in circumference, approximately 2.5−3 cm wide and 4 mm thick. Obtaining the correct colour In order to obtain the colour alternation comparable to the original pattern, part of the processed wool was entrusted to Karine Schreck, an artisan dyer. In accordance with result of the chemical analysis, madder was used to obtain a similar red to that of the headdress. This dye requires a raised temperature to be maintained for a certain length of time, which has the effect of causing a slight felting of the wool. The result was a noticeable difference in volume between the dyed and undyed wool. These two batches of contrasting coloured wool were then divided into strands of the same dimensions and density as the original. To obtain similar pile yarn, the undyed wool was divided into eight strands, while the red wool was only divided into four. Each strand was rolled between the palms of both hands to make finished strands, approximately 1 cm in width. The ground warp was made from natural linen similar to that of the original. However, the weft was cotton, a choice justified as it has a better ability to lock in position the loops of the pile warp. The weaver’s tools Square tablets of boxwood (Buxus) perforated at the corners and measuring between 38 and 46 mm along the sides were reproduced on the basis of archaeological examples from Antinoopolis held in the Louvre (inv. no. AF 1536; Collingwood 1982, 15). A rectangular weaving frame was made to measure to contain the full length of the bourrelet since its loops of pile prevent it from being rolled up on a beam. This lightweight frame was constructed because space was required underneath to hold the wool strands ready and to enable control of the reverse of the work. The use of two wooden formers, 15 mm in diameter, made it possible to create regular loops. They are the equivalent of ‘velvet irons’ used in the manufacture of velvets. Knitting needles were also used to ensure the rigidity necessary to maintain the counter-loops. Finally a weft-beater was required to tamp in the weft (Fig. 2.8).

Production of the copy Fig. 2.8. Tools and raw material used in the experimental reconstruction. (Photo: F. Médard)

Weaving process Following the archaeological example, the ground warp was threaded on

2.  The use of weaving tablets in the production of headdresses in Egypt

31

eight tablets. Each carried two linen threads placed in diagonally opposite holes, threaded in a S-direction and stretched on the frame. The warp threads from each tablet were knotted together 10 cm from the warp bar in order to stabilise the start of the weaving and, once the work was complete, to allow them to be brought together in a more substantial knot, which helps to create the shape of the bourrelet (Fig. 2.9). The wool strands that make up the pile warp cannot be stretched on account of their fragility. They are, therefore, left free to form the series of loops and counter-loops which give Fig. 2.9. The start of the reconstruction with the volume to the work. During weaving, warp mounted on tablets. (Photo: F. Médard) each strand of wool is wound over a cylindrical former, then left on the back of the band while the tablets are rotated to anchor it; the change of colour is the most delicate part of the operation (Fig. 2.10). Shaping and finishing Unlike a conventional velvet, an additional process must be implemented to ensure the solidity of the work. A light manual felting will serve to stabilise it, to ensure the cohesion of the counter-loops and to give the bourrelet its final appearance. The felting operation involved spraying the bourrelet with boiling soapy water to cause the scale structure of the wool fibres to open. It was then rolled with both hands flat, here between two layers of bubble wrap, but it would work equally well with fine basketry mats, to consolidate it completely (Fig. 2.11). This rolling action is applied particularly to the ends in order to obtain a similar shrinkage to the original and to ensure greater solidity. Wool felts very quickly, a few seconds are enough. For this reason throughout the process the dimensions of the bourrelet were controlled so that it did not become misshapen. Once the desired effect had been obtained, the item was rolled up in a towel to remove any dampness and to increase the felting. The operation lasts 30 minutes and the felting loses about a centimetre from the original length of the bourrelet. Once the felting finished, the ‘petals’ of colour were arranged in a wavy check pattern identical to that noted on the original. This is because the fibres are compressed in the centre by the warp binding point and the felting becomes

32

Claire Gérentet-de-Saluneaux and Fabienne Médard

Fig. 2.10. Working the tablets and wool strands to create the textile. Top left: the back; centre left: the front; bottom left: the back; right: weaving in progress. (Photos: F. Médard)

concentrated around this anchor point causing the fibres to undulate slightly (Fig. 2.12).

Conclusion

Technical elements The choice of the right wool for this task is essential in order to obtain a satisfying result. A useful project would be to research the characteristics of Egyptian livestock in the first centuries AD in order to be as accurate as possible in respect of the quality of the fibres and their impact on the final appearance of products.

2.  The use of weaving tablets in the production of headdresses in Egypt Felting, although one step among many, turned out to be of paramount importance and several observations were made in respect of it during the making of the copy. The dyed and undyed wool reacted differently to the felting: the red wool shrank less quickly than natural wool because dyeing had made the fibres less porous and therefore less susceptible to modification. This influenced the appearance of the final product: it was important to take it into account in order to anticipate and adjust any potential distortion. The locking of the counter-loops intuitively put in place during weaving proved to be very effective for the rest of the operation and, as such, may testify to a similar practice when the bourrelet was made in antiquity. Without felting, the loops come apart and the weaving breaks up, whereas felting provides support and cohesion to the work. At the end, the fixing threads were removed from the counter-loops and the latter remained perfectly in place locked into the weave. This second finding could be absolute proof of an original felting finish. It is the only known example to date that combines these two techniques. Other examples of tablet weaves with warp pile were made much later, for example the girdle from the tomb of Prince Sigismondo Pandolfo Malatesta in Rimini dating to the fifteenth century (Vial 1971) and from the nineteenth century doublefaced Central Asian bands, which is a topic that Peter Collingwood discussed in his book (Collingwood 1982, 352−61). More information about the Central Asian bands collected by Henri Moser

33

Fig. 2.11. Felting the bourrelet. Experimental work by Dominique Lejean. (Photo: F. Médard)

Fig. 2.12. Adjusting the felted loops. (Photo: F. Médard)

34

Claire Gérentet-de-Saluneaux and Fabienne Médard

Fig. 2.13. Comparison between the original bourrelet and the experimental reconstruction. (Photos: F. Médard)

2.  The use of weaving tablets in the production of headdresses in Egypt

35

in the 1880s that he donated to Bern Historical Museum is published in Lerber 1990 and four of these tablet-woven bands with pile are described by Marga and Heribert Joilet (Joilet & Joilet 1975, 126−27). Evaluation The experiment was undertaken on the basis of rigorous and accurate scientific observation. Nine weaving trials were necessary before achieving the right result. All the data has been scrupulously recorded so that the operation is reproducible. The result, in addition to its resemblance to the original model, presents similarities of detail, which lead one to think that, if the techniques and our ‘helping hands’ are not exactly the same as those of former times, they are not very far away (Fig. 2.13). The originality and creativity of this bourrelet is due to the technical combination of tablet weaving and felt. No previous instance of this kind appears to have survived. Thanks to these far-off artisans for demonstrating ‘technical innovation in the softness of wool’, it is now possible through experimental research to share the potential here.

Bibliography

Bénazeth, D. 2011. Accessoires vestimentaires dans la collection de textiles coptes du musée du Louvre. In A. De Moor & C. Fluck (eds), Dress accessories of the 1st millennium AD from Egypt, Proceedings of the 6th conference of the research group ‘Textiles from the Nile Valley’ Antwerp, 2−3 October 2009 (Tielt, Lannoo), 13−33. Calament, F. 2005. La révélation d’Antinoë par Albert Gayet. Histoire, archéologie, muséographie. Institut Francais d’Archéologie Orientale, Bibliothèque d’études coptes 18, 2 vol (Paris). Collingwood, P. 1982. The Techniques of Tablet Weaving (London, Faber & Faber). Durand, M. & Saragoza, F. (eds) 2002. Egypte, la trame de l’histoire. Catalogue d’exposition, Musée Départmental des Antiquités-Rouen, 19 October 2002−20 January 2003 (Paris, Somogy-Editions d’Art). Gayet, A. 1898. Catalogue des objets recueillis à Antinoé pendant les fouilles de 1898 et exposes au musée Guimet du 22 mai au 30 Juin 1898 (Paris, E. Leroux Editions, Paris). Joilet, M & Joilet, H. 1975. Brettchenweben (Bern, Haupt). Lerber, K. von 1990. Die Brettchengewebe der Sammlung Moser im Historischen Museum Bern, Schweizerische Arbeitslehrerinnen-Zeitung, 2/90 (Bern). Linscheid, P. 2011. Frühbyzantinische textile Kopfbedeckungen. Typologie, Verbreitung, Chronologie und soziologischer Kontext nach Originalfunden (Wiesbaden, Reichert Verlag). Lintz, Y. & Coudert, M. (eds) 2013. Antinoé, momies, textiles, céramiques et autres antiques. Histoire des collections du musée du Louvre (Paris, Editions d’Art Somogy). Santrot, J. (ed.) 2001. Au fil du Nil. Couleurs de l’Egypte chrétienne. Catalogue d’exposition, Musée Dobrée, 19 October 2001−20 January 2002 (Nantes, Somogy-Editions d’Art). Seiler-Baldinger, A. 1991. Systematik der textilien Techniken (Basel, Basler Beiträge zur Ethnologie). Van Raemdonck, M., Verhecken-Lammens, C. & De Jonghe, D. 2011. The mummy of the ‘embroideress’ and the contents of her grave. In A. De Moor & C. Fluck (eds), Dress accessories of the1st millennium AD from Egypt. Proceedings of the 6th conference of the research group ‘Textiles from the Nile Valley’ Antwerp, 2−3 October 2009 (Tielt, Lannoo), 222−35. Vial, G. 1971. Un ruban de velours tissé ‘aux cartons’ (provenant de la tombe de Malatesta). Bulletin du CIETA 34, 54−74.

Chapter 3 Evidence of tablet weaving from Viking-age Dublin Frances Pritchard

Introduction Excavations carried out by the National Museum of Ireland in the 1970s and early 1980s yielded a considerable amount of evidence relating to craft-working and everyday life in Dublin in the tenth and early eleventh centuries when the town was under Scandinavian rule and a thriving mercantile centre. Craft equipment, including small, square weaving tablets and short weaving battens recovered from several of the sites, indicates that bands were tablet woven in the neighbourhood. In addition, there are many wool bands woven in this technique, including examples applied as bindings to heavyweight wool cloths that were probably cloaks. Various gold and silver-brocaded silk bands also attest to the ostentatious wealth of some inhabitants in the town.

Background It is well known that tablet weaving was a popular method for making bands in the Viking period and that they have been found in excavations from sites, especially burials, throughout the Viking world. Indeed, Peter Collingwood in his inspirational book, The Techniques of Tablet Weaving, drew attention to the astonishing preservation of a brocaded band with a linen warp in the process of production from the Oseberg ship burial dating to the mid-ninth century (Collingwood 1982, 16−17, pl. 5). This find, which was recovered from a bed in the prow of the ship, is important for indicating that the technique was practised by those of high status as well as on a domestic level, a practice that continued for many centuries as demonstrated, for example, by Eleanor (c. 1161−1214), queen of Castile and daughter of Henry II of England, who in 1197 is reputed to have made a matching stole and maniple both interwoven with her name by means of tablets that are preserved in the Museo de Real Colegiato de San Isidoro, Léon (May 1957, 98; Jasperse 2017, 532−34).

38

Frances Pritchard

The technique of tablet weaving was already known in Ireland before the arrival of the Scandinavian invaders in the ninth century. Evidence in support of this in the early medieval period comes from two sites, Lagore Crannog in Co. Meath and a hill fort at Deer Park Farms, Co. Antrim. From the former site, which is considered to have been a native Irish royal residence, there are two tablet-woven textiles, one interpreted as an experimental band made from a combination of wool and animal hair woven with four tablets dates to the seventh century and the other, which is possibly part of an incomplete starting border made with three four-hole tablets, dates to the ninth century (Start 1950, 216−17, 224, figs 114 & 115, pl. 19.3). However, the latter find overlaps with the start of Scandinavian influence and Norse elements have been detected among the artefacts from the later periods of the site showing that some cultural interaction was taking place (Wallace 2016, 472). More significantly for indicating local production was the recovery of a square tablet of leather pierced with four holes, 36 × 39 mm, from an eighth-century deposit at Deer Park Farms (Wincott Heckett 2011, 143).

Evidence for local manufacture in Dublin The evidence from Dublin is more strongly assisted by the remarkable conditions of preservation, which resulted in what is considered the most complete archaeological record of any town in Europe for the tenth and eleventh centuries (Wallace 2016). Among different types of craftworking represented in particular districts of the town were amberworking, for which there was a tenth-century workshop in Fishamble Street; various bone, antler and walrus ivory workshops including one dating to the eleventh century that specialised in the production of gaming pieces at Christchurch Place; and fine metalworking that involved practising decorative motifs on pieces of bone and stone in same vicinity, which was close to where the cathedral was built a short time later (Wallace 2016, 291−301, 392−96). By contrast there is no evidence that tablet weaving was concentrated in specialised workshops and it appears to have been an activity undertaken on a household basis. Nevertheless, there is more material evidence for tablet weaving preserved from Dublin than from any other urban site in the Viking period and this provides a strong contrast to towns such as Hedeby and Birka where only isolated examples of weaving tablets have been recovered (Andersson 2003, 167). Little equipment is needed to carry out the task. The basic tools are a pack of tablets and some form of beater. Currently, at least 34 square tablets have been recorded from various sites in Dublin, including five different buildings on the Fishamble Street site (Pritchard 2014, 230). The tablets are chiefly made from thin, sawn plaques of wood or bone with one example made from horn and another possibly antler (Wallace 2016, 268, fig. 7.39) (Fig.  3.1). In size they range from 30  ×  31  mm to 41  ×  45  mm and among them is a group of five bone tablets from an eleventh-century deposit at High Street (National Museum of Ireland 1973,

3.  Evidence of tablet weaving from Viking-age Dublin

39

47, no. 230; E71:8963) and another group of six wooden tablets from Fishamble Street (E172:11367). Not all the tablets in each of these two groups are identical in size, three of the wooden tablets are 37 mm2, another is 38 mm2, and the other two are 34 × 35 mm and 35 × 38 mm; the bone tablets include four of which are 35 mm2 and the fifth is smaller being 31  mm2. Although the difference in size is small, Lise Ræder Knudsen has indicated that a difference of only 1 mm in a set of tablets of regular shape would have an impact on a finished band and, therefore, she suggests that they may have been used separately or perhaps for a differently patterned border of a band (Ræder Knudsen 2010, 155). However, the dimensions of the square, beech-wood tablets were not identical for the patterned band in the process of being woven that was placed on the ship burial at Oseberg (Grieg 1928, 180−81; Christensen & Nockert 2006, 143−45). The sizes of the holes on different tablets is also significant as the four tiny holes on a bone tablet with stamped ring and dot decoration mean that only very fine thread such as silk could easily have passed through the perforations, whereas it is noticeable that the diameter of the holes on the wooden tablets, in particular, are larger (Fig. 3.1). Several short wooden weaving swords have also been recovered from sites in Dublin including Fishamble Street, High Street and Christchurch Place (Ó Riordáin

Fig. 3.1. Assorted weaving tablets from Dublin excavations. (Photo: Valerie Dowling © National Museum of Ireland)

40

Frances Pritchard

Fig. 3.2. Short weaving sword, length 184 mm, from Fishamble Street, Dublin. NMI E190:908. (Drawing: John Murray © National Museum of Ireland)

1971, 80, fig. 22; Lang 1988, 76−77, figs 88 & 89) (Fig. 3.2). Ranging in length from 184 mm to 267 mm their distinctive shape with a curved blade, thicker upper edge and indentation next to a pointed tip gave them particular advantages as has been demonstrated by Katrin Kania in the course of her experimental work (https://www.pallia.net/en/blog/2018/09/21/finding-out-things). Thus once the weft has been inserted and tablets turned, the curved blade enables the weft to be pressed into position ‘very precisely yet gently, using a rolling motion of the curve across the fell’. The combination of a thin, blunt blade and thick upper edge helps with opening the shed and the pointed tip helps picking up cords, particularly in patterned work and enables any errors to be more easily rectified. Remarkably a miniature wooden pendant carved in the shape of weaving sword was also recovered from the Fishamble Street site (E172:13524; Wallace 2016, 292, fig. 8.20). A further indication of local production comes from analysis of the fleece types undertaken by Penelope Walton Rogers at the Anglo-Saxon Laboratory, York. A wad of unprocessed fleece (1.5 gm) recovered from outside one house proved to be a Semi-Fine fleece type (Walton Rogers 2013a) (Fig. 3.3). This fleece type was not common until much later in the medieval period and the fleece may have been imported, although, at present, information about sheep reared in Ireland is extremely limited. This wool appears to have been selected for special products including the tabby-woven veiling

3.  Evidence of tablet weaving from Viking-age Dublin

41

used for wool scarves and caps in Dublin and Birka-type textiles in Scandinavia (Walton 1988, 152−53; Walton Rogers 2013b). In respect of tablet-woven bands, a similar Semi-Fine fleece type was identified by Michael Ryder in both the warp and weft of a band of high quality woven with 16 tablets that was recovered from the aisle of a house at Fishamble Street (Fig. 3.4).

Wool bands Not all the textiles from the various sites have yet been conserved and analysed. However, from Fishamble Street there Fig. 3.3. Raw fleece of Semi-Fine type, staple length 50−55 mm from a deposit dating to 970 to 990 at are more than 30 all-wool tablet-woven Fishamble Street, Dublin. NMI E172:7428. (Photo: bands. Most are warp-twined bands Michael Pollard) with all cords equally twined between successive picks, which Collingwood identifies as the simplest form of tablet weaving (Collingwood 1982, 112). The plied warp yarn was usually threaded through all four holes of each tablet from opposite sides, a method that prevents the yarn from adjacent tablets catching against each other when the tablets are turned (Collingwood 1982, 74−5), and the tablets were given quarter turns together as one pack so that alternate cords twist in Z and S directions forming regular chevrons (Fig. 3.4). However, a few have a mixed sequence of Z and S-cording, one band for example made with 12 tablets includes three pairs of cords twisted in the same direction (SZSZSSZZSZZS) (E172:10571, Fig. 3.5) and another made with 17 tablets includes a sequence of four tablets threaded in a Z direction (E172:3267). Packs of between six and 26 tablets Fig. 3.4. Detail of a tablet-woven were used for these simple bands which were band, width 12 mm, made with 16 proficiently made with a relatively even tension four-hole tablets using wool of Semiand approximately 7 to 9 picks per cm, although Fine fleece type from the north aisle a few are twice as fine with up to 17 picks per cm of a house, late tenth/early eleventh (Fig. 3.4). In width they range from 6 mm to 35 mm. century at Fishamble Street, Dublin. The narrowest band was made with four tablets all NMI E172:11815. (Photo: Michael threaded in the same S direction, which has caused Pollard)

42

Frances Pritchard

Fig. 3.5. Coarse tabby-woven cloth bound with a wool tablet-woven band, width (when unfolded) 33 mm, made with 12 tablets from a deposit dating to 960 to 980 at Fishamble Street, Dublin. NMI E172:10571. (Photo: Michael Pollard)

it curl into a corkscrew shape (E172:10940). Now in two lengths totalling 370 mm, it may have been used to fasten a garment or furnishing. Other bands may initially have been used as garters or girdles before being recycled. Thus for comparison, a band, 15 mm in width, has been interpreted as a garter from the chieftain’s burial dating to the 970s at Mammen, Jutland (Ræder Knudsen 1991, 149; Munksgaard 1991), and a patterned tablet-woven band, 16−17 mm in width, was among several worn as girdles by a woman in a grave dating to the eleventh century at Kirkkomaki, Turku in south west Finland (Penna-Haverinen 2010, 195). It is noticeable, however, that many reconstructions of female Viking-period dress, from the tenth as well as the ninth century, omit girdles due mainly to the scant amount of evidence preserved but they are illustrated for men in the form of belts (Mannering 2017, 158, 162, 164, figs 7.10, 7.11, 7.17 & 7.18). Some of the wider bands in Dublin are folded double and show signs of heavy wear along the centre fold indicating that they were used as bindings. In the five examples where traces of the cloth have been recorded inside the fold, the fabric is a coarse grade tabby-woven cloth (Fig. 3.5). It, therefore, seems probable that they are the remains of cloaks with the tablet-woven edgings providing a contrast of colour, as well as preventing the cut edges of the cloth from fraying. Dye analysis undertaken by Penelope Walton Rogers on one band used as a binding indicated the presence of a red yarn dyed with a species of wild madder or bedstraw (E172:13676).1 Generally, however, the acidic conditions of preservation, which has resulted in the textiles becoming stained brown from the tannins in the soil, means that no patterning from the use of differently coloured warp threads can be detected by eye

3.  Evidence of tablet weaving from Viking-age Dublin

43

Fig. 3.6. Patterned, all-wool tablet-woven band, maximum width 46 mm, from John’s Lane, Dublin. NMI E173:158. (Photo: Michael Pollard)

on the bands, although from microscopic examination it appears some threads were spun from naturally pigmented wool, which would have imparted some degree of colour variation. A few flat bands appear to have been made with the warp yarn threaded through only two holes of each tablet and rotated with half turns forwards and back to form a warp-faced tabby weave. One of these bands (E172:3087), 18 mm in width, has bare weft loops remaining on either side and this may have resulted from each edge tablet having been threaded with a contrasting plant fibre yarn, which has disintegrated. A small number of tubular, tablet-woven bands, 5 to 8 mm in diameter, have been recovered (Pritchard 1992, 100). They are unpatterned and were made in a similar manner to the flat, four-hole-threaded, warp-twined bands described above except that the weft thread was passed through the shed from the same side after each rotation of the tablets and was then pulled tight to form a tube (Collingwood 1982, 367−68). A similar method was employed to edge the opening of what was probably a pouch made from several recycled scraps of woven silk (E172:7821), although only two tablets were used and the wool weft was inserted with a needle through the cloth. A more complex band was recovered from a disturbed level at John’s Lane. Now in four fragments totalling 190 mm in length and 46 mm in width, it is a weft-patterned band made with 45 tablets using two-ply (Z-spun/S-ply) wool thread (Fig. 3.6). The borders are formed from four alternately threaded four-hole tablets on each side and rotated continuously with quarter turns resulting in typical warp-twined cords. The centre field has a geometric pattern of diagonals, diamonds and small crosses worked in a dark-coloured thread, possibly dark blue, on what appears to be a red ground, although negative dye results were obtained. It has a three-dimensional appearance partly due to the use of a supplementary weft of doubled two-ply thread,

44

Frances Pritchard

which extends only across the central field and floats from one motif to another on the back so that the band is not reversible. The ground appears to combine areas of 3.1 twill with tabby suggesting that the 37 tablets used to create the central field were threaded through two holes and manipulated individually. Further analysis is needed to establish the precise threading and turning sequence but this is hindered through the degree of wear that the band has suffered. A three-dimensional effect often results from threads degrading and, although there are no obvious traces of a missing bast fibre warp, the possibility cannot be dismissed, and it may mean that the tablets used in the patterned zone also had four threads rather than two making it more similar to a double-sided band from the chieftain’s burial at Mammen where a plant fibre warp yarn has also disappeared due to the adverse soil conditions (Ræder Knudsen 1991, 149).

Silk bands brocaded in metal thread Lightweight silk goods were plentiful in the town ranging from tabby-woven ribbons, silk scarves and caps, to recycled strips of samite, which appear mainly to have

Fig. 3.7. Knotted skein of raw silk from Fishamble Street, Dublin. NMI E190: 3498. (Photo: Michael Pollard)

45

3.  Evidence of tablet weaving from Viking-age Dublin

trimmed the edges of linen garments (Pritchard 1988). Imported silk skeins and balls of silk thread were available for making silk products locally (Fig. 3.7), or for sewing and embroidery, and as well as a length of plied silk thread remaining attached to a needlecase, some is wound round what must have been a treasured pot sherd of Roman-period samian ware (Wallace 2016, 268, fig. 7.40). Precious metal was also available in the town. Small strips of gold were recovered from John’s Lane (E173:211 & 212) and seven fragments of gold wire up to 40 mm in length were excavated from a deposit at Fishamble Street (E41:3196). However, as all the narrow strips of metal sheeting used in the brocaded tablet-woven bands was S-twisted round a silk core, it seems unlikely that any of this gold was destined for use in tablet weaving. Nevertheless, the local inhabitants did have access to supplies of gold and silver foil, droplets of gold have been detected on crucibles, and much metalworking was carried out in the neighbourhood (Wallace 2016, 344). There are at least six bands with geometric patterns brocaded in metal thread. Four gold bands and one silver band have a similar basic structure using fine S-twisted silk thread for the warp inserted through all four holes and the tablets were given quarter turns in a forward direction for the ground weave (Table 3.1). The tablets were not consistently threaded from opposite sides, perhaps because silk yarn is smoother than wool and the threads are less likely to become entangled side by side, Table 3.1. Table of metal-brocaded bands Accession no.

Metal

E71:11124

Dimensions mm

Total no. Picks Threading of tablets per sequence in cm central field

No. of tablets in central field

Stave borders

L

W

gold

257 (3 frags.)

8.5

19

28

9 all S

9

2 outer (S, Z threaded), 1 inner (Z threaded)

E122:13272 (Fig. 3.9)

gold

c.23 (2 frags.)

5.5

16

28

alternately Z&S

16

none

E122:13448 (Fig. 3.9)

gold

31

8.5

25

24

S,Z,S,Z,3S, alternately Z&Sx4

15

2 outer (Z, S threaded), 1 inner (Z threaded)

E190:1194 (Fig. 3.8)

gold

16

9

31

44

8S,Z,3S, Z,S,Z,S,Z

17

3 outer (S, Z, Z on left; S, Z, S on right), 1 inner (S on left; Z on right)

E172:10679 (Fig. 3.10)

silver

150

42

61

20

alternately S&Z

41

1 outer (Z threaded), 2 inner (Z threaded)

46

Frances Pritchard

or perhaps this helped in memorising the pattern. One of the gold bands has half the pack of tablets in the central field threaded in a S direction and another has all nine of the tablets used for the central zone Z-threaded. Most were woven using an odd number of tablets and this meant that they also used an odd number of tablets for the patterned field. The tie-down points of the brocading weft forming the pattern is consistently carried out by one thread from a tablet. The ground weft is always silk, probably dyed red as from the reverse the bands now possess Fig. 3.8. A & B. Gold-brocaded tablet-woven band, width a drab pinkish tone, and the 9 mm made with 31 tablets from Fishamble Street, Dublin. brocading weft extends to the NMI E190:1194. The dots in the diagram represent the gold inner cord of the outer stave brocading weft. (Photo: Valerie Dowling © National Museum borders, where such borders are present (Fig. 3.8). of Ireland, diagram: Christina Unwin) The bands are preserved as very short pieces and two are in fragments. In addition, two are folded at one end with hems of 3 mm and 8 mm respectively. Not all are expertly woven and a simple error in turning the tablets is apparent on one of the bands recovered from a late eleventh-century deposit inside a house at Christchurch Place (E122:13448, Fig. 3.9). Moreover, the quality of the gold appears to vary in its purity and only in one band does the metal still gleam (Fig. 3.8). The metal strip is also very thin and has become extremely worn on most of the bands. It is S-spun round a core of silk, thereby contrasting with the tablet-woven bands recovered from graves in Birka, Sweden, and further east in the Shestovica cemetery near Chernigov, Ukraine, and Pskov, Russia, where drawn wire, usually silver and often used double, was selected for the brocading (Geijer 1938, 80−83; Mikhailov 2008, 18−19). The narrow width of the gold bands suggests that they were probably worn round the head either in association with caps or veils or as headbands. They were a very characteristic feature of dress within the Viking world in the tenth century, and burial evidence, for example at Birka, indicates that they were restricted to wealthy individuals as a marker of status (Geijer 1938; Hägg 1982). This is further

3.  Evidence of tablet weaving from Viking-age Dublin

47

supported by the discovery of a male burial in a tenth-century boat grave at Ness, Hamarøy, Nordland, Norway excavated in 2011 that yielded part of a tablet-woven band brocaded in silver filé thread attached to the man’s skull (Vedeler 2014, 42−43). The silver band from Dublin is considerably wider than the gold bands, perhaps reflecting the cheaper cost of the metal, and the weft has not been so densely beaten (Table 3.1 & Fig. 3.10). All the tablets were threaded alternately in pairs. In view of its width, it also has a secondary pattern of chevrons within the Fig. 3.9. Diagrams of the patterning of two stave borders on each side. The pattern in the gold-brocaded tablet-woven bands from central field repeats after every 52 brocading a house dating to the eleventh century at wefts apart from where it changes along its Christchurch Place, Dublin. NMI E122:13272 length from one of lozenges and crosses to a (left), E122:13448 (right). (Diagrams: Christina Unwin) diagonal lattice interlace, which was a popular pattern on such bands (Pritchard 1988, 154) and was recorded at Birka (Geijer 1938, 83, fig. 20). At one point along the Dublin band (indicated on the diagram by an arrow) all the tablets were reversed to prevent the warp threads becoming overtwisted. As with the gold thread, the silver thread is similarly S-spun round a core of silk, although most of the metal has degraded to such an extent that it is preserved only as small black specks. X-ray fluorescent analysis undertaken by Justine Bayley, formerly of English Heritage, indicated that the silver content is very pure. Unfortunately, although silver was extremely popular within the Viking world and much hack silver, silver ingots and silver bullion has been recovered from hoards in Ireland, tracing the geographic source of the metal is problematic due to so much silver being melted and reused (Griffiths 2010, 103−105). A comparable silver-brocaded band at least 53 mm in width with a secondary border pattern in addition to that of the central field, was recovered from a burial dating to the late tenth century at Hørning, Jutland (Voss 1991, 194−95, fig. 12; Ræder Knudsen 2005, 38, 42−43, table 7.1). The woman was evidently of a high status as she was placed in a carriage for the burial and there the silver band was interpreted as decorating the edge of a gown. Silver bands also trimmed the edges of kaftan-type jackets and tunics at Birka (Hägg 1986). The fact that all these metal-brocaded bands in Dublin are woven in a similar way suggests that they are likely to have been produced by makers who learnt the technique in the same community. This is reinforced by a survey that Ræder Knudsen undertook of brocaded tablet-woven bands from tenth-century contexts, chiefly in Denmark, Sweden and England, in which she identified subtle differences in terms of technique and the materials used with the amount of silk thread often being limited, thus pointing to different workshop practices. The Dublin bands of this type compare

48

Frances Pritchard

Fig. 3.10. A & B. Silver-brocaded tablet-woven band, width 42 mm, made with 61 tablets from Fishamble Street, Dublin. NMI E172:10679. (Photo: Jon Bailey, diagram: Christina Unwin)

3.  Evidence of tablet weaving from Viking-age Dublin

49

most closely with those excavated from graves in Denmark except that there flax thread was often substituted for silk thread where it would not have been visible (Ræder Knudsen 2005, 42−43, table 7.1). There are at least three further silk bands, which may have been patterned with a brocading weft of gold or silver thread as they are also characterised by fine S-twisted silk warp yarn. However, no ground or brocading weft remains indicating that, by contrast, they may possibly have incorporated some type of plant fibre, which was more susceptible to decay in the anoxic environment. One of these bands was found in association with a narrow strip of open-weave silk, 120 mm in width (Wincott Heckett 2003, 38−39, fig. 24). Another tablet woven band from Dublin is also patterned with a silver filé thread but it differs technically. The silver has a tiny percentage of copper in the metal and the lamella was spun round a plant fibre core, probably either flax or nettle. Scarcely any of this thread is preserved and similarly the plant fibre threads used in the warp have disintegrated leaving a ghost of the pattern (Fig. 3.11). Both a second warp and a second weft are silk, of which the latter has a slight S-twist. The band is 55 mm in width and was made with about 70 tablets. It has been cut at both ends leaving a piece 190 mm in length patterned along the centre with a row of alternating saltire crosses and lozenges separated by bars of tabby weave. It is edged on each side with a chevron-patterned border, while the cords of the inner borders are no longer fully intact and appear to have been threaded with two silk and two plant fibre threads in diagonally opposite holes. The main field with its areas of plain weave suggests that it used a two-hole threading system as discussed by Collingwood in relation to some bands possibly dating around the eighth to tenth centuries at the abbey of St Maurice, Switzerland (Schmedding 1978, 171−74, cat. nos. 140, 142; Collingwood 1982, 208−10). Consequently, it appears to have been produced in a different workshop from any of the other Dublin bands.

Fig. 3.11. Patterned tablet-woven band, width 55 mm, made with threads of silk, silver filé and a plant fibre from Fishamble Street, Dublin. NMI E172:14373. (Photo: Michael Pollard)

50

Frances Pritchard

Starting borders No tablet-woven starting borders have been identified from among the many hundreds of textiles examined from the site. However, only a very small number of borders, as opposed to selvedges, are preserved and this may reflect the fact that most of the textiles were recycled before finally been discarded as rubbish.

Conclusion This survey of tablet weaving in tenth and eleventh-century Dublin shows that it was a skill carried out within the town. The wool products were used as bindings to strengthen the edges of garments in a similar way to narrow bands in tabby and chevron twill that were woven from combed (worsted) yarn (Pritchard 2017, 119). Other tablet-woven wool bands were probably used as garters to hold leg wrappings in place, or girdles, for which there is a long history of examples made in this technique, although few have been preserved from Viking-age graves. More elaborate, patterned bands woven from silk and metal thread were also in circulation, and at least some of these may have been woven locally. This is because the size of the holes in some of the Dublin tablets would only have been suitable for very fine thread, especially silk, and there appears to have been no difficulty in obtaining supplies of silk thread. On the basis of tenth-century burials in Norway, Denmark, Sweden and Russia, it is probable that most of the narrow metal-brocaded bands were worn as headbands by either men or women although some may have been applied as trimmings to the edges of men’s and women’s garments. The significance of the material lies not only in the quantity of it but also that it was recovered from urban rubbish deposits thereby reflecting everyday life rather than having any special funerary associations. It thus provides an insight into the lifestyle, household activities and craft skills of some of the town’s Hiberno-Scandinavian inhabitants.

Acknowledgements I am very grateful to Dr Patrick Wallace, former director of the National Museum of Ireland, for enabling me to analyse many of the Dublin textiles; to Katrin Kania for discussing recording procedures with me; and to Clare McNamara, Image Library Officer, National Museum of Ireland, for obtaining the images that I requested at a very difficult time. Note

1 The method of dye testing was solvent extraction followed by absorption spectrophotometry.

3.  Evidence of tablet weaving from Viking-age Dublin

Bibliography

51

Andersson, E. 2003. Tools for Textile Production from Birka and Hedeby. Birka Studies 8 (Stockholm, Riksantikvarieämbetet). Christensen, A.E. & Nockert, M. (eds) 2006. Osbergfunnet IV. Tekstilene (Oslo, PDC Tangen). Collingwood, P. 1982. The Techniques of Tablet Weaving (London, Faber and Faber). Geijer, A. 1938. Birka III. Die Textilfunde aus den Gräbern (Uppsala, Almqvist & Wiksells). Greig, S. 1928. Kvindearbeitet. In A.W. Brøgger & H. Schetelig (eds), Osebergfundet Bind II (Oslo, Universitetets Oldsaksamling), 173−98. Griffiths, D. 2010. Vikings of the Irish Sea (Stroud, The History Press). Hägg, I. 1982. Some remarks about Birka Clothing. In L. Bender Jørgensen & K. Tidow (eds), Archäologische Textilefunde, Textilsymposium Neumünster (Neumünster, Textilmuseum Neumünster), 249−65. Hägg, I. 1986. Die Tracht. In G. Arwidsson (ed.), Birka II:2. Systematische Analysen der Gräberfunde (Stockholm, Almqvist & Wiksell), 51−72. Japerse, J. 2017. Matilda, Leonor and Joanna: the Plantagenet sisters and the display of dynastic connections through material culture. Journal of Medieval History 43, 523−47. Lang, J.T. 1988. Viking-Age Decorated Wood. A Study of its Ornament and Style. Medieval Dublin Excavations 1962−81, Series B, vol.1 (Dublin, Royal Irish Academy). Mannering, U. 2017. Iconic Costumes: Scandinavian Late Iron Age Costume Iconography (Oxford & Philadelphia, Oxbow Books). May, F.L. 1957. Silk Textiles of Spain: Eighth to Fifteenth Centuries (New York, Hispanic Society of America). Mikhailov, K. 2008. New finds of Viking Age textiles in Ukraine and Russia. Archaeological Textiles Newsletter 47, 17−20. Munksgaard, E. 1991. Kopien af dragten fra Mammengravern. In M. Iversen (ed.), Mammen. Grav, kunst og samfund i vikingetid. Jysk Arkæologisk Selskabs Skrifter 28 (Højbjerg, Aarhus Universitetsforlag), 151−53. National Museum of Ireland, 1973. Viking and Medieval Dublin, National Museum Excavations, 1962−1973. Catalogue of Exhibition (Dublin). Ó Riordáin, A.B. 1971. Excavations at High Street and Winetavern Street, Dublin. Medieval Archaeology 15, 73−85. Penna-Haverinen, S. 2010. Patterned tablet-woven band – in search of the 11th century textile professional. In E. Andersson Strand, M. Gleba, U. Mannering, C. Munkholt & M. Ringgaard (eds), North European Symposium for Archaeological Textiles X (Oxford & Oakville, Oxbow Books), 195−200. Pritchard, F. 1988. Silk braids and textiles of the Viking age from Dublin. In L. Bender Jørgensen (ed.), Archaeological Textiles, Report from the 2nd NESAT Symposium 1984 (Copenhagen, Arkaeologisk Institut), 149−61. Pritchard, F. 1992. Aspects of the wool textiles from Viking age Dublin. In L. Bender Jørgensen and E. Munksgaard (eds), Archaeological Textiles in Northern Europe, Report from the 4th NESAT Symposium. Tidens Tand 5 (Copenhagen, Det Kongelige Danske Kunstakademi), 93−104. Pritchard, F. 2014. Textiles from Dublin. In N. Coleman & N. Lokka (eds), Kvinner I Vikintid (Oslo, Scandinavian Academic Press), 224−40. Pritchard, F. 2017. Twill weaves from Viking Age Dublin. In M. Bravermanová, H. Březinová & J. Malcolm-Davies (eds), Archaeological Textiles – Links between Past and Present, NESAT XIII (Liberec & Prague, Technical University of Liberec, Faculty of Engineering/ Institute of Archaeology of the CAS, Prague), 115−23.

52

Frances Pritchard

Ræder Knudsen, L. 1991. Det uldne brikvaevede band fra Mammengraven. In M. Iversen (ed.), Mammen. Grav, kunst og samfund I vikingetid. Jysk Arkæologisk Selskabs Skrifter 28 (Højbjerg, Aarhus Universitetsforlag), 149−50. Ræder Knudsen, L. 2005. Brocaded tablet-woven bands: same appearance, different weaving technique, Hørning, Hvilehøj and Mammen. In F. Pritchard & J.P. Wild (eds), Northern Archaeological Textiles NESAT VII (Oxford, Oxbow Books), 36−43. Ræder Knudsen, L. 2010. Tiny weaving tablets, rectangular weaving tablets. In E. Andersson Strand, M. Gleba, U. Mannering, C. Munkholt & M. Ringgaard (eds), North European Symposium for Archaeological Textiles X (Oxford & Oakville, Oxbow Books) 150−56. Schmedding, B. 1978. Mittelalterliche Textilien in Kirchen und Klöstern der Schweiz. Schriften der AbeggStiftung 3 (Bern, Stämpfli & Cie). Start, L.E. 1950. Textiles. In H. Hencken (ed.), Lagore Crannog: an Irish Royal residence of the 7th to 10th centuries A.D. Proceedings of the Royal Irish Academy 53, 203−24. Vedeler, M. 2014. Silk for the Vikings (Oxford & Philadelphia, Oxbow Books). Voss, O. 1991. Hørning-graven. En kammergrav fra o. 1000 med kvinde begravet i vognfading. In M. Iversen (ed.), Mammen, Grav, kunst og samfund i vikingetid. Jvsk Arkæologisk Selskabs Skrifter 28 (Højbjerg, Aarhus Universitetsforlag), 189−203. Wallace, P.F. 2016. Viking Dublin. The Wood Quay Excavations (Co. Kildare, Irish Academic Press). Walton, P. 1988. Dyes and wools in Iron Age textiles from Norway and Denmark. Journal of Danish Archaeology 7, 144−58. Walton Rogers, P. 2013a. A fragment of unprocessed fleece from Viking Age Dublin, E190:7428. Anglo-Saxon Laboratory Report, 11 January 2013. Walton Rogers, P. 2013b. Fleece types in further textiles from Viking Age, Dublin. Anglo-Saxon Laboratory Report, 3 October 2013. Wincott Heckett, E. 2003. Viking Age Headcoverings. Medieval Dublin Excavations 1962−81, Series B, vol. 6 (Dublin, Royal Irish Academy). Wincott Heckett, E. 2011. The textiles. In C.J. Lynn & J.A. McDowell (eds), Deer Park Farms: The Excavation of a Raised Rath in the Glenarm Valley, Co. Antrim. Northern Ireland Archaeological Monographs 9 (Belfast, TSO), 354−66.

Chapter 4 The so-called ‘Palermo bands’ and their technique Regula Schorta

In The Techniques of Tablet Weaving, Peter Collingwood presents an illustration of a beautiful weft-brocaded band with warp threads raised above the brocading to make animals and birds in four colours (Fig. 4.1) (Collingwood 1982, 349, pl. 205).1 Collingwood does not seem to have studied its structure, given that the accompanying text is very brief and confined to emphasising the band’s relationship to a tabletwoven orphrey with a similar warp-patterning technique, which is itself analysed in detail (Collingwood 1982, 345–48).2 In fact, the four-colour-and-gold band is not a warp-twined tablet weave, but a simple warp-weft interlaced structure – a fact that Peter Collingwood would certainly have understood immediately, had he examined the band in person. This four-colour-and-gold band is a particularly beautiful and well-preserved representative of a rather large group of bands. These bands have been colloquially referred to as ‘Palermo bands’ since the early twentieth century, when Otto von Falke mentioned them briefly in his Art History of Silk Weaving and attributed them to Palermo workshops for historic and stylistic reasons (von Falke 1913, 1:122).3 For convenience, the term ‘Palermo bands’ will be used in this paper as well, since it is short and commonly used. The first and most important criterion for treating the bands as a group is not, however, their presumptive place of production – which will be returned to later – but the distinctive technique used to make them. The Palermo bands’ resemblance to gold-brocaded, tablet-woven bands with regard to their feel, usage, material, high warp density and sometimes also decor, has often led – and sometime still leads – to confusion regarding the technique used to manufacture them. Also, the steep angle of the basic twill is strongly reminiscent of the cords produced by tablet weaving. Already in 1955, however, Sigrid MüllerChristensen carefully distinguished between ‘woven bands’ and ‘tablet-woven

54

Fig. 4.1. Gold-brocaded band with animals and birds in four warp colours. Twelfth/ thirteenth century. Vienna, MAK – Museum für Angewandte Kunst, inv. no. T 752. (© MAK, photo: Aslan Kudrnofsky)

Regula Schorta bands’ (Müller-Christensen 1955, 11, and, for instance, cat. nos. 16, 35). In the interim, detailed analyses have been published by scholars such as Gabriel Vial (1995) and Daniël De Jonghe. The latter made a careful analysis of the girdle for the imperial sword (Reichsschwert) in Vienna (Fig. 4.2), which shall serve as a model here (De Jonghe 1991, 71–75). The girdle is a band 6.3 cm wide and 189 cm long, with fittings of gilded silver (Imperial Treasury, Kunsthistorisches Museum, Vienna, inv. no. WS XIII 9; Nobiles Officinae 2004, 280– 82). There is no lining or other additions. The band is divided lengthwise into three sections, all with a golden ground and set off from each other by narrow plain silk stripes. The selvedges also consist of narrow stripes without brocading. The broad middle section is decorated with ten motifs, most of them appearing only once. The narrow side sections both have the same inscription, CRISTVS RIEHGNAT CHRISTVS INQPARAT DEVS, with the letters spaced widely and quite evenly. The basic structure of all of the Palermo bands is that of a warp-faced twill, often an over four, under one (4/1) twill, as with the girdle of the imperial sword, and either in S or in Z direction. There are also 3/1 twills and, more rarely, 2/1 twills. The warp ends are always silk, Z-twined from two ends, and usually arranged in consecutive ends of several colours. Plaincoloured, brownish stripes form the narrow lines separating the individual sections. Judging from the majority of medieval gold-brocaded bands – which, in fact, combine and enhance the gold with red – it can be assumed that the brownish silk is actually discoloured from red (see Hofenk de Graaf 2004, 147). The gold-covered main sections of the bands are multi-coloured. The warp is striped thread by thread, following a strict rule: there is always one colour fewer than the number of threads in

4.  The so-called ‘Palermo bands’ and their technique

55

Fig. 4.2. Girdle for the imperial sword (Reichsschwert), detail. Twelfth/thirteenth century, with later fittings. Vienna, Kunsthistorisches Museum, Imperial Treasury, inv. no. WS XIII 9. (© KHMMuseumsverband)

the twill unit. The girdle of the imperial sword is thus striped with four colours: one end is brownish (previously red), one green, one pale purple and one white. The effect of such a striped twill with high warp density is close to an optical illusion – the eye tends not to see the overall structure, but rather simply combines the floats of each colour, in particular if there is one colour that is visually dominant. As a result, the visual appearance is that of a twill, with its diagonal running in the opposite direction to that of the actual structure (Figs 4.3 & 4.4). On the paper model, where warp yarns cannot be squeezed, the optical illusion of diagonal colour stripes is less evident and the actual structure is easy to recognise. On the bands themselves, however, the effect is striking, but seldom seen: if a band is well preserved, the gold completely covers its front face; if it is not well preserved and the gold is worn away, the colours have usually deteriorated too, often to such a degree that they all uniformly look brown. The same effect is also visible on the reverse (Fig. 4.5). While, at first glance, the twill looks S-directed, the diagonal is actually Z, but far steeper, and not that easy to follow, because it is blurred by the constant colour change. As has been mentioned, the front of these bands is covered by a gold lancé-weft. The gold thread always consists of a lamella of either gold or gilded silver, wound in S-direction around a silk core of often yellowish colour. It is always single, and bound in twill by select warp ends. One of the colours of the warp, often the red one, always binds the lancé-weft in the technique known in French as à liage repris. Since the colours follow each other in strict sequence and since there are four colours in

56

Regula Schorta

Fig. 4.3. Palermo band weave structure; where the gold lancé weft is worn away, the 4/1 S twill with four warp colours is visible. Detail of Fig. 4.16. (© Victoria and Albert Museum, photo: Silvija Banić)

Fig. 4.4. Palermo band weave structure; 4/1 S twill with four warp colours. (Paper model: author)

the girdle, every fourth warp end binds the gold. It is, of course, advisable to position the gold binding points in such a way that they do not interfere with the binding done by the same warp ends with the foundation wefts. This is most easily achieved by placing the gold binding point close to the middle of the long warp float of the ground twill, and by using the same twill – 4/1 in our case. This automatically results in a diagonal opposite to the one of the ground structure (Fig. 4.6). The basic structure of Palermo bands is thus a foundation weave in warp-faced twill and a lancé-gold weft effect bound in weft-faced twill, both twills with the same unit but opposite diagonals. To create the coloured silk motifs on this gold surface, certain warp ends are brought to the front (Fig. 4.7). Because the warp density is so high, the ends of one colour only – which represent no more than one-fourth of all of

4.  The so-called ‘Palermo bands’ and their technique the warp ends – nicely cover the surface. The movement of the threads is, however, somewhat awkward, since the underlying structure is left unchanged – the warp ends forming motifs must re-enter the structure in order to be bound by the foundation weft. In the upper section of the paper model (Fig. 4.8), the black warp ends are lifted above the lancé weft, while the wefts are unnaturally spaced in order to reveal the black ends re-entering the ground twill. As soon as the lancé wefts lie closely next to each other, the effect in the colour motifs is that of a mock twill or ‘twill without binding points’, with diagonals marked only by lines of depressions. This diagonal runs once again in the opposite direction to that of the foundation twill – according to the same rule that applies to the binding of the lancé weft: in the case of the Vienna girdle, it is every fourth end out of a twill of five that forms the motif. In fact, the effect is not a twill, but a simple ‘over all’, which, in French, is termed ‘masse’. It is not entirely correct to describe this effect as twill, but in order to evoke the effect of diagonals, a formula such as ‘5/0 mock twill’ might nevertheless perhaps be envisaged. The structural analysis of Palermo-band weaves can, therefore, be summarised as follows: • there is a warp-faced foundation twill formed by a warp consisting of consecutive ends of different colours and a weft • there is an overall effect of a lancé gold weft, bound in weft-faced twill by the ends of only one of the warp colours • there are motifs formed by lifting all of the ends of one of the warp colours above the gold wefts • there is a consistent mathematical relationship between these structures, based on the fact that the number of warp colours is always one fewer than the number of warp ends in a single unit of the foundation twill.

57

Fig. 4.5. Palermo band weave structure; 4/1 S twill with four warp colours, reverse. (Paper model: author)

Fig. 4.6. Palermo band weave structure; 4/1 S twill with four warp colours, lancé weft bound in 1/4 Z twill by every fourth warp end. (Paper model: author)

There are further variations, beyond this basic structure. An analysis of the orphrey band of the chasuble of St Valerius in Barcelona was undertaken by Gabriel Vial (Fig. 4.9) (Museu del Disseny, Barcelona, inv. no. MTIB

58

Fig. 4.7. Mitre from St Peter’s abbey in Salzburg, detail of gold band. Twelfth/thirteenth century. Riggisberg, AbeggStiftung, inv. no. 233. (© Abegg-Stiftung, photo: Christoph von Viràg)

Fig. 4.8. Palermo band weave structure; 4/1 S twill with four warp colours, lancé weft bound in 1/4 Z twill by every fourth warp end, with warp effects raised above the lancé weft. (Paper model: author)

Regula Schorta 5572; Vial 1995). The band is 5.3 cm wide and has a pattern of rather distorted and difficult-to-recognise birds, and foliage. There is nothing peculiar about its foundation weave, a 3/1 Z-twill with three colours, today dark brown, cream and pink. The gold lancé-weft is, however, bound by the pink warp ends in a 1/6 S-twill. This is not an obvious choice, given that the two structures – i.e. the four-end and the seven-end twill – at some point or another always interfere with each other. The result is a slight, easily recognisable deviation of the gold weft from its horizontal line. In the literature, several other combinations of bindings have been described. However, it seems that many of them do not stand up to scrutiny. The mock twill of the colour effects, in particular, often leads to misunderstandings. Very careful analysis is necessary, while results deviating from the general scheme are best double-checked.

Patterns Patterns and pattern arrangements of the Palermo-type bands follow characteristic typologies. In her PhD thesis on the episcopal mitre, Heidi Blöcher (2012) discusses them in detail. What follows is a brief summary of her findings.4 Bands with small motifs, placed at regular intervals, form a first group (Fig. 4.10). The width of the bands varies, broad ones are up to 9 cm wide, displaying, one on top of the other, two rows of identical plants, quadrupeds, birds or geometric elements in the middle section, while another set of motifs decorates the narrower sections. A key object is the mitre found in the grave of bishop Amedée II in Lausanne, Switzerland (Blöcher 2012, 259–60; Flury-Lemberg 1988, 248–50, 468), because it is dated to before 1159. Very similar bands have been found in an unidentified bishop’s grave in Halberstadt (Blöcher 2012, 250–51, Borte 2), and also in the grave of archbishop Andreas Sunesson, who died in 1228 in Lund, Sweden (Branting & Lindblom 1929, 98–99, pl. 117A; Blöcher 2012, 263–65). Of course, burial dates can never be more than a terminus post quem non, a latest possible date, whereas it has long been known that old and unfashionable textiles

4.  The so-called ‘Palermo bands’ and their technique were sometimes used as funeral vestments even for dignitaries. Representative bands for the second group as established by Heidi Blöcher are often large, up to 12 cm wide, and characterised by diagonals that articulate the surface of the bands (Fig. 4.11). The band used for a mitre found in Mainz cathedral, from an unknown bishop’s grave (Schorta & Wilhelmy 1998; Blöcher 2012, 275–76), shows a careful arrangement of S- and Z-inclined geometric bands, framing triangles of inhabited scrolls. There are also scrolls without animals and animals without scrolls – the whole setup is carefully planned, but does not follow a defined repeat. The narrow side sections are sometimes set with little animals and ornaments, similar to the first group, while sometimes there is a continuous scroll, as on the circulus band decorating the imperial mitre found in the grave of Emperor Henry VI in Palermo (British Museum 1878, 0907.1-3; Granger-Taylor 1994). The year of his death, 1197, dates this group to the twelfth century, too.

59

Fig. 4.9. Variation of weave structure; orphrey band of the chasuble of St Valerius. Second half of the thirteenth century. Barcelona, Museu del Disseny, inv. no. MTIB 5572. (Abegg-Stiftung, photo: Christoph von Viràg)

Fig. 4.10. Bands with small motifs, placed at regular intervals; mitre from the grave of bishop Amedée II of Lausanne (d. 1159). Lausanne, Musée Historique. (Abegg-Stiftung, photo: Anne Javor)

60

Regula Schorta

Fig. 4.11. Band with inclined geometric bands and inhabited scrolls; mitre from the grave of an unknown bishop. Second half of the twelfth century. Mainz, Erzbischöfliches Diözesanmuseum. (Abegg-Stiftung, drawing: Geertje Gerhold)

A third group of bands is distinguished by rather loosely distributed, large and often figurative elements in the main section of the bands (Fig. 4.12). The bands decorating the hem of the Utrecht alb (Het Catharijneconvent, Utrecht, inv. no. OKM t91), probably a secular vestment (Defoer 2004; Keupp 2014, 75 note 41), can be attributed to this group, but perhaps the most beautiful bands are those found on two mitres, one in New York, the other in Riggisberg. The former, in the Metropolitan Museum, comes from Neustift monastery close to Bressanone, South Tyrol, and is said to go back to the monastery’s founder, bishop Hartmann I of Bressanone (d. 1164). A lion, a mermaid, two centaurs and a figure that is difficult to identify are to be found on its horizontal bands (Blöcher 2012, 286–88, Borte 1). The band of the Riggisberg mitre, from the convent of St Peter’s at Salzburg, shows a regular sequence of alternating narrow transverse strips, and either a tree-like scroll ornament or a mermaid holding her two fishtails in her hands (Otavský & Wardwell 2011, 154–63; Blöcher 2012, 310–13). This band is as obviously vertically oriented as the one from New York is horizontal – it seems that at least some of the bands were made with a clear direction of use in mind. This would also apply to two further bands, which are closely related to each other. They show the same bestiary – albeit on a much smaller scale – and are set in a framework of interlinked medallions. Although intended for horizontal use, one of these bands has been used for the lappets of a mitre preserved in Salzburg cathedral (Blöcher 2012, 323–27, Borte 3). A nearly identical band (Fig. 4.13) decorates the shoes of the imperial vestments in Vienna (Imperial Treasury, Kunsthistorisches Museum, Vienna, inv. no. WS XIII 13; Nobiles Officinae 2004, 275–79). While the shoes themselves were remade in the seventeenth or eighteenth century, both the red fabric and the gold band of their upper date to the twelfth century. A fourth group of bands is dominated by colourful and versatile patterns of lozenges, lattices, triangles and crosses (Fig. 4.14), all depicted in a linear manner very much reminiscent of tablet-woven bands or counted-needlework fillings. A mitre in the Benedictine monastery of Disentis in Switzerland preserves high-quality bands of this type (Blöcher 2012, 240–43). Potentially noteworthy is the comparatively thin gold thread used, a feature not usually seen on the other types of bands discussed so far. Of course, there are many interrelations and hybrid bands that fall in between these groups. As much as the lattice-design bands seem to be different from all of the others, the vertical tituli bands of the Neustift mitre in New York mentioned

4.  The so-called ‘Palermo bands’ and their technique

61

Fig. 4.12. Band with large figurative elements; mitre from St Peter’s abbey in Salzburg. Twelfth/ thirteenth century. Riggisberg, Abegg-Stiftung, inv. no. 233. (© Abegg-Stiftung, photo: Christoph von Viràg)

above combine them with small vegetal and animal motifs, such as tree-like scrolls or a quadruped with its head turned and facing the viewer (Blöcher 2012, 286–88, Borte 2).5 There are also bands that merge the typical elements of groups one and two, for instance the magnificent orphrey bands on the so-called chasuble of St Wolfgang from the monastery of St Emmeram in Regensburg (Domschatzmuseum Regensburg, inv. no. L 1982/3; Herrmann 2002). The monochrome patterned silk fabric from which the garment is made dates from the mid-twelfth century (Schorta 2001, 296–98) which fits quite well with the mitre bands of these groups. While all of the bands discussed so far seem to date from between the mid-twelfth and the early thirteenth century, a fifth group might be assigned a later date. The bands from this group all show the 1/6 twill binding of the lancé weft already described, while

62

Regula Schorta

Fig. 4.13. Band with bestiary in a framework of interlinked medallions. Twelfth/thirteenth century, stitched to the shoes of the imperial vestments. Vienna, Kunsthistorisches Museum, Imperial Treasury, inv. no. WS XIII 13. (© KHM-Museumsverband)

the ground weft is often linen, rather than the usual silk. A prominent representative of this group is the mitre in St Michael’s Church in Bamberg, allegedly attributed to bishop Otto I of Bamberg (d. 1139), but actually dating to the late thirteenth century (Blöcher 2012, 189–92). This would correspond with the St Valerius vestments in Barcelona, which were donated to the cathedral of Roda de Isábena in Huesca, Spain, around 1275 (Otavsky & ‘Abbās 1995, 192–93), whose chasuble was decorated with such a band (Flury-Lemberg & Illek 1995, 59–62; Vial 1995).

Lampas bands One further technical variation must be mentioned. Instead of using one out of the three or four warp colours of the Palermo band structure for tying down the lancé gold weft, a binding warp is added (Fig. 4.15). Such a band has, for instance, come to light among the textiles found in the ossuary of Bamberg Cathedral (Diözesanmuseum, Bamberg, inv. no. M54; Schorta 1985, 82, 90, 183), others were excavated in the cathedral of Basel (Historisches Museum Basel, inv. nos. HMB 2000.106.5.a–b, 2000.106.8 & 2000.105.4.bl.; Bayer & Schorta 2013, 177–78). Their structure consists of a 3/1 Z twill for the foundation weave, a warp in three colours, the typical 4/0 S mock twill

4.  The so-called ‘Palermo bands’ and their technique of the animal, and a 1/3 Z twill binding of the gold lancé weft. Where the gold weft is lost, the loops that once bound it still protrude very clearly. Counting the number of warp ends binding the gold and the number of warp ends forming the foundation twill gives the warp proportion of one binding warp end to nine main warp ends. This is definitely a lampas weave. Another example of such a Palermo lampas-type band is present on the mitre worn by the corpse of the fourteenth-century bishop of Basel, Johann Senn von Münsingen (d. 1365) (Historisches Museum Basel, inv. no. HMB 1975.184.12.m–r; Bayer & Schorta 2013, 195–97). For burial, he has obviously been dressed with an almost two-hundred-year-old headgear whose relationship to the Palermo bands is evident, even if a certain stiffness and awkwardness in the design cannot be denied. Addressing the question of how often, when and where Palermo-type lampas structures appear is a task for future research. Ordinary twill-based lampas bands, with either silk or gold lancé weft, are not rare, and they have long been considered products of German looms. Many are woven to shape for stoles and maniples, some with patterns of loosely placed elements reminiscent of the Palermo-type bands, others with medallions enclosing quadrupeds or birds (for an overview see von Wilckens 1991, 97–103). They have also been used side by side with Palermo type bands in one vestment – for instance, the fanons of the imperial mitre of Emperor Henry VI (Granger-Taylor 1994), whose horizontal circulus band, as mentioned

63

Fig. 4.14. Band with pattern of lozenges; lappets of a mitre. Thirteenth century. Disentis, Switzerland, Benedictine abbey. (Photo: Heidi Blöcher, Berlin)

Fig. 4.15. Fragment of Palermo-type gold band with lampas weave structure, reverse with visible binding warp floats. Twelfth/thirteenth century. Bamberg, Diözesanmuseum, inv. no. M54. (Bayerisches Landesamt für Denkmalpflege, photo: Hannelore Herrmann)

64

Regula Schorta

before, is a Palermo-type band of the second pattern group. A particularly beautiful maniple with a long, woven inscription mentioning a certain Alebertus entered the Victoria and Albert Museum from the Bock collection (V&A 8588-1863; Pregla 2009). Its further provenance is unknown, but to judge from its bead-embroidered endpieces, one might think of one of the nunneries in Lower Saxony.6 It is peculiar with regard to its technique, which is a lampas weave in the middle and side sections and a simple twill in the dividing stripes with geometric motifs. The metal thread of the lancé weft is similar to that of the Palermo-type bands, but it is always doubled, and it lies on the back of the band where it is not used on the front, bound by the binding warp in the lampas sections and floating unbound in the twill parts. Even if their arrangement is very different, the main warp’s colours – red, green, white and violet – are comparable to the Palermo-type bands, and so are the pattern elements, quadrupeds, birds and stylised floral elements. Inscriptions also occur along the sides of Palermo bands, for instance on the girdle for the imperial sword. It is not clear whether there is a relationship between these various lampas bands and the Palermo-type bands – and if so, of what kind. They are a strong indication, however, of a flourishing band-weaving tradition in Central Europe. This tradition very probably also encompassed bands of the Palermo type, as is suggested by an 18.5 cm wide and 126 cm long orphrey band in the Victoria and Albert Museum (Figs  4.16 & 4.17) (V&A 1250-1864; Cole 1911, 451 with fig.  39b). It was also acquired from Franz Bock, has long been accessible in the old study collection, but has never been photographed until now, perhaps due to its mediocre state of preservation. Between two narrow side sections, with the usual Palermo type animal and vegetal motifs, there is a broad middle section with six New Testament scenes, one on top of the other. From top to bottom, these are the Crucifixion,

Fig. 4.16. Orphrey band with New Testament scenes and woven inscription ODILIA ME FECIT. Lower Saxony, around 1200. London, V&A Museum, acc. no. 1250-1864. (© Victoria and Albert Museum, photo: Silvija Banić)

4.  The so-called ‘Palermo bands’ and their technique

65

Ecclesia and Synagogue, the three Marys at the grave, Christ triumphant, the Annunciation, and four angel’s heads to the sides of a star, probably part of the Nativity. The weave is a typical, non-lampas Palermo band 4/1 S-twill woven with a warp striped in red, green, white and blue-violet and with the gold lancé weft bound in 1/4 Z-twill. The inscription ODILIA ME FECIT, at the lower right is in beautiful large letters in red on gold. The name of the weaver7 indicates German rather than Sicilian provenance,8 and also many iconographical details link the scenes to the art of Lower Saxony around 1200, as it is known from manuscript illumination, in particular. Bands with biblical multi-figured scenes are not very common, but some do exist (von Wilckens 1991, 101). A rather broad stole with New Testament scenes

Fig. 4.17. Crucifixion; detail of orphrey band in Fig. 4.16. (© Victoria and Albert Museum, photo: Silvija Banić)

66

Regula Schorta

has been found in a Trier bishop’s grave (Braun 1907, 591 fig.  279; Groß-Morgen 2004, 31), and a beautiful red and gold set of stole and maniple was in Utrecht (Het Catharijneconvent, inv. nos. OKM t92a & b),9 but has unfortunately been lost (von Wilckens 1991, 100, fig. 106). Technically, these bands are described as lampas weaves, comparable to the silk lampas bands mentioned before. With regard to both style and technique, they are more distant from the Palermo-type band with the name of Odilia.

Conclusion These comparisons with loom-woven lampas bands that are possibly German bring the subject back to the starting point: the designation of the bands with the fascinating twill structure as ‘Palermo bands’. As mentioned before, Otto von Falke argued for their Sicilian origin using stylistic and, above all, historical arguments. Two such bands are, indeed, part of the secular treasury of the Holy Roman Empire, which is found today in Vienna: the girdle for the imperial sword and the trimming of the coronation shoes. Neither, however, can be traced farther back in history than to the inventories of the fifteenth century. There is no proof that they were already part of the Norman set of vestments, and it is noteworthy that the seam trimming of the coronation mantle, whose embroidery is undoubtedly a product of the Sicilian court workshop, is a beautiful tablet weaving, unrelated to Palermo-type bands. The secular mitre from the grave of Emperor Henry VI (d. 1197) in Palermo, by contrast, combines several types of bands (Granger-Taylor 1994). The Palermo-type band of the horizontal circulus has been mentioned before, and so has the lampas band of the lappets. Both are framed by narrow tablet-woven braids, while a fourth band, now lost, once formed the vertical titulus. While nothing is known about its technique, a careful drawing of it was published in 1784: it carried an Arabic inscription bearing good wishes – nothing astonishing at the Sicilian court under Norman rule, but thus far unparalleled, and, of course, a very strong argument for the Sicilian production of these bands (Daniele 1784, 45–46 & pl. H). According to the imprint of the band on the mitre’s silk fabric, it was only around 3 cm wide – which would allow for medallions of a maximum diameter of 2.5 cm. This is minute, even for the finest of the Palermo-type bands, and it is at least doubtful whether it would have been possible to weave an Arabic inscription of this length into such a tiny medallion. The description of the band, as published in 1784, does not tell us anything more than that there were ‘certi scudetti rotondi, ne’ quali son queste parole cufiche’, i.e. that there were certain round shields, or shapes, within which these cufic words were found (Daniele 1784, 46). There is no mention whatsoever of whether these words were woven, embroidered, or perhaps executed in yet another technique or even material. To my knowledge, no other woven band of the Palermo type with an Arabic inscription has been described so far. It is impossible to attribute

4.  The so-called ‘Palermo bands’ and their technique

67

the whole group of bands, which are so clearly defined by their weaving technique, to the workshops of Palermo based on this band of unknown technique and its supposed Arabic inscription. Rather than assuming that Sicily was the place of production of all these bands, their provenance should be discussed individually. Despite the very uniform technique, there might have been several centres where they were woven. As is so often the case, closer scrutiny does not result in an answer, but in a whole set of new questions. It seems that only the careful and very precise technical analysis of these bands – and other, related structures – will perhaps lead to a better understanding of their interdependencies, affiliations and relationships. In the meantime, ‘Palermo-type bands’ might perhaps be used as a moniker, for convenience and out of established habit – so long as it is not taken literally, as a designation of provenance.

Acknowledgements The author thanks Barbara Pregla, Halle/Saale, for most helpful discussions and for allowing insight in her manuscript of the catalogue raisonné of the textile treasures of Halberstadt cathedral, which includes a chapter about these bands. She also thanks Frances Pritchard for thoughtfully editing the text, the anonymous reviewer for useful suggestions, and Silvija Banić for providing photos of the Odilia band. Notes

1 Three fragments have been preserved: The Cleveland Museum of Art, acc. no. 1939.46 (purchase from the J.H. Wade Fund); Victoria & Albert Museum, no. 8569-1863; Vienna, MAK – Museum für Angewandte Kunst, inv. no. T 752. 2 This band is preserved in three parts: Riggisberg, Abegg-Stiftung, inv. no. 304; The Cleveland Museum of Art, acc. no. 1931.444 (gift of Dr. C.S. Reber); Cologne, Museum Schnütgen, inv. no. N 321. 3 Von Falke argues on the basis of analogies in the motifs and an occasional shared awkwardness that seems to relate the bands to tapestries with golden ground, such as the partial lining of the coronation mantle in Vienna, Kunsthistorisches Museum, Imperial Treasury, inv. no. WS XIII 14. 4 Current location and museum number are given only for objects not in Blöcher 2012. 5 The gold background in this case is not bound in simple twill but in a small lattice pattern based on twill. 6 Kroos 1970, 33 with note 47, cautiously suggests a provenance from St Marienberg in Helmstedt, because the oldest source for part of the text of the inscription is a hymn in a breviary from that monastery. 7 The formula – with the band itself speaking about its maker – points to a weaver’s inscription rather than a donor’s. Claussen 1985, 265. 8 The name gained popularity in the Alsace area with the rise of the cult of St Odile (Ottilia) in the fourteenth and fifteenth century, but there are also records from twelfth-century northern Germany. 9 For details see http://adlib.catharijneconvent.nl/ais54/Details/collect/26977, last accessed July 30, 2020.

68

Bibliography

Regula Schorta

Bayer, A. & Schorta, R. 2013. Katalog der Gräber und Grabinventare [all entries on textiles]. In H.R. Meier & P.-A. Schwarz (eds), Die Grabfunde des 12. bis 19. Jahrhunderts aus dem Basler Münster. Repräsentation im Tod und kultureller Wandel im Spiegel der materiellen Kultur. Materialhefte zur Archäologie in Basel 23 (Basel, Archäologische Bodenforschung des Kantons Basel-Stadt), 135–238. Blöcher, H. 2012. Die Mitren des Hohen Mittelalters (Riggisberg, Abegg-Stiftung). Branting, A. & Lindblom, A. 1929. Medeltida vävnader och broderier i Sverige, vol. 2 Utländska arbeten (Uppsala, Stockholm, Almqvist & Wiksell). Braun, J. 1907. Die liturgische Gewandung: Im Occident und Orient nach Ursprung und Entwicklung, Verwendung und Symbolik (Freiburg i. Br., Herdersche Verlagsbuchhandlung). Claussen, P. 1985. Künstlerinschriften. In A. Legner (ed.), Ornamenta ecclesiae. Kunst und Künstler der Romanik. Exhibition catalogue Cologne, Josef-Haubrich-Kunsthalle (Cologne, SchnütgenMuseum), vol. 1, 263–76. Cole, A.S. 1911. Weaving. Archaeology and art. In The Encyclopædia Britannica, 11th ed. (Cambridge, Encyclopædia Britannica), vol. 28, 448–55. Collingwood, P. 1982. The Techniques of Tablet Weaving (London, Faber and Faber). Daniele, F. 1784. I regali sepolcri del Duomo di Palermo (Naples, Stamperia del Re). De Jonghe, D. 1991. De Textieldocumenten uit Sint-Truiden. Technologische Bevindingen. In E. Deconinck (ed.), Stof uit de Kist. De middeleeuwse textielschat uit de abdij van Sint-Truiden (Leuven, Peeters), 63–105. Defoer, H.L.M. 2004. De zogenaamde albe van Bernulphus, een onderdeel van een keizerlijk ornaat. In J.-C. Klamt (ed.), Representatie. Kunsthistorische bijdragen over vorst, staatsmacht en beeldende kunst, opgedragen aan Robert W. Scheller (Nijmegen, Uitgeverij Valkhof Pers), 111–24. Flury-Lemberg, M. 1988. Textilkonservierung im Dienste der Forschung. Ein Dokumentarbericht der Textilabteilung zum zwanzigjährigen Bestehen der Abegg-Stiftung. Schriften der Abegg-Stiftung 7 (Riggisberg, Abegg-Stiftung). Flury-Lemberg, M. & Illek, G. 1995. Spuren kostbarer Gewebe. Riggisberger Berichte 3 (Riggisberg, Abegg-Stiftung). Granger-Taylor, H. 1994. [Cat. no.] 190 Royal mitre from the tomb of the Holy Roman Emperor Henry VI. In D. Buckton (ed.), Byzantium. Treasures of Byzantine Art and Culture. Exhibition catalogue London, British Museum (London, British Museum Press), 172–75. Groß-Morgen, M., Rothbrust, B. & Weber, W. 2004. Zu ewigem Gedächtnis und Lob. Die Grabstätten der Trierer Bischöfe in Dom und Liebfrauen. Exhibition catalogue Trier, Bischöfliches Dom- und Diözesanmuseum. Museumsführer 3 (Trier, Bischöfliches Dom- und Diözesanmuseum). Herrmann, H. 2002. Beobachtungen zur Technik von ‘Palermitaner’ Goldborten an der sogenannten Wolfgangskasel des 12. Jahrhunderts aus Regensburg. In S. Martius & S. Ruß (eds), Historische Textilien. Beiträge zu ihrer Erhaltung und Erforschung. Veröffentlichung des Instituts für Kunsttechnik und Konservierung im Germanischen Nationalmuseum 6 (Nuremberg, Verlag des Germanischen Nationalmuseums), 43–50. Hofenk de Graaff, J.H. 2004. The Colourful Past. Origins, Chemistry and Identification of Natural Dyestuffs (Riggisberg, Abegg-Stiftung & London, Archetype Publications). Keupp, J. 2014. Des Kaisers alte Kleider. Zum Kontext herrscherlicher Textilgeschenke im Hochmittelalter. In I. Siede & A. Stauffer (eds), Textile Kostbarkeiten staufischer Herrscher. Werkstätten – Bilder – Funktionen (Petersberg, Michael Imhof Verlag), 67–76. Kroos, R. 1970. Niedersächsische Bildstickereien des Mittelalters (Berlin, Deutscher Verlag für Kunstwissenschaft). Müller-Christensen, S. 1955. Sakrale Gewänder des Mittelalters. Exhibition catalogue Munich, Bayerisches Nationalmuseum (Munich, Hirmer).

4.  The so-called ‘Palermo bands’ and their technique

69

Nobiles Officinae 2004. W. Seipel (ed.), Nobiles Officinae. Die königlichen Hofwerkstätten zu Palermo zur Zeit der Normannen und Staufer im 12. und 13. Jahrhundert. Exhibition catalogue Vienna, Kunsthistorisches Museum (Vienna, Kunsthistorisches Museum). Otavsky, K. & ‘Abbās Muḥammad Salīm, M. 1995. Mittelalterliche Textilien I. Ägypten, Persien und Mesopotamien, Spanien und Nordafrika. Die Textilsammlung der Abegg-Stiftung 1 (Riggisberg, Abegg-Stiftung). Otavský, K. & Wardwell A.E. 2011. Mittelalterliche Textilien II. Zwischen Europa und China. Die Textilsammlung der Abegg-Stiftung 5 (Riggisberg, Abegg-Stiftung). Pregla, B. 2009. [Cat. no.] IV.8 Manipel mit Goldborte und Perlstickerei. In M. Puhle (ed.), Aufbruch in die Gotik. Der Magdeburger Dom und die späte Stauferzeit. Exhibition catalogue Magdeburg, Kulturhistorisches Museum (Mainz, Philipp von Zabern), vol. 2, 147–50. Schorta, R. 1985. Anmerkungen zur Technik einzelner Gewebe der Bamberger Grabfunde. In Textile Grabfunde aus der Sepultur des Bamberger Domkapitels. Colloquium, Bamberg, Schloss Seehof, 22–23 April 1985 (Munich, Lipp), 80–90. Schorta, R. 2001. Monochrome Seidengewebe des Hohen Mittelalters. Untersuchungen zu Webtechnik und Musterung (Berlin, Deutscher Verlag für Kunstwissenschaft). Schorta, R. & Wilhelmy, W. 1998. [Cat. no.] 43 2. Mitra. In H.-J. Kotzur (ed.), Hildegard von Bingen 1098–1179. Exhibition catalogue Mainz, Bischöfliches Dom- und Diözesanmuseum (Mainz, von Zabern), 123–27. Vial, G. 1995. Chasuble de Saint Valère, bordure. In Flury-Lemberg & Illek 1995, 237–41. von Falke, O. 1913. Kunstgeschichte der Seidenweberei, 2 vols. (Berlin, Wasmuth). von Wilckens, L. 1991. Die Textilen Künste. Von der Spätantike bis um 1500 (Munich, Beck).

Section II Sprang

Chapter 5 Hairnets with gold tube beads from the Roman Rhineland and their textile technique Petra Linscheid

Introduction It was in 1994 that I met Peter Collingwood during the Early Textiles Study Group conference in Manchester. Wondering about some non-woven structures, Peter opened my eyes to the fascination of the sprang technique. Some months later, I was fortunate to take part in a course to learn sprang plaiting from Peter’s close friend and colleague Noémi Speiser in Richterswil, Switzerland. Since that time, the sprang technique holds a permanent place in my research, starting with the beautiful sprang headcoverings from Early Byzantine Egypt. In the meantime, researching on excavated local textiles from the Rhineland, the evidence of sprang is much more difficult to find, since northern European textiles tend to survive just in minute traces. I often miss Peter Collingwood’s expertise in identifying the structures that I see. Hairnets were a popular headcovering for women in the Roman period. The most precious examples were decorated with golden elements, in most cases gold tube beads. In the western Roman provinces, mostly the non-organic parts of the hairnets survive. The following paper aims to consider in which textile technique these hairnets were produced. Starting from my own investigations on a hairnet from RheinbachFlerzheim, three other hairnets from the Roman Rhineland will be introduced. Then I will turn to comparable finds from other western Roman provinces, focusing on the question of textile technique.

The hairnet from Rheinbach-Flerzheim In 1982 and 1985, the office for the preservation of monuments in the Rhineland (Landschaftsverband Rheinland [LVR]  – Amt für Bodendenkmalpflege) excavated

74

Petra Linscheid

Fig. 5.1. About 250 gold tube beads from Rheinbach-Flerzheim tomb 7. (© LVR-LandesMuseum Bonn, photo: P. Linscheid)

a group of tombs, belonging to a nearby Villa Rustica at Rheinbach-Flerzheim, district of Rhein-Sieg, 30 km south of Cologne. The tombs, which date to the second to third centuries AD, contained rich grave goods (Gechter 1992). Three of the tombs preserved textile remains, including a gold thread tapestry (Linscheid 2011a, 70−72). In the cremation grave no. 7, belonging to a girl aged 14−16 and dated by the archaeologist Constanze Lösch to the first quarter of the third century AD (Lösch 2011), approximately 250 gold tube beads were preserved (Fig. 5.1). The find is housed today in the LVR-LandesMuseum Bonn with the inventory number 82.275. Presumably the gold tubes had been placed in a now deteriorated wooden casket together with

5.  Hairnets with gold tube beads from the Roman Rhineland and their textile technique 75 other grave goods. They consist of gilded bronze foil and were recognised by Constanze Lösch as parts of a hairnet (Lösch 2012). The find was recently investigated by the author in the framework of the project, Textile Bodenfunde im Bestand des LVR-LandesMuseums Bonn, funded by the Ministry of municipal affairs Northrhine-Westfalia. The research aimed to find traces of threads in order to understand the textile construction Fig. 5.2. Single gold tube bead from Rheinbachof the hairnet. Flerzheim. (© LVR-LandesMuseum Bonn, photo: The c. 250 gold tube beads measure R. Vogel) 1.0−1.5  mm in width and 7.00  mm in length (Fig.  5.2). In addition there is a smaller group measuring just 5 mm in length. The tubes consist of a rectangular gilded bronze foil, rolled up along the short edge into a longitudinal cylinder, with the edges slightly overlapping. Deposits cover most of the gold surface, which is visible in just a few places. The original position of the gold tubes in the tomb was not recorded, but in several cases two or three tubes are adhering to each other in a parallel, staggered alignment (Fig.  5.3). This position reveals the original arrangement of the gold tube beads on the net. A few tubes preserved traces of Fig. 5.3. Gold tube beads in parallel staggered threads due to metal corrosion. The alignment. (© LVR-LandesMuseum Bonn, photo: P. Linscheid) tubes were threaded on a plied thread, consisting of three Z-spun yarns plied in S-direction (Fig. 5.4). The diameter of the plied thread is 0.5 mm. The bronze corrosion gave the thread a greenish colour. The raw-material could not be determined. The same plied thread appears running parallel alongside and between the gold tube beads (Fig. 5.5). This means that in the hairnet bare textile parts were visible and the tubes were positioned just in special areas of the net. Single fragments show several plied threads running parallel and crossing or interlinking each other. The parallel alignment and staggered arrangement of the gold tube beads would match perfectly with an interlinked sprang structure of the textile net.

76

Petra Linscheid

The reconstruction with gold tubes threaded on the sprang warp threads shifted them exactly to this position (Fig. 5.6). Minute fragments of white glass mixed with the tubes might have belonged to glass beads, alternating with the gold tube beads in special areas. In between the gold tubes 10 fragments of a gold thread could be observed, with a thread diameter less Fig. 5.4. Plied thread from the hairnet. (© LVR- than 0.1 mm, consisting of a gold wire, LandesMuseum Bonn, photo: R. Vogel) wound spirally in Z-direction around a lost core thread (Fig.  5.7). The gold thread obviously served as the weft in a narrow tabby-woven band, 1.5 cm in width. The mixing of the gold thread with the gold tube beads suggests, that the gold band belonged to the hairnet. Since the hairnet was found in a cremation grave, there is no indication of its position on the woman’s body. Even the original shape and dimensions of the hairnet cannot be determined. The band made from gold thread instead would serve best as a loose band for fastening or as a browband, indicating that the net had been made up in the shape of a receptacle and was not used as a rectangular veil. An amber hairpin, found among the grave goods, might have served to fix the hairnet in place on the woman’s head.

The hairnet from Rommerskirchen The closest comparison to the RheinbachFlerzheim hairnet is a find at Rommerskirchen, a site situated near Cologne, about 50 km northeast of Rheinbach-Flerzheim. The LVR-Amt für Bodendenkmalpflege im Rheinland discovered a group of tombs belonging to a Villa Rustica near Rommerskirchen in 2005 (Gechter 2007). Two coffins were lifted and transported to the conservation workshops of the Landesmuseum Bonn to document and retrieve the content of the sarcophagus under optimal conditions.

Fig. 5.5. Plied thread alongside and between the gold tube beads. (© LVR-LandesMuseum Bonn, photo: P. Linscheid)

5.  Hairnets with gold tube beads from the Roman Rhineland and their textile technique 77 One of the tombs contained a lot of grave goods, mainly glassware, which dated the cremation burial to the second quarter of the third century AD. The textile remains, including a blue and golden tapestry weave, were conserved and studied by Annemarie Stauffer (Stauffer 2011). From the same tomb, the remains of a hairnet were retrieved, which were conserved and investigated by Gabriele Schrade from the Cologne University of Applied Sciences (Schrade 2007; Schrade 2011; Schrade 2013). From this net a total of 46 gold tube beads survived, each one measuring 6  mm in length, consisting of a small bronze foil with gilding, rolled up in a cylindrical shape. The 46 tubes covered an area of c. 18 × 18 cm on the bottom of the sarcophagus and therefore must have decorated the main part of the net. White and blue glass beads and metal plates made up a browband, which was fixed to the net. The existence of a browband and the concentration of the gold tube beads in a limited area make it plausible that the net was in the shape of a receptacle and not a veil. There were few textile remains. Gabriele Schrade identified threads from wool or fine animal hair, used as S-plied thread from two Z-spun yarns. The textile construction technique was hard to determine, but G. Schrade observed linked threads (Fig. 5.8) resembling the interlinking of sprang technique. She explicitly excluded a knotting technique (Schrade 2011, 56). The find included four bone and one metal hair pin to fix the hairnet on the woman’s head.

The hairnet from Erkelenz-Borschemich

Fig. 5.6. Reconstruction of gold tube beads in interlinked sprang. (© P. Linscheid)

Fig. 5.7. Fragments of band made from In 2013 a group of tombs of the first to second gold thread. (© LVR-LandesMuseum centuries AD was excavated belonging to a Villa Bonn, photo: R. Vogel) Rustica in Erkelenz-Borschemich, district of Heinsberg, about 50 km northwest of Cologne. In the cremation burial IV of a woman, dated to the early second century AD, many precious grave goods were found (Schuler 2017). The remains of gold thread from a tapestry and golden tubular beads from a hairnet were conserved and investigated by Annemarie Stauffer (Stauffer 2017).

78

Petra Linscheid

Fig. 5.8. Hairnet from Rommerskirchen, linked threads. (© LVR-Amt für Bodendenkmalpflege im Rheinland, photo: G. Schrade)

The hairnet was decorated with at least 284 gold tube beads, measuring 2.0−4.5 mm in length and 1 mm in width. In contrast from the gilded bronze tubes in RheinbachFlerzheim and Rommerskirchen, the tubes in Erkelenz possessed a high gold content (88%) providing an impressive lustre. The gold tube decoration was enriched by additional gold beads. Unfortunately, there is no documentation of the finds in situ providing evidence for the original position of the tubes and beads, and no textile remains were preserved from which to deduce the textile technique for making the hairnet.

The hairnet from Krefeld-Gellep Since the 1960s, the large Roman-Frankish cemetery of Krefeld-Gellep, situated 60 km north of Cologne, is being excavated. The cremation burial no 3639, dated to the second century AD and belonging to a woman 20−24 years old, contained gold threads and gold tube beads which were conserved and investigated by HansJürgen Hundt at the Römisch-Germanisches Zentralmuseum Mainz (Hundt 1989). Approximately 2500 gold tubes were preserved measuring 2.9−3.7  mm in length

5.  Hairnets with gold tube beads from the Roman Rhineland and their textile technique 79

Fig. 5.9. Gold tube beads from Krefeld-Gellep. (© Museum Burg Linn, Krefeld, photo: Bildarchiv Römisch-Germanisches Zentralmuseum Mainz)

and 0.3−0.7 mm in width (Fig. 5.9). They were constructed from a rectangular gold or gilded foil sheet rolled into a cylindrical shape. Analysis of the gold content has not been undertaken. Hans-Jürgen Hundt recognised the tubes as coming from a haircovering or veil, comparing it to a find in Cádiz (see below). No threads remained with the gold tubular beads to provide evidence for the textile technique of the hairnet. To sum up: in at least four women’s tombs of the Roman Rhineland, dating to the first to third centuries AD, hairnets decorated with gold tube beads were found. They all belonged to cremation burials, which means that they lack information about the position of the find on the body. The identification as headcoverings is based on comparable finds in Roman Italy and Spain (see below). Annemarie Stauffer has pointed out the different qualities of the tubular beads: while in Erkelenz-Borschemich they consist of gold of a high purity, in RheinbachFlerzheim and Rommerskirchen a cheaper material, namely gilded bronze, was used (Stauffer 2017, 82). Another difference might be in the quantity of gold decoration used in the hairnets, varying between 46 gold tube beads in Rommerskirchen and

80

Petra Linscheid

2500 in Krefeld-Gellep. A smaller number of golden tubes means that the hairnet showed larger areas of textile ground. Nevertheless, thousands of gold tubes probably decorated a large part of the hairnet’s surface. The textile technique, in which the hairnets in the Rhineland were produced, could be traced only in two examples: the hairnets in Rheinbach-Flerzheim and Rommerskirchen were probably made in sprang technique.

Hairnets with gold tube beads from Roman Italy The fashion centre for hairnets with gold tube beads was located in first to third century AD Roman Italy, as indicated by the large number of finds in that province. Five examples have been unearthed so far in the city of Rome: in tombs 2, 3 and 4 from the necropolis of Vallerano, in tomb A at Via di Tor Sapienza and tomb 12 at Via delle Vigne Nuove (Bedini, Rapinesi & Ferro 2004, 85−86; Rapinesi et al. 2014, 60−67). In Tomb 2 of Vallerano c. 13,000 gold tubes were retrieved, measuring 1.0–8.0 mm in length, together with 128 different types of gold beads and a band made from gold thread. In tombs 3 and 4 of Vallerano some dozen gold tubes were observed, with the burial at Via di Tor Sapienza 51 examples and in the tomb at Via delle Vigne Nuove c. 500 gold tubes were found. The tubes measured 3.0−8.0 mm in length and were partly combined with other types of gold beads. From tomb 31 of the northern cemetery of Albenga in north-west Italy, a cremation burial of the first century AD, 800 gold tubes were retrieved together with two other types of beads (Rapinesi, Ferro, Bedini & Giuliani 2014, 67). An important find is the hairnet from Milan, belonging to an inhumation burial situated in the territory of the Universita Cattolica, dated to the first half of the third century AD (Maspero & Rottoli 2005). This hairnet was embellished by 3096 gold tubular beads 2.0–5.0 mm in length combined with 173 other types of beads. The position of the tubes around the cranium of the burial revealed, that the hairnet in this example had the shape of a receptacle, a bonnet, and not a veil. With the hairnets from Rome and Albenga, no trace of threads survived to provide evidence for the textile framework of the hairnet. In the Milan hairnet, however, the remains of a plied thread from three yarns was observed, running through the tubes. The plied thread with a diameter of 0.4−0.5  mm was of undeterminable material. In what textile technique these threads were used to create the net could not be determined. The position and orientation of the beads in the area of the woman’s head provided no regularity either, which would help to identify the textile technique (Maspero & Rottoli 2005, 76, fig. 5−6). A tubular bead with a continuous length of bare thread gives the impression that the hairnet had visible textile areas and was not entirely covered by gold tube beads (Maspero & Rottoli 2005, 68). Unfortunately, for none of the hairnets with gold tube beads found in Italy is the textile technique known. It might have been the sprang technique, though, because there is evidence for this technique from a splendid hairnet plaited in sprang

5.  Hairnets with gold tube beads from the Roman Rhineland and their textile technique 81

Fig. 5.10. Fragments of the hairnet from Cádiz. (© Carmen Alfaro Giner, photo: Bildarchiv RömischGermanisches Zentralmuseum Mainz)

exclusively from gold thread (Linscheid 2011b, 80, fig. 31; Rapinesi et al. 2014, 65, no. 5). This hairnet, dated to the first century AD, was probably found at Rome.

The hairnet with golden tubular beads from Roman Medina Sidonia in Spain Before 1979, fragments of a hairnet with gold tube beads were found inside a funerary urn of the first century AD in Medina Sidonia, today Cádiz in the South of Spain. This important find was investigated by Carmen Alfaro Giner (1983−84; 1984, 150−151 no. 142, pl. XXXVI, 4−5). The textile net was remarkably well preserved from having been in contact with a bronze mirror inside the urn. Seven net fragments survived, altogether covering an area of 27 × 16 cm (Figs 5.10 & 5.11). The net is produced with overhand knots in half hitch appearance (Alfaro Giner 1983–1984, 80, fig. 1; 1984, 149, fig. 137; cf. Emery 1966, 34, fig. 20), creating lozengeshaped meshes of 3.0−5.0 mm in size. The thread used is linen, Z-plied from two yarns, the latter with 0.1−0.15 mm thread diameter each. A total of c. 50 gold tubes, 3.0 mm in length and 1.0 mm in width are present in at least four of the net fragments in a

82

Petra Linscheid

Fig. 5.11. Fragment from Cádiz with gold tube beads. (© Carmen Alfaro Giner, photo: RömischGermanisches Zentralmuseum Mainz)

Fig. 5.12. Knotting technique and possible position of gold tube beads in the hairnet from Cádiz. (Drawing with additions: P. Linscheid, after Alfaro Giner 1983−84, 80, fig. 1)

linear arrangement. Obviously, the net was just partially adorned with gold tube beads, while large textile areas remained bare. The exact position of the gold tubes on the net is not clear. There are two possibilities (Fig.  5.12): the gold tubes were placed over the knots (Fig.  5.12. A), which means they were attached after the net was finished, or else the gold tubes were threaded on before netting started and they sit on the length of thread between the knots (Fig. 5.12. B). No original edges or traces of shaping are preserved on any of the fragments. Therefore, there is no evidence to reveal whether the net was made up into a receptacle or whether it was used in a rectangular shape as a veil.

Sprang and knotted hairnets from the Roman Period For Peter Collingwood, the primary interest in these hairnets would not be in the golden decorative elements, but in the textile technique in which the net was made

5.  Hairnets with gold tube beads from the Roman Rhineland and their textile technique 83 and in the way the beads were incorporated into the net fabric. As demonstrated above, the textile technique of most hairnets with gold tube beads is unknown, because the textile threads have deteriorated. For the Rhineland, there are two examples providing evidence: the hairnets from Rommerskirchen and from Flerzheim which were probably worked in the sprang technique. However, the hairnet from Cádiz in Spain was made by knotting. Therefore, Roman hairnets with gold tube beads and other golden decorative elements were produced in sprang as well as in knotting techniques. Hairnets with gold decoration were the luxury version of textile hairnets made from wool and linen. Wool and linen hairnets in the Roman period appear mostly in sprang technique and more rarely in knotting. The best evidence is provided by a wicker basket from Roman Hawara, in Egypt, found in a woman’s grave of the late second or early third century AD, containing three hairnets from linen, two of them in sprang and one in knotting technique. The find is housed today in the Petrie Museum of Egyptian Archaeology, University College London, inv. 28009 a–c (University College 2001; Linscheid 2011b, 78−79, figs 28−29, 81, fig. 32). The hairnets with gold tube beads in the Rhineland reveal that Roman fashions in headdress styles reached the province of Germania Inferior on the northern edge of the Roman Empire. More than that, the finds provide evidence for a textile technique which is otherwise not documented in the Roman Rhineland: the sprang technique.

Acknowledgement I am grateful to Carmen Alfaro for discussing the hairnet from Medina Sidonia with me. Furthermore I owe thanks to Gabriele Schrade for her information on the threads in the Rommerskirchen hairnet and to Susanne Willer, LVR-LandesMuseum Bonn, for her permission to publish the finds.

Bibliography

Alfaro Giner, C. 1983−84. Notas sobre una redecilla romana de Medina Sidonia (Cádiz). Boletín del Museo de Cádiz 4, 77−81. Alfaro Giner, C. 1984. Tejidos y Cestería en la Penínsola Ibérica (Madrid, Consejo Superior De Investigaciones Científicas). Bedini, A., Rapinesi, I. & Ferro, D. 2004. Testimonianze di filati ornamenti in oro nell’abbigliamento di età romana. In C. Alfaro, J.P. Wild & B. Costa (eds), Purpureae Vestes I. Actas del I. Symposium Internacional sobre Textiles y Tintes del Mediterràneo en època romana (Universitat de València), 77−88. Emery, I. 1966. The Primary Structures of Fabrics: An Illustrated Classification (Washington D.C., The Textile Museum). Gechter, M. 1992. Der römische Gutshof von Rheinbach-Flerzheim. In Landschaftsverband Rheinland, Rheinisches Amt für Bodendenkmalpflege, Rheinisches Landesmuseum Bonn (eds), Spurensicherung. Archäologische Denkmalpflege in der Euregio Maas-Rhein (Mainz, Zabern), 452−60. Gechter, M. 2007. Die mittelkaiserzeitlichen Gräber von Rommerskirchen. In LVR-Amt für Bodendenkmalpflege im Rheinland (ed.), Archäologie im Rheinland 2006 (Stuttgart, Theiss), 140–42.

84

Petra Linscheid

Hundt, H.-J. 1989. Zierelemente aus Goldfolie bzw. Goldlahn aus Grab 3639. In R. Pirling, Das römisch-fränkische Gräberfeld von Krefeld-Gellep 1966−1974. Germanische Denkmäler der Völkerwanderungszeit, Ser. B, 13 (Stuttgart, Steiner), 42−43. Linscheid, P. 2011a. Textile Bodenfunde der römischen Kaiserzeit und der Merowingerzeit im Bestand des LVR-LandesMuseums Bonn. In LVR-Amt für Bodendenkmalpflege im Rheinland (ed.), Textilien in der Archäologie (Treis-Karden, Print Concept), 69−79. Linscheid, P. 2011b. Frühbyzantinische textile Kopfbedeckungen. Typologie, Verbreitung, Chronologie und soziologischer Kontext nach Originalfunden. Spätantike, Frühes Christentum, Byzanz 30 (Wiesbaden, Reichert). Lösch, C. 2011. Ein außerordentlich reiches Mädchengrab des dritten Jahrhunderts aus RheinbachFlerzheim. Archäologische Informationen 34/1, 117−22. Lösch, C. 2012. International vernetzt – das zweite römische Prunkhaarnetz im Rheinland. In LVR-Amt für Bodendenkmalpflege im Rheinland (ed.), 25 Jahre Archäologie im Rheinland 1987−2011 (Stuttgart, Theiss), 120−22. Maspero, A. & Rottoli, M. 2005. Il microscavo e le analisi di laboratorio: metodologie e risultati. In M.P. Rossignani, M. Sannazaro & G. Legrottaglie (eds), La Signora del Sarcofago, Una sepoltura di rango nella necropoli dell’Università Cattolica. Contributi di Archeologia 4 (Milan, Vita e Pensiero), 55−81. Rapinesi, I.A., Ferro, D., Bedini A. & Giuliani, M.R. 2014. Elementi di oreficeria nelle acconciature femminili da rinvenimenti in sepolture nell’area di Roma. Approfondimenti sulla tecnologia di fabbricazione. In C. Alfaro Giner, J. Ortiz García & M. Antón Peset (eds), Tiarae, diadems and headdresses in the Ancient Mediterranean Cultures. Symbolism and technology, Monografías del SEMA de València III (València, Universitat de València), 51−67. Schrade, G. 2007. Haarnetz und Stirnband aus einem Frauengrab bei Köln. Archaeological Textiles Newsletter 45, 11−15. Schrade, G. 2011. Untersuchungen eines Kopfschmuckes aus einem mittelkaiserzeitlichen Steinsarg aus Rommerskirchen. In LVR-Amt für Bodendenkmalpflege im Rheinland (ed.), Textilien in der Archäologie (Treis-Karden, Print Concept), 53−60. Schrade, G. 2013. Analyses of Remains of a Headdress from a Late Roman Female Burial. In J. Banck-Burgess & C. Bübold (eds) NESAT XI. The North European Symposium for Archaeological textiles. Poster-Appendix (Rahden, Marie Leidorf). Schuler, A. 2017. Ein außergewöhnlicher römischer Bestattungsplatz bei Borschemich. Rheinische Ausgrabungen 75. (Darmstadt, Zabern). Stauffer, A. 2011. Antiker Luxus aus römischen Särgen. In LVR-Amt für Bodendenkmalpflege im Rheinland (ed.), Textilien in der Archäologie (Treis-Karden, Print Concept), 61−68. Stauffer, A. 2017. Die Goldtextilien. In A. Schuler, Ein außergewöhnlicher römischer Bestattungsplatz bei Borschemich. Rheinische Ausgrabungen 75 (Darmstadt, Zabern). 77−83. University College 2001. Digital Egypt for Universities, Hawara, Roman tomb group: https://www. ucl.ac.uk/museums-static/digitalegypt/hawara1/roman/girlgrave.html.

Chapter 6 Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt Anne Kwaspen and Kristin South

Introduction The ongoing excavations at Fag el-Gamous, an early Christian necropolis on the edge of the Fayum oasis in Egypt, continue to reveal well-preserved and extensively wrapped burials dating from the Roman into the early Medieval period, making this site an important source for the in situ study of archaeological textiles from the first millennium AD. Among the many articles of clothing found in the burials of Fag el-Gamous are multiple hairnets made in sprang technique. This study includes a description of the technical features of both the common hairnets with side seams and the more exceptional hairnets without side seams. Of particular note are the short hairnets without side seams found at Fag el-Gamous: to date, only one example of this exceptional type has been published from other venues (Selem & Al-Khalek 2010, 97), but at Fag el-Gamous there are an additional five short hairnets without side seams, four of which are from the same burial. The discovery of these five unusual hairnets prompted a study of their technical details and led to research into a plausible way of wearing this type of hairnet.

Hairnets at the necropolis of Fag el-Gamous Fag el-Gamous is an early Christian necropolis located on the eastern edge of the Fayum oasis, about 80 km south of Cairo. Excavations at the site, ongoing since 1981, have uncovered over 1200 burials, most of which have included at least tatters of textiles, and many are preserved well enough to display the full original burial kit. Of those with extensive textiles, most are wrapped in multiple layers of undyed linen with minimal wool decoration, but approximately 5% also include clothing items (tunics, socks, hairnets, scarfs and footwear) (South & Kwaspen 2021).

86

Anne Kwaspen and Kristin South

Fig. 6.1. Hairnet FeG 2005 SW 13 on skull. (© BYU Egypt Excavation Project)

Amongst the preserved burials, hairnets are found either packed between layers of wrappings or on the head of the deceased (Fig. 6.1). Those on the head may have been placed as they would have been used in daily life, but in some burials multiple hairnets are layered over the head. It is assumed that this is a burial practice and that it was not customary to wear more than one hairnet at the same time (Linscheid 2011, 62; Pritchard 2006, 131). In some areas of the necropolis, unauthorised digging has disturbed burials closer to the surface, and multiple hairnets have also been found in these disturbed areas, loose and out of context in the sand. No radiocarbon dating has yet been performed on headwear from Fag el-Gamous. However, the depth of the preserved burials (between 0.60 m and 1.50 m) indicates that the sprang hairnets at this cemetery can probably be dated from the fourth to eighth century AD, the same period as the published results of radiocarbon dating of sprang hairnets from many sites in Egypt in European museum collections (De Moor et al. 2002, 2732; De Moor et al. 2014, 103−20). Sprang hairnets were made in a variety of sizes in height and width (Linscheid 2011, 233−84). In every case where the sex of the burial could be determined through an examination of diagnostic features of the skull and pelvic bones, the hairnets of Fag el-Gamous are found exclusively on female burials. With a fairly large number of hairnets found in situ thus far, there are no exceptions to this rule. To work backwards from a hairnet to assume the physical sex of an individual is not acceptable,

6.  Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt

87

Fig. 6.2. Sprang structures. (© Anne Kwaspen)

nevertheless the use of hairnets seems to have been a regular feature of female gender identification in this society. It is generally assumed that only adult women wore this type of headwear, although some very small examples, which would have fitted better on the head of a younger girl, have been found (Kwaspen 2011, 71; Linscheid 2011, 54; James & Médard 2017, 150).

Technical features of sprang hairnets from Egypt

Sprang Sprang is the name given to a stretchy, net-like textile structure made on a vertical frame by one set of threads (warp) only. These threads are stretched around the upper and lower end of the frame. The structure is created by twisting the threads around each other, by pulling the back-layer threads to the front and dropping the frontlayer threads to the back (Collingwood 1974, 31−33; Hald 1980, 245−50; d’Harcourt 1959, 80−82; Van Reesema 1920−1921, 6−12). Of the three basic structures of sprang – interlinking, interlacing and intertwining (Frame 1986, 67−82)  – only the first two are used in the hairnets found at Fag el-Gamous (Fig. 6.2). Interlinking (spiral) is the most common structure. Some monochrome hairnets have a combined structure of interlinking and interlacing (oblique). The patterning of the polychrome hairnets is made by extra coloured threads that twine around the threads of the basic structure

88

Anne Kwaspen and Kristin South

and the technique of double interlinking (Collingwood 1974, 153−83; Kwaspen 2011, 70−95). The most typical aspect of sprang technique is that, because the threads are fixed at both ends, two parts of the structure are produced simultaneously in exact symmetrical mirror images at the top and at the lower end. When these two halves of the structure meet at the centre line on the frame, the threads are worked into a row of finishing chain stitches. To avoid the unravelling of the work it is sufficient to chain and fix one layer of threads, but in many instances the threads on the front layer are worked together and the threads on the back layer are also made into a second chain (Kwaspen 2014, 276). Material Among the hairnets excavated in Egypt, all are made in one of three combinations of wool and/or linen: the coloured hairnets have a sprang structure made in wool. Many other hairnets are made in natural-coloured linen. The third variation has the sprang structure created with a combination of wool and linen threads. This third type has not yet been recorded from Fag el-Gamous. Linen is abundantly present in the burial finds of Fag el-Gamous as a wrapping material (South 2017, 88−107). However, as a material for clothing, linen occurs much less frequently than wool (South & Kwaspen 2021). To date, this is also the case for the sprang hairnets. Only four of the 48 hairnets recovered were made entirely in linen (FeG 2005 SE 37; FeG 2006 SE 48; FeG 2009 NE 8.5; FeG 2009 NE 59). Unfortunately it was not possible to analyse these four linen hairnets in detail for this article. All sprang structures of the wool hairnets analysed from Fag el-Gamous have S-spun, Z-plied wool threads (S2Z). The use of S2Z thread is a very characteristic feature of the sprang hairnets from Egypt, although four hairnets made from Z-spun, S-plied threads were identified from Karanis in the collection of the Kelsey museum (Galliker & James 2019, 264). The threads and cords used for the other parts of the hairnets, such as the browbands and drawstrings, are also made from wool, although the yarn composition of these threads and cords varies. Shape and specific parts of the hairnets Egyptian hairnets differ according to how the completed sprang structure is finished. The first and most numerous group, both in wool and in linen, are hairnets with two side seams. Within this category, as described below, hairnets with seams can be made into either rectangular or cone-shaped hairnets. The second group consists of hairnets without side seams. These are much less common, although both wool and linen examples are known in this type. This second group can be subdivided into short hairnets and hairnets with two long tail panels at the back.

6.  Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt

89

Sprang hairnets with side seams Forty-three hairnets made with side seams have been found at Fag el-Gamous (Fig. 6.3). To create a hairnet of this type, the sprang structure is folded horizontally at the centre and the sides are placed next to one another and stitched so that the hairnets have no seam allowance. Closed with a stitch that appears very similar to the interlinking sprang, the seam itself is often scarcely noticeable. In other cases, however, the sewing is executed with coarse stitches using a bundle of S-spun threads in a contrasting colour. The half of the hairnet that corresponds to the upper half of the frame, with twists in Z-direction, becomes the back of the hairnet and the lower part, with twists in S-direction, becomes the front of the hairnet.1 The structure on the lower part of the frame is often packed more firmly, with the result that the front part of a hairnet is often shorter than the back. The 43 sprang hairnets with side seams Fig. 6.3. Different parts of a hairnet with seams. a = central part; b = decorative side from Fag el-Gamous can be classified not only border; c = decrease of width at the top; according to material used and shape, but d = warp loops on the back of the hairnet; also according to structure and decoration. e = browband; f = chained securing on the With regard to the interplay of structure and central line; g = chained ridge; h = drawstring decoration, a distinction is made between at top of hairnet; i = drawstring through monochrome hairnets and polychrome ones. back loops. (© Anne Kwaspen) The monochrome hairnets form the largest group with four made from linen and 17 from wool. Of the wool, nine are red, one brown, one purple, two blue and five yellow (or of natural, undyed wool colour). The decoration of these hairnets is done by a combination of sprang structures. This means a variation of twists in S- or Z-direction and dense or open structure. An open structure can be created with slits, obtained by not twisting certain threads in certain rows or by the technique of multiple twisting (Collingwood 1974, 95−102, 132−51; Kwaspen 2014, 278–81). Four monochrome hairnets from Fag el-Gamous have an open structure decoration.2 The others are patterned by alternating S- and Z-twists. Five hairnets have a particular pattern of this type, in which large diamonds are formed with smaller diamonds set into them vertically. The large diamond figures are always bordered with intersecting lines (Fig. 6.4).3

90

Anne Kwaspen and Kristin South

Fig. 6.4. Diamond pattern obtained by alternating twists in S-direction and Z-direction. (© BYU Egypt Excavation Project)

Fig. 6.5. Fragment of hairnet 2006 SE 44 B. (© BYU Egypt Excavation Project)

Eighteen hairnets from Fag el-Gamous are polychrome wool hairnets. All have one basic colour with decorative motifs formed by threads in other colours. Eight hairnets4 with red as their basic colour show a striking resemblance to the five monochrome hairnets with the diamond patterning. They have exactly the same pattern in the middle part of the structure. The difference is that these hairnets additionally have side borders with yellow and green threads that form spiral figures. The other ten polychrome hairnets have coloured decorative elements spread over the entire structure. The basic colour of these hairnets is blue (five), red (four) and purple (one) (Fig. 6.5).5 Three fragments are too small to distinguish their decorative design and colour use. The shape of hairnets with side seams, rectangular or cone-shaped, is determined by the method used to

6.  Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt

91

decrease the width at the top of the hairnet. During the production of the sprang structure, as the centreline is approached on the frame, threads are grouped for several rows to decrease the number of threads that have to twist. Twisting the threads becomes more and more difficult as the upper and lower parts meet (Collingwood 1974, 128). Rectangular hairnets have a minor reduction of the width on the top by several rows that are twisted with two threads paired together. For a cone-shaped hairnet, the reduction process is more extensive. After some rows (often two) Fig. 6.6. Decreasing number of two or four threads are taken together, the number of threads. (© Anne Kwaspen) threads taken together increases from four or six until sometimes more than 40 threads are twisted together to form a narrow, stiffened end or a tassel (Fig.  6.6). To date, only one cone-shaped hairnet of this type has been found among the textiles from Fag el-Gamous, a monochrome hairnet in natural, undyed wool with patterning obtained by twists in S and Z direction (FeG-2013-NW-51.1). The most frequently used reduction in width of hairnets from Fag el-Gamous is the first four to Fig. 6.7. Detail of hairnet FeG 1981 eight rows of two threads taken together followed by NE. (© BYU Egypt Excavation three to twelve rows of four threads bundled. This Project) results in a trapezoidal shape, somewhere between the rectangular and cone-shaped hairnets. They are classified here with rectangular hairnets due to their closer similarity in shape (Fig. 6.7). As mentioned above, the threads must be secured at the completion of the sprang process to avoid a total unravelling of the structure. As the centre of the frame is reached in the interlinking process, the threads must be worked in a row of chain stitches. The threads on the front layer are made into a chain together and the threads on the reverse layer are mostly made into a second chain. These two chains together form a tunnel in the centre of the fabric through which a drawstring can pass. In rectangular hairnets, the drawstrings can run in two different ways through the tunnel: either a single closed loop is formed on the top of the hairnet, or the drawstring makes a double loop, with one loop outside the hairnet and one inside (Kwaspen 2011, 90). Pulling the drawstring gathers the top of the hairnet together. The double loop easily holds this shape in place when the cord is pulled. With only one loop the cord must be knotted to give the same result. In some cases the central tunnel is too narrow for the drawstring, so for such hairnets the drawstring just runs along inside of the hairnet, or is inserted into the structure

92

Anne Kwaspen and Kristin South

Fig. 6.8. Fixing warp to a browband. (© Anne Kwaspen)

next to the chained ridge. In cone-shaped hairnets, the centre-line constricts to a point at the top of the stiffened end or tassel and the drawstring at the top is mostly a single loop. When warping the frame, the lower end of the warp can be fixed in two ways (Kwaspen 2014, 276−77) (Fig. 6.8). By far the more common practice is to set up the warp with a separately woven, plaited or sprang band that is tied onto the lower end of the frame. The warp threads are then directly attached to this band by a thread that makes a stitch through the band, then three or four warp threads are looped around this thread, which is next stitched again through the band and so on. The band determines the width of the hairnet and probably helped to keep the hairnet in shape while creating the sprang structure. The second method affixes the warp around a thin wool cord or even only a two-ply wool thread. The first twists of the interlinking sprang are packed next to this cord so the loop ends on the front are only the very small size of the cord. When finishing the hairnet, this sprang edge is also sewn onto a separately made band, with stitches going through this band and around the inserted cord. This separately woven, plaited or sprang band becomes a browband. When woven, the weave is warp-faced tabby. Of the 42 hairnets with side seams excavated at Fag el-Gamous only eight hairnets still have an extant browband. Four of these are warp-faced tabby-woven bands, three are braided, and one is a pile weave with long loops.6 In all cases, the warp of the sprang structure is fixed directly to these browbands. In general the end loops at the other side of the structure are between 1 and 1.5 cm long. Longer loops, up to 3 cm, can be grouped together with four to six loops and then worked in a chain (Kwaspen 2011, 90). The drawstrings on the back of the hairnet, inserted in these end loops, are always laced in the same way (Fig.  6.9). Knotted on one end of the browband, the drawstring goes through the back loops, further forming a big loop before disappearing again into the loops from the other side, to finish in a knot on the other outer end of the browband. In the case of two strings, rather than one big loop outside, it ends in two string ends. Due to the

6.  Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt

93

connection to the browband, the hairnet fits comfortably on the head when the back drawstring is pulled tight at the neck. Hairnets without side seams The hairnets without side seams can also be divided into two types, namely hairnets with two long tail panels at the back and a simple, short version without panels (Fig. 6.10). Unlike the hairnets with seams, where the shape is obtained by side seams to create a head-sized shape, hairnets without side seams are formed from the rectangular structure being made into a head-cover by gathering the sprang structure together using a single drawstring. The main feature of hairnets without side seams, therefore, is firstly the Fig. 6.9. Lacing drawstring at the back of the hairnet. (© Anne Kwaspen) gathering of the top or bottom of the sprang structure into a tight circle or a narrow U-shape. This shape is obtained by cinching the end loops of the structure together over the inserted drawstring (Fig. 6.11). The gathered part forms the top of the finished hairnet and a further characteristic is the absence of a browband attached to the front. Another important difference is that the sprang structure of the hairnets without side seams is made with many more warp threads and is therefore much wider. The number of threads of the seamless hairnets is around 500 and higher, while the hairnets without seams have approximately between 220 and 420 warp threads.7 Both types of hairnets without side seams were made in wool as well as in linen, although there is one example with back panels in the Metropolitan Museum of Art, New York, in which the sprang structure is constructed of wool and linen yarn (The Metropolitan Museum of Art NY: 90.5.33). All the known sprang hairnets without side seams have patterning formed by a combination of open and dense sprang; the open structure providing an even more elastic structure. Finds of sprang hairnets without side seams are much rarer than the numerous hairnets with seams. The only 16 that have been published all belong to museum collections.8 Fifteen of these are hairnets with two long panels at the back; only one, which is in the Textile Museum Cairo, is of the short style (TM 139: JE55205). The discovery of four short hairnets of wool at the necropolis of Fag el-Gamous is therefore an important new source for the study of this type of headcovering (Fig.  6.12).

94

Anne Kwaspen and Kristin South

Fig. 6.10 A. Diagram of hairnet with tail panels. B. Diagram of short hairnet without side seams. (© Anne Kwaspen)

A fifth hairnet made of linen can be identified as seamless based on structural details in photographs. Hairnets with long back panels No sprang hairnets with back panels have been found at Fag el-Gamous but this type of headwear was also made in woven fabric, and fragments of three head-covers in this variation have been found at the site. Woven head-covers of this type are known made from linen in tabby weave. They were also woven with stripes created by alternating weft-faced tabby and groups of bare warp ends (Nauerth 1989, 173; Selem & Al-Khalek 2010, 104; Linscheid 2011, 288−90). Three finds from Fag el-Gamous are woven with a striped pattern made in this manner (FeG-1992-SW-39.3; FeG-2005-SW26A; FeG-2005-SW-26B) (Fig. 6.13). When tailed hairnets are produced in sprang, a long interlinking structure is made,9 with the bottom part on the frame becoming the gathered top part of the finished hairnet. Based on measurements of the more or less intact hairnets, this part averages 30 cm in height and is large enough to cover the head. After the structure is finished, the top part on the frame is split vertically in two by unravelling one interlinking thread to the centre line. The length of the two resulting panels is considerably longer

6.  Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt

95

Fig. 6.11. Circle at top of hairnet obtained by cinching the end loops of the structure together over the inserted drawstring. Sprang structure made by Anne Kwaspen

than the created mirror image, which is more densely packed at the bottom of the frame when it is being made. Although the top part on the frame of a sprang structure is always longer, the discrepancy in length for these back panels is remarkable and clearly done on purpose because when the hairnet is worn, the panels need to be long enough to cross and tie over the head.10 The drawstring at the top of a hairnet without side seams fulfills a functional role after the sprang structure is made.11 When the end loops of the structure are cinched to form the top, the two ends of the inserted drawstring hang loose. These string ends are then inserted into the selvedge sides of the sprang, making a spiral pattern along the edge until reaching the centre line of the structure. At that point, the outer ends of the cord hang down on the outside, where they can function again as a drawstring when tied at the back of the head to hold the hairnet in place on the head when worn (James & Médard 2017, 158). In her book on early Byzantine headcoverings, Petra Linscheid studied the way these hairnets with tail panels were worn. She published a portrait sculpture from the collection of the Musée Saint-Raymond in Toulouse, in which the woman wears a

96

Anne Kwaspen and Kristin South

Fig. 6.12. Purple hairnet FeG-1992-SW-39.1; yellow hairnet FeG-1992-SW-39.5; yellow hairnet FeG-1992-SW-39.11; red hairnet FeG-1992-SW-39.19A. (© BYU Egypt Excavation Project)

headcovering that is highly likely to be a hairnet with two back tail panels (Fig. 6.14). On the front top, the gathering of the cloth in a narrow U-shape is clearly visible. The panels are draped over the top of the head. How the panels were fixed or tied cannot be deduced from the sculpture. Short hairnets without side seams Four short seamless hairnets in wool and one in linen have been found in burials at Fag el-Gamous.12 This study of the technical features and construction of these short hairnets is based on only three of the four wool hairnets. The fourth one (FeG-1992SW-39.19A) was not adequately preserved for technical analysis as it was found felted to another hairnet, a rectangular red example with side seams. The linen hairnet has only been available to study from photographs. The total length of the sprang structure of each of the three short wool hairnets varies between 28 cm and 33 cm, which roughly corresponds to the length of the top

6.  Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt

97

Fig. 6.13. Fragment of the top of a woven hairnet with tail panels that are not visible here, FeG 1992 SW 39.3. (© BYU Egypt Excavation Project)

half (bottom part on frame) of the hairnets with tail panels. This shows that the full length of the sprang structure of the short hairnets is sufficient to cover the head. In both the hairnets with side seams and the hairnets with tail panels, the part that was produced at the bottom of the frame became the part that was worn on the forehead. This rule does not apply to short seamless hairnets, probably because it does not matter for the fit. This becomes clear when examining the twist direction: the structure of the top of the purple hairnet is twisted in Z-direction, while the twists at the top of the yellow hairnets are made in S-direction. The decorative pattern of the sprang structure is different for each of the three hairnets, but is always made with a combination of open and dense interlinking. The purple hairnet, in addition, has horizontal stripes that have been formed by alternately twisting single threads and then twisting rows of two threads taken together. The three hairnets are sufficiently preserved to be able to study and compare the lacing of the drawstring around the edges of the sprang structure. It can be deduced that only one drawstring was used and that it was always inserted in the same way (Fig. 6.10. B). The cord was passed through the end loops on one side, then both ends of the cord were inserted into the selvedges in the same manner as the hairnets with tail panels (Fig. 6.15). Unlike the longer hairnets, though, the drawstring continued around the fourth side of the short hairnets, where one end went through the second set of end loops from left to right, while the other went through from the other direction (Fig. 6.16).

98

Anne Kwaspen and Kristin South

Fig. 6.14. Marble Portrait Sculpture Ra 82, Musée Saint-Raymond – Musée d’Archéologie de Toulouse. (© Daniel Martin)

Fig. 6.15. Fragment of selvedge with inserted yellow drawstring. Hairnet FeG 1992 SW 39.1. (© BYU Egypt Excavation Project)

To shape the structure, the top edge was gathered together to form a small circle or narrow U-shape, similar to the longer tailed hairnets (Fig.  6.17). This circle- or U-shape was then sewn in place to secure the shape. When the hairnet was worn, the rest of the drawstring was pulled tight so that the hairnet fitted around the head. With two ends of the drawstring running through the back loops and the bottom structure, it was finally cinched together at the back of the head. To experiment with the fit of this type of short hairnet a sprang structure of 496 warp threads and a total length of 32 cm was made. By fitting the hairnet on a model, it was possible to experiment with the way the drawstring ends could be used and tied. Bringing the drawstrings several times over the top of the head and tying the ends at the bottom resulted in a headcover that looks very similar to that on a sculpture of a woman’s head in the collection of the Metropolitan

6.  Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt

99

Fig. 6.16. Fragment of end loops at the back of hairnet FeG 1992 SW 39.5, with two drawstring ends inserted. (© BYU Egypt Excavation Project)

Fig. 6.17. Fragment of the top of hairnet FeG 1992 SW 39.1. (© BYU Egypt Excavation Project)

Museum of Art, New York (Linscheid 2011, 124) (Fig. 6.18). It is, therefore, considered that this is a very plausible way in which these hairnets could be worn (Fig. 6.19).

100

Anne Kwaspen and Kristin South

Fig. 6.18. (Top) Marble Portrait Bust of a Woman with a Scroll, The Cloisters collection 66.25, Metropolitan Museum of Art. (© Metropolitan Museum of Art) Fig. 6.19. (Bottom) Fitting of a modern hairnet made according to the size of short hairnets without side seams. (© Anne Kwaspen)

6.  Sprang hairnets from the necropolis of Fag el-Gamous in the Fayum, Egypt

101

Conclusion Research thus far has enabled 48 sprang hairnets from Fag el-Gamous to be identified, some in very poor condition and others fully or mostly preserved. Together they illustrate nearly the whole range of types of hairnets in sprang that were made and used in Egypt in the first millennium AD. From this single site, hairnets in wool and linen, and in three of the four main shapes – rectangular and conical hairnets with side seams, and short hairnets without side seams – have been recovered. The remaining type, long tailed hairnets without side seams, has only been identified in a woven structure, not in sprang, but it may yet be found as a systematic search continues of the photographic record, the field books and remains of hairnets from this site. This survey of the technical aspects of sprang headwear from Fag el-Gamous will be supplemented in future with a more detailed study of the archaeological context of the finds; full analyses of as many of the hairnets as are sufficiently documented; and an interpretation of what their presence may have meant to those who buried these complex sprang structures with their loved ones.

Acknowledgements We thank the Ministry of Antiquities in Egypt for their ongoing support of the work of the Brigham Young University Egypt Excavation Project. Anne Kwaspen’s research has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement 844179. Notes

1 Determined by Anne Kwaspen in analysing of the sprang hairnets collections of the Phoebus Foundation (Katoen Natie collection) (45 pieces) and the Louvre (37 pieces). 2 Red hairnet FeG-2003-SW-8 and blue hairnet FeG-2009-NE-34B; red hairnet FeG-1987-NW-7B; yellow hairnet FeG-1987-NW-7A. 3 Hairnets FeG-1981-NE; FeG-1981-582-NC-29; FeG-1984-SE-17; FeG-1992-SW-0A; FeG-1992-SW39.19B. 4 Hairnets FeG-1987-NE-6A1; FeG-1994-0Z; FeG-1994-0E; FeG-2003-SW-13A; FeG-2003-SW-13B; FeG-2006-SE-23A; FeG-2006-SE-24; FeG-2009-NE-7. 5 Blue hairnets: FeG-1992-SW-39.12; FeG-1992-SW-39.14; FeG-1992-SW-39.15; FeG-2006-SE44B; FeG-2006-SW-4/5/6. C; red hairnets: FeG-1994-0D; FeG-2005-SW-13; FeG-2005; FeG-2006-SE-23B; purple hairnet: FeG-2009-NE-34A. 6 Woven browband: FeG-1992-SW-39.12; FeG-2003-SW-13A; FeG-2003-SW-13B; FeG-2006-SE-23B; braided browband: FeG-2005-SW-13; FeG-2005-UB; FeG-2006-SE-44; piled browband: FeG-2009NE-34B. 7 This is based on numbers of warp threads from the hairnets analysed by Anne Kwaspen in the collections of the Phoebus Foundation (Katoen Natie collection) (45 pieces) and the Louvre (37 pieces). 8 Textile Museum Cairo: TM 138 (JE55206), TM 139 (JE55205); The Metropolitan Museum of Art NY: 30.3.56, 33.10.22/23, 90.5.33; Brooklyn Museum NY: 37.1767E, 37.1977E, 37.1770E, 37.1769E; Hermitage St. Petersburg: 13113; Musée de Confluences Lyon: 900.12.60; five hairnets of SMBK Berlin. See catalogue Linscheid 2011.

102

Anne Kwaspen and Kristin South

9 The intact examples published by Linscheid 2011 and James & Médard 2017 have a length between 68 cm and 88 cm. 10 The authors discussed this feature with sprang artist Carol James, who made a replica of the hairnet of Mummy inv no. 900.12.60 Musée de Confluences Lyon. James confirmed that it took a deliberate effort to make the panels much longer and looser than the other part of the sprang structure. 11 The authors identified this feature in the hairnets TM. 138 (JE 55206) of the Textile Museum Cairo and 900.12.60 Musée de Confluences Lyon. Further detailed analysis is necessary on other examples of this type to find out if this is a general feature. 12 Purple FeG-1992-SW-39.1; Yellow FeG-1992-SW-39.5; Yellow FeG-1992-SW-39.11; Red FeG-1992SW-39.19A; Linen FeG-2009-NE-59.

Bibliography

Collingwood, P. 1974. The Techniques of Sprang. Plaiting on Stretched Threads (London, Faber and Faber). De Moor, A., Verhecken-Lammens, C. & Van Strydonck, M. 2002. Radiocarbon dating of Coptic woollen caps in sprang technique. Bulletin du CIETA 79, 27−32. De Moor, A., Fluck, C., Van Strydonck, M. & Boudin, M. 2014. Radiocarbon dating of linen hairnets in sprang technique. British Museum Studies in Ancient Egypt and Sudan 21, 103−20. d’Harcourt, R. 1959. Textiles of Ancient Peru and Their Techniques (Seattle-London, University of Washington Press). Frame, M. 1986. Nasca sprang tassels: structure, technique and order. The Textile Museum Journal 25, 67−82. Galliker, J. & James, C. 2019. Textile interrelationships: Karanis sprang hairnets in daily life. In A. De Moor, C. Fluck & P. Linscheid (eds), Egypt as a Textile Hub. Textile Interrelationships in the 1st Millennium AD (Tielt, Lannoo), 258−75. Hald, M. 1980. Ancient Danish Textiles from Bogs and Burials. A Comparative Study of Costume and Iron Age Textiles (Copenhagen, The National Museum of Denmark). James, C. & Médard, F. 2017. Appreciation of the ancient craftsmen through the recreation of a byzantine sprang bonnet from the Antinoopolis necropolis. In A. De Moor, C. Fluck & P. Linscheid (eds), Excavating, Analysing, Reconstructing. Textiles of the 1st Millennium AD from Egypt and Neighbouring Countries (Tielt, Lannoo), 148−61. Kwaspen, A. 2011. Sprang hairnets in the Katoen Natie Collection. In A. De Moor & C. Fluck (eds), Dress Accessories of the 1st Millennium AD from Egypt (Tielt, Lannoo), 70−95. Kwaspen, A. 2014. Features and analysis of sprang hairnets. In C. Alfaro, J. Ortiz & M. Anton (eds), Tiarae, Diadems and Headdresses in the Ancient Mediterranean Cultures (Valencia, University of Valencia), 273−82. Linscheid, P. 2011. Frühbyzantinische textile Kopfbedeckungen (Wiesbaden, Reichert). Nauerth, C. 1989. Die koptischen Textilien der Sammlung Wilhelm Rautenstrauch im Städtischen Museum Simeonstift Trier (Trier, Städtischen Museums Simeonstift). Pritchard, F. 2006. Clothing Culture: Dress in Egypt in the First Millennium AD (Manchester, The Whitworth Art Gallery). Selem, M. & Al-Khalek, S. 2010. Egyptian Textiles Museum (Cairo, Ministry of Culture). South, K. 2017. The use of basket-weave linen in burials of the necropolis of Fag el-Gamus, Egypt. In A. De Moor, C. Fluck & P. Linscheid (eds), Excavating, Analysing, Reconstructing. Textiles of the 1st Millennium AD from Egypt and Neighbouring Countries (Tielt, Lannoo), 88−107. South, K. & Kwaspen, A. 2021. The tunics of Fag el-Gamus (Egypt): a survey of types. In M. BustamanteÁlvarez, E.H. Sánchez López & J. Jiménez Ávila (eds), Purpureae Vestes VII. Redefining Ancient Textile Handicraft: Structures, Tools and Production Processes (Valencia, University of Valencia), 469−77. Van Reesema, E.S. 1920−1921. Egyptisch vlechtwerk (Amsterdam, Van Holkema & Warendorf).

Chapter 7 Tight-fitting clothing in antiquity and the Renaissance: Research and experimental reconstruction Dagmar Drinkler and Carol James

Introduction Collingwood left us an encyclopedic treatment of the technique ‘sprang’. He listed material evidence for the use of sprang as a textile technique since early antiquity. By examining pottery, sculptures, portraits, and texts, this paper entertains the possibility that sprang was used repeatedly in Europe and Asia Minor to create tightfitting clothing. Designs on many Greek vases as well as sculptures from the fifth century BC depict people wearing tight fitting, highly decorated clothes. While most of these are hose, some are also full body suits. The question arises as to how these form-fitting textiles were made. Woven cloth is not sufficiently elastic and there is no evidence that knitting was known at this time. The written record contains no information and the textile and costume research sources yield almost no insight. Experimental work to re-create these clothes reveals that all motifs found can be easily and efficiently rendered using the sprang technique. Likewise, Renaissance portraits depict individuals wearing form-fitting hose. An added development featured on these garments is a line mid-thigh, above which the pattern changes. The written record is confusing: a sample text describing knitting features a sprang frame. Experimental reconstruction of such hose provide compelling reasons that sprang, as opposed to knitting, may have been used as the method of choice for these garments. The earliest depictions of sprang, along with its tools, can be found on vases from very early times (Jenkins & Williams 1985). In fact, men wearing tight-fitting and strikingly patterned garments, particularly hose, appear on a great number of ancient works of art. Extant pieces of sprang are found worldwide (Collingwood 1974, 37−42), dating from the Bronze Age in the second millennium BC up to the

104

Dagmar Drinkler and Carol James

beginning of the nineteenth century. The earliest find is a hairnet excavated from a woman’s grave at Borum Eshøj, near Aarhus (Denmark) dating c. 1400 BC (Hald 1980, 245, 251). In Egypt more than 400 sprang headcoverings of the fourth to seventh century AD have been found (Linscheid 2011) and an example of tubular sprang, which has sometimes been interpreted as a legging and which has been radiocarbon dated to AD 425−535 was recovered from a bog at Tegle, Norway (Halvorsen 2010, 97−99). Examples of sprang items which have not been excavated include military sashes preserved from the period from 1600 to the beginning of the nineteenth century. Probably the earliest is dated 1602 and belonged to Elector Christian II of Saxony (Schuckelt 2010, 115), and the most famous is associated with George Washington c. 1779 (Collingwood 1974, pl. 67). With growing industrialisation in textile production at the beginning of the nineteenth century, sprang techniques generally fell into disuse apart from in remote regions presumably because this technology cannot be implemented mechanically. The textiles under discussion must have been stretchable, but the few available written sources give no information about this aspect of clothing. Experimental work could be important to further the understanding of potential methods for the creation of objects. Reverse engineering using methods known at the time is not a guarantee that this is how it was done, but leads to new avenues of investigation and allows a certain appreciation of the sophistication of the textile workers of the time. The project described here brought about a greater appreciation of certain correlations between colour design, requirements of use, and specific properties of the textile technique but more testing is desirable in the future to accord with scientific protocols (Andersson Strand 2010).

Tight-fitting clothes from Greek vase paintings and polychrome sculptures From the fifth century BC onwards depictions in Greek art show Greeks in mostly aggressive confrontation with individuals or groups of people, the enemies standing out by their form-fitting hose or often even ‘suits’ that cover the entire body. These garments characterise ‘foreigners’ or ‘strangers’, which refer to the Scythians and the Persians as well as the mythical Amazons. Greek vase paintings depict close-fitting clothes (Wünsche 2008) (Fig. 7.1), as do sculptures like the representation of Paris as an archer from the western gable of the Temple of Athena Aphaia (Brinkmann 2003) (Fig. 7.2. A), today in the Glyptothek in Munich, or the so-called Persian horseman in the Acropolis Museum in Athens (Brinkmann et al. 2008) (Fig. 7.3. A). The Greeks never depict themselves in this manner. The question arises as to the manner in which these form-fitting clothes were produced. Little is known about this. The ancient historian Herodotus reports that the Scythians and the Persians wore leather breeches (Calmeyer 1975, 474; Gleba 2008). In Scythian graves in the Altai mountains of Siberia felted stockings have been found (Rudenko 1970, 95) and unpatterned trousers of goatskin

7.  Tight-fitting clothing in antiquity and the Renaissance

105

Fig. 7.1. Pelike 2351, 440‒435 BC, Munich, Staatliche Antikensammlungen. (© Archives Staatliche Antikensammlungen und Glyptothek, Munich)

(Simpson & Pankova 2017, 350). Leather, felt and woven cloth cannot account for the elasticity that must be present in the close-fitting garments. The elastic textiles that cling to the body cannot be reproduced by any of the weaving techniques known to us. Other techniques, such as nålbinding, produce stretchable textiles, but they lose this quality when patterns are worked in. The sprang technique A method for creating elastic textiles is the sprang technique (Collingwood 1974), the existence of which can be verified for the time around 1400 BC. Sprang, originally a word from Swedish (in Old Swedish språngning) for openwork textiles, has been used for several decades now exclusively for the braiding method with stretched threads. Towards the end of the 1960s, it was agreed in textile circles to use this Swedish term worldwide for all objects manufactured using this technology, including Coptic headgear and South American works. Other terms include knitting, network, Egyptian plaiting, braiding, crawling, lace weaving, openwork, Coptic lace, Coptic plaiting (Collingwood 1974, 34). Parallel threads are stretched around a frame and then twisted, interlinked or intertwined to produce a plaited structure. To understand the elastic quality of

106

Dagmar Drinkler and Carol James

a

b

Fig. 7.2. A. Reconstruction of Paris as an archer from the western gable of the Aphaia-Tempel, c. 490‒480 BC by V. Brinkmann 2003, Munich, Glyptothek. (© Archives Glyptothek, Munich). B. Double zigzag pattern as found on the Paris sculpture in one-layer sprang. (© Drinkler)

a

b

Fig. 7.3. A. Reconstruction of so-called Persian horseman in the Acropolis Museum in Athens, c. 490 BC, by O. Primavesi 2008. (© Dieter Rehm, Munich). B. Two-layered sprang with long lozenges pattern as found on the Persian horseman. (© Drinkler)

7.  Tight-fitting clothing in antiquity and the Renaissance

107

sprang, it is necessary to explain the difference between a woven textile and sprang. A woven fabric is based on two systems of threads crossing each other at more or less right angles. The vertical system is called the warp, and the horizontal one is called the weft. Both warp and weft are necessary to produce a woven fabric. In sprang a single system of parallel vertically stretched threads is required. The warp is one continuous thread that goes around the upper and lower ends of a frame, respectively. A simple frame construction is sufficient, along with a tensioning device. The engagement of the threads as they form cloth, results in a certain uptake in length. The requirement for the initial size of the frame is determined by the length of the finished textile plus the take-up, which is usually about 30% (James 2011). For the sprang method the threads are divided into two layers, one in front of the frame and one at the back under constant tension. Both hands are needed and at least two rods for separating or fixing the two layers of threads respectively. The left hand separates the threads of the two layers, while the right hand picks up the threads one after the other, with the fingers twisting them (Fig. 7.4). This is done by picking up a thread from the back layer and then pushing a thread from

Fig. 7.4. Picking up a thread from the reverse and then pushing a thread from the front to the back. (© Staatliche Antikensammlungen und Glyptothek, Munich, Renate Kühling)

108

Dagmar Drinkler and Carol James

the front layer to the back. This means that each stitch is composed of two threads: one thread from the front layer, and one thread from the back. Each row is worked separately. In fact when one row has been worked a mirror image of that row is created simultaneously. After finishing each row the twisted threads have to be fixed by inserting a separating rod between the row that has been worked and its mirror image. Then a second rod is inserted into the same space and used to push the entire mirror image row of stitches to the opposite end of the stretched threads, thus adding a new row to the plaited textile at that end as well. This mirror image structure means that one section is in S-direction and the other is in Z-direction. Both sections of the sprang textile grow towards each other. At the end of this process the un-worked portion of threads become so short that no more plaiting is possible. To finish the threads must be fixed in their position in the middle of the piece to ensure that it cannot unravel. It is possible to insert a horizontal thread into this space, at the expense of elasticity at this point in the cloth. A more elastic method to secure the plaited work is to crochet a row of chain stitches across these threads. Both methods create a continuous textile with strong borders and a relatively compact, visually contrasting middle section. Simple, schematic drawings showing the interlinking of threads in sprang can be found in every publication on this topic. They suggest a loose and openwork structure because the textile is always shown as if under tension, to explain how every single thread is positioned in the course of working. When relaxed, however, sprang textiles very often have tight and compact surfaces that have ridges similar to a woven twill. The extent of elasticity in sprang is related to the material used (e.g. wool, linen, silk, or cotton), the thickness of the threads, the twist angle, the structure produced and finally the method of working. Depictions on Greek vases indicate that the ancient Greeks knew this technique. The vases show women holding a frame with a zigzag-patterned textile on it (Clark  1983; Jenkins & Williams 1985). It can be assumed with certainty that these are frames for sprang, on which the headwear of Greek women was produced. This headwear had the same patterns as the hose. In modern literature however these frames are mostly referred to as embroidery frames. The patterns of the tight-fitting hose and their execution in sprang Patterns found in depictions of such garments in Greek art can be reconstructed easily and efficiently in this technique by various methods of working: single-layer, two-layer or multi-layer sprang. It is also possible to create a pattern by alternating between single-layer and two-layer sprang. The pattern of the textile is determined by the arrangement of threads in different colours that are stretched around the frame before starting to work. Single-layer sprang is suitable for stripes, small zigzag and rhomboid patterns, which look the same on both sides of the fabric (Fig. 7.5). A variation of this is the

7.  Tight-fitting clothing in antiquity and the Renaissance

a

109

b

Fig. 7.5. A. Simple vertical stripes are obtained by interlinking, by intertwining the colours alternately one can obtain a small zigzag pattern. B. By intertwining only one colour alternately one can obtain a small rhomboid pattern. (© Drinkler)

double zigzag pattern that can be found on the sculpture of Paris (Fig. 7.2. B).1 In two-layer sprang, following the pattern, the threads of one colour are placed over the threads of the other colour, thus producing two layers that are not connected (Fig. 7.6). As the threads move diagonally in plaiting, patterns with diagonal lines can easily be created and are especially suitable for sprang. By interchanging the layers, a coloured pattern can be obtained. When working in this manner with only two colours, the pattern is in the opposite colour scheme on the reverse side of the fabric. If the second layer is worked in several colours, a diamond pattern in different colours can be obtained. A variation of this is the pattern that can be observed in the colour reconstruction of the so-called Persian horseman (Fig. 7.3. B). In three-layer sprang a third colour is simply added when stretching the threads around the frame in the beginning. Thus front and reverse will each show one colour, while the third colour is sandwiched between the outside layers (Fig. 7.7).

110

Dagmar Drinkler and Carol James

Fig. 7.6. The double zigzag pattern in a two-layer sprang. The hose are completely made as hose, and  the top consists of two separate sleeves and a felted waistcoat. (© Staatliche Antikensammlungen  und Glyptothek, Munich, Renate Kühling)

a

b

Fig. 7.7. A. Situla, c. 360 BC, MuseoNazionale Archaeologico, Naples. (© Jastrow, wikipedia, free domain).  B. Zigzag pattern in a three-layer sprang. (© Drinkler)

7.  Tight-fitting clothing in antiquity and the Renaissance

111

By interchanging several layers, patterns in dress depicted in ancient works of art can be re-created. To date we have successfully re-created in sprang all such patterns that we have found. Assuming the same size thread is used, a piece of two-layer sprang is twice as thick and heavy as a piece of the same size in singlelayer sprang. With the use of finer threads the number of stitches per cm increases both widthwise and lengthwise. There is also the possibility of making a pattern by alternating between single-layer and two-layer sprang. Such cloth requires a larger than otherwise necessary number of threads in width, as the overall width and elasticity of the single-layer sprang are limited by the narrower two-layer areas. If there is only one colour of threads visible in a horizontal row, then the other colour(s) are all in the reverse layer. To create a pattern combining single-layer and two-layer cloth, both colours are visible for a few rows, then only one colour is visible for a few rows (Figs  7.8 & 7.9). This method, varying between singlelayer and two-layer cloth, causes uneven borders. In small pattern repeats this would have little effect when wearing the hose. In larger Fig. 7.8. By alternating one- and two-layered sprang, patterns, it can cause horizontal you can produce a dotted pattern. (© Drinkler) stripes that are less elastic than the rest of the textile. In a vase painting depicting hose with a larger repeat it is clearly visible how the horizontal stripes draw in, thus deforming the shape of the thighs of the garment (Fig. 7.10).

112

Dagmar Drinkler and Carol James

Fig. 7.9. This pattern variation is an interconnection of horizontal and vertical stripes. It is another pattern obtained by alternating one- and two-layered sprang. If the front of the textile shows only one colour, the other colour is at the back. To make a pattern, both colours are united in one layer. (© Drinkler)

a

b

Fig. 7.10. A. Detail on Fig. 7.1. B. Uneven border in one- and two-layered sprang alternately. (© Drinkler)

7.  Tight-fitting clothing in antiquity and the Renaissance

113

Sprang in the Renaissance Following the Greek and Roman era, sprang was primarily used for headwear. It was not until the end of the fifteenth century that sprang reappeared in fine art. Panel paintings depict sprang headwear as well as women in front of sprang frames. From the beginning of the sixteenth century, there are many depictions of women in front of a vertical frame (von Wilckens 1991, 336−38, abb. 378). Even the Virgin Mary is often shown in this position. Only rarely do the women hold a shuttle in their right hand for inserting a weft. In most cases, the hands are depicted in the position that is very typical for sprang (see Nutz et al. this volume, Fig. 9.5). Mostly this work is described as ‘weaving’, or, if the author was not really sure, it is called ‘women’s handwork’. Striped hose During the fifteenth century the tight-fitting, one-piece pair of men’s hose became fashionable at the Burgundian court. Written sources tell us that these were made from leather, wool or linen fabric (Zander-Seidel 1990, 181−85; Fingerlin 1997, 101; Niekamp & Woś Jucker 2008). To achieve a perfect fit around the leg, hose made from these relatively inelastic materials were produced with vertical back seams and gussets for the buttocks.2 These hose either had two legs of the same colour or each leg in a different colour. Towards the end of the fifteenth century men’s formfitting, coloured hose with vertical stripes appeared. This fashion can be seen in panel paintings, sculptures, and graphic art between around 1490 and 1530. The men who are depicted wearing these colourful striped hose are almost always outcasts of society such as hangmen and torturers. Here again the question remains. How were these garments made? It is unlikely that such a perfect fit without any creases could be achieved for hose sewn together from strips of coloured fabrics. Working vertical stripes in knitting is, at best, awkward and time consuming. The easiest way to create a coloured pattern in knitting is to work horizontal stripes, but this never appears in stockings in fine art of those times. In sprang however, vertical stripes can be produced quite easily. The threads are stretched around a frame in the desired width and colours of the stripes and then worked in single-layer sprang by simple interlinking. In paintings from this period, for example by Lucas Cranach the Elder 3 and Hans Holbein the Elder (Fig.  7.11), a second type of hose can be observed: simple vertical stripes from the toe up to above the knee, then continuing in alternating colours and different patterns. These patterns can also be created in sprang. For the pattern change in the top section of the hose, a second layer of threads is added to the sprang piece by simple threading. It is easy to work one row of chain stitch in crochet, onto which the additional vertical threads are inserted. This method creates a thin horizontal line, which is usually not visible in depictions of hose. But there is one case where such a line can clearly be seen: it is a pair of hose worn by a man on the left side panel of the altarpiece of Gerolzhofen by Tilman Riemenschneider (Fig. 7.12). The sprang sample made

114

Dagmar Drinkler and Carol James

Fig. 7.11. Hans Holbein the Elder, Basilica San Paolo fuori le mura (detail), c. 1504, oil on wood, Augsburg, Staatsgalerie Altdeutsche Meister, Inv. No. 5332. (© Munich, Bayerische Staatsgemäldesammlungen)

for comparison shows that when stretching it, the single-layer sprang in the striped legs is less elastic horizontally than the two-layer sprang in the top section. This quality is not only an advantage for fit and comfort when wearing the hose, but the two-layer sprang is also less transparent at this delicate part of the garment. The conspicuously patterned hose of a gondolier in a Venetian painting by Vittore Carpaccio has garnered much attention in the history of dress (Fig. 7.13. A). The upper section of the hose shows a trompe l’oeil cubic pattern, which can be produced in two-layer sprang in two colours. The third colour that is necessary for the three-dimensional effect is created by a blend of the two colours. When

7.  Tight-fitting clothing in antiquity and the Renaissance

a

115

b

Fig. 7.12. A. Tilman Riemenschneider, altarpiece of Gerolzhofen, left panel, c. 1515, Munich, Bayerisches Nationalmuseum, Inv. No. MA 1963. The executioner is wearing tight-fitting hose with vertical stripes along the legs and diagonal stripes at the bottom. B. Vertical stripes in simple interlinking sprang and in the upper part diagonal strips in a two-layer sprang. (© Drinkler)

the sprang sample was worked, the third colour was obtained by two-coloured hatching (Fig. 7.13. B). The shoes worn with the striped hose It is striking to note that when striped hose and stockings came into fashion, the shape of the shoes changed at the same time. The long pointed poulaines were replaced almost without exception by square-toed shoes worn with the striped hose. These so-called ‘cow-mouth’ shoes were probably much more comfortable but this seems to be only one aspect of this new fashion. Hose produced in sprang are provided with a square end of the foot without a shaped toe-cap, which consequently requires a square-cut shoe. So far historiography has not explained the origin of this shoe fashion sufficiently. But worn together with sprang hose the square-toed shoes with very wide openings would cling to the feet.

116

Dagmar Drinkler and Carol James

a

b

Fig. 7.13. A. Vittore Carpaccio, The Miracle of the Relic of the Cross at the Ponte die Realto, 1494, tempera on canvas, Venice, Gallerie dell’ Accademia. (© Published with kind permission of Ministero per i Beni e le Attività Culturali Venezia). B. Sprang work of the trompe l’oeil cubic pattern, which can be produced in two-layer sprang in two colours. (© Drinkler)

Experimental reconstruction To produce a pair of hose both legs are worked in one sprang piece. Work progresses from ankles to waistline. The piece is finished in its middle section. The threads are cut and knotted immediately to keep the waistband of the hose elastic. Then the vertical seams of the legs are joined and the hose are sewn together. By following the edge of the plaiting with the sewing thread, the seam is not visible (Fig. 7.14). A project to explore the construction was undertaken, looking at diverse aspects of Renaissance hose, the creation of a foot, and the detail of the diamond pattern seen in the gondolier in the painting by Carpaccio (Figs  7.13 & 7.15). The warp selected was a tightly twisted merino sock yarn from Yarn2Dye4, in two colours: natural white and indigo-dyed blue, 200  g of each. Allowing for shrinkage of 30%, the length of each warp thread section had to begin at about 360 cm. Working this length of warp creates certain technical problems. The false-circular warp method described by Peter Collingwood eliminates many of these problems (Collingwood 1974, 256−61). Such a ‘false circular’ warp was established using only 1.5 m length of the frame.

7.  Tight-fitting clothing in antiquity and the Renaissance

117

Fig. 7.14. Hose worked in two-layered sprang in two colours. (© Drinkler)

The foot Creation of a foot requires two different lengths, a longer portion that will cover the sole of the foot, and a shorter portion to cover the top of the foot (Fig. 7.16). The work began as two pieces to be sewn into a foot upon completion. The seam at one side of the foot will be continuous with the inside seam of the hose (Fig. 7.17). Following this plan, the hose began as two warps. The portion for the sole of the foot was 7 cm longer than the portion intended as the top of the foot. Work started on the longer portion. With 6 cm of cloth completed, the remaining un-worked warp was the same length on both pieces (there had been a 1 cm uptake in the warp), and the two warps were now worked side by side, maintaining a split between the two (Fig. 7.18). When the split measured 17 cm, work proceeded across the full width of the two warps, closing the split (Fig. 7.19). The design at the thigh Work progressed up the length of the leg. At 70 cm, it is time to consider adding threads for the supplemental warp. Diverse methods were explored to imitate the diamond shape in the gondolier’s hose. The method that finally

118

Dagmar Drinkler and Carol James

Fig. 7.15. Hose worked to explore patterns in the painting by Carpaccio, worn with low cut, ‘cow mouth’, slip-on shoes. (© James)

worked was to ‘sew’ the new warp thread in place. A length of thread was measured corresponding to the amount required for the addition in that colour zone. A mark was placed at the centre of this length. A needle and sewing thread attached that centre to the desired location in the cloth for the

7.  Tight-fitting clothing in antiquity and the Renaissance

119

Fig. 7.16. Schema for creating the foot of the hose. (© James)

Fig. 7.17. Invisible seam in sewing up interlinking sprang. (© James)

point of the diamond. The new thread was incorporated into the cloth, and after pushing the cross to the mirror-image side, the far ends were sewn to the corresponding place on the mirror-image side. Subsequent additions on either side of this first stitch were ‘sewn’ in on both ‘Z’ and ‘S’ sides (Fig.  7.20).

120

Dagmar Drinkler and Carol James

Fig. 7.18. The hose begins as two separate pieces of cloth, worked side-by-side, one shorter than the other. (© James)

To help support the end loops of this supplemental warp, a thread was passed through the loops on the inside of the hose. Unfortunately, the length was not sufficient to completely double the colour zone, nor had provisions been made for additions in all colour zones. When the remainder of the supplemental warp was added, it proved quite difficult to push the cross down. The addition of the supplemental warp creates a need for greater width. With twice the threads, twice the width space is required (Fig. 7.21). It became apparent that there could be functional reasons for aspects

7.  Tight-fitting clothing in antiquity and the Renaissance

121

Fig. 7.19. The two separate pieces of cloth join to form a single piece that will be the leg. (© James)

of the gondolier’s design. Had there been sufficient length to complete double the colour zone, and had the additions been made on all stripes (only three were attempted), then we speculated that the transition to the greater width of the cloth would have been eased. Shaping Interlinked sprang is a very elastic cloth, but this elasticity is insufficient to allow for the crotch shaping. Threads were cut to shape the cloth. Knots are placed at both S and Z sides of the cloth and had to sit tightly against the cloth to prevent

122

Dagmar Drinkler and Carol James

Fig. 7.20. The supplemental warp threads, sewn in place, loops reinforced with a separate thread. (© James)

unravelling (Fig.  7.22). Greater length along the centre back seam is provided by use of ‘short rows’ (rows that do not run the full width of the piece) (Fig.  7.23). The cloth measures an extra 7 cm along the centre back seam. To create a casing for a drawstring at the waistline, threads were sorted into their separate colour layers and one colour layer was worked longer than the other. This allows the knots that end the work to hide on the inside of the waistband. The finished pair of hose presented bear striking resemblance to the hose worn by the gondolier in the painting by Carpaccio. ‘Cow mouth’ shoes A pair of low-cut, slip-on shoes was constructed using tanned leather and an estimated pattern to approximate the shape of the broad, square-toed shoes. It was anticipated that a tie would be required across the instep to keep the shoe on the foot. The square toe of the hose required special attention to be intentionally tucked into the toe of the shoe. Once on the foot, the shoe stayed snug. Considerable walking around was possible without losing the shoes.

7.  Tight-fitting clothing in antiquity and the Renaissance

123

Fig. 7.21. The addition of the supplemental warp creates a need for more space. With twice the threads, twice the width space is required. (© James)

124

Dagmar Drinkler and Carol James

Fig. 7.22. Detail of shaping for the crotch. © James

Fig. 7.23. Detail of the allowance for the rise in the centre back. (© James)

7.  Tight-fitting clothing in antiquity and the Renaissance

125

Subsequent research Subsequent experimentation has verified suspicions concerning the diamond shape on the thigh. Use of a sufficient length of supplemental warp completely filled the colour zone, greatly eased the increase in width, resulted in the lower edge of the diamond shape on the gondolier’s thigh. The finished garment presents a striking resemblance to the hose worn by Carpaccio’s gondolier (Fig. 7.24). However, testing the use of the technique in this way by others experienced in sprang and selecting

Fig. 7.24. Finished pair of hose using sufficient supplemental warp at the gradual addition. (© Bastian Krack)

126

Dagmar Drinkler and Carol James

wool yarn closer to that available in the past is desirable to help endorse the experiment.

Conclusion Because of the lack of extant garments, it cannot be proved whether the striped hose in the Renaissance or hose in antiquity were produced using sprang, but the parallels are striking. In both periods there are depictions of women working on small frames, on which hairnets and caps can be produced in sprang. There are no hints at the fabrication of entire hose. The wearers of the hose in ancient art are generally not the Greeks but foreigners such as the Scythians and the Persians. In the Renaissance the wearers of striped hose are in the beginning only outcasts or criminals, including the soldiers who brought Christ to the cross. Written sources do not mention anything that could be related to hose being produced by plaiting. Synonyms for weaving according to Roget’s Thesaurus include braiding, interlinking, knitting, and plaiting (Roget 2001). The German dictionary of the Brothers Grimm (Grimm & Grimm 1854−1960) makes clear that the words to describe textile techniques like knitting, netting, knotting, or plaiting, cannot clearly be separated from one other. The term ‘knitting’ (‘Stricken’) was generally employed for the production of meshwork. But how this was achieved remains unclear. No tools or devices, let alone knitting needles, are ever mentioned in the early texts. Since little is known about the origin of knitting, possibly in the early days the notion of knitting included the sprang technique. The interlinking structure provides natural elasticity. When pulled laterally the stitches open, accommodating greater width, and the length shortens somewhat. In interlinking double cloth, the two layers do not sit perfectly one beneath the other. When the interlinking stitches are pulled laterally and open, there is another layer of stitches to block the sightline through the hole, preventing transparency. Woven double cloth is no wider than single layer cloth. It was surprising to note that the sprang double cloth sat much wider at rest than the single layer cloth. This is due to the space required for the threads to pass from upper layer to lower and vice versa. This forces a slightly looser tension in the interlinking stitch. The double cloth tested was noticeably more elastic than comparable single-layer sprang. The double cloth at the hips of these hose is doubly practical: it provides extra width as well as opacity, making sprang an excellent choice for this type of garment. The reconstruction of upper garments has not yet been attempted but it should be possible to produce them using the same method. Depictions of such garments clearly show that the sleeves were mostly set in separately. In rare cases no seam is visible at the shoulders, which possibly indicates that it is a continuous sprang piece from hip to wrist with a low V-neckline. The production of form-fitting garments as described here cannot be proven by surviving examples from antiquity, although there is no doubt that sprang was known

7.  Tight-fitting clothing in antiquity and the Renaissance

127

to the ancient Greeks and that patterns of the elastic textiles depicted in ancient art can be executed efficiently in sprang. It is hoped that this research will provide a new stimulus to textile and dress research. This very early and simple method of creating a flat textile, namely plaiting with its special technique of sprang, should be restored to its place of importance among other textile techniques. Notes

1 The double zigzag pattern on the Paris sculpture can only be traced as a weathered relief. There are no remnants of the original pigments. The choice and number of colours were arbitrarily selected by the artist. I chose to use two colours in my reconstruction of the cloth. Also the sculpture of Paris as archer clearly depicts the elasticity of the pattern. At the thigh the pattern repeat is distinctly shorter than at the ankles, corresponding exactly to a real braid in sprang. Unfortunately the coloured casts do not reflect these differences in the stretched material as does the original. 2 There are many paintings that clearly show the cut of sewn hose. Only the piper and drummer from Albrecht Dürer, Three Kings Altar, outside right wing, c. 1504 in Cologne, Wallraf-Richarz Museum, is shown as an example, as well as the representation of a pair of leather hose: Tobias Stimmer, Portrait of Jacob Schwytzer, 1564, Public Art Collection, Kunstmuseum Basel. 3 So-called Katharinenaltar, 1506, oil on wood, Dresden, Gemäldegalerie Alte Meister, Gal. No. 1906.

Bibliography

Andersson Strand, E. 2010. Experimental textile archaeology. In E. Andersson Strand, M. Gleba, U. Mannering, C. Munkholt & M. Ringgaard (eds), North European Symposium for Archaeological Textiles X (Oxford & Oakville, Oxbow Books), 1−3. Brinkmann, V. 2003. Der Prinz und die Göttin. Die wiederentdeckte Farbigkeit der Giebelskulpturen des Aphaia-Tempels. In V. Brinkmann & R. Wünsche (eds), Bunte Götter (Munich, Staatliche Antikensammlungen und Glyptothek), 84−97. Brinkmann, V., Koch-Brinkmann, U., Piening, H. & Trianti, I. 2008. Der schöne Feind. Neue Untersuchungen am ‘Perserreiter’ von der Athener Akropolis. In V. Brinkmann (ed.), Bunte Götter. Die Farbigkeit antiker Skulpturen (Frankfurt am Main, Liebieghaus), 95‒101. Calmeyer, P. 1975. Hose, Reallexikon der Assyriologie und vorderasiatischen Archäologie (Berlin & New York, Walter de Gruyter & Co.). Clark, L. 1983. Notes on small textile frames pictured on Greek Vases. American Journal of Archaeology 87, 91−96. Collingwood, P. 1974. The Techniques of Sprang. Plaiting on Stretched Threads (London, Faber and Faber). Fingerlin, I. 1997. Seltene Textilien aus Kloster Alpirsbach im Nordschwarzwald. Waffen- und Kostümkunde 1997, Vols. 1 & 2, 99–122. Gleba, M. 2008. You are what you Wear: Scythian Costume as Identity. In M. Gleba, C. Munkholt, & M.-L. Nosch (eds), Dressing the Past (Oxford, Oxbow Books), 13−28. Grimm, J. & Grimm, W. 1854−1960. Deutsches Wörterbuch (Leipzig, Hirzel). Hald, M. 1980. Ancient Danish Textiles from Bogs and Burials. A Comparative Study of Costumes and Iron Age Textiles (Copenhagen, The National Museum of Denmark). Halvorsen, S.W. 2010. Norwegian peat bog textiles: Tegle and Helgeland revisited. In E. Andersson Strand, M. Gleba, U. Mannering, C. Munkholt & M. Ringgaard (eds), North European Symposium for Archaeological Textiles X (Oxford & Oakville, Oxbow Books), 97−103. James, C. 2011. Sprang Unsprung (Winnipeg, Manitoba, Sash Weaver).

128

Dagmar Drinkler and Carol James

Jenkins, I. & Williams, D. 1985. Sprang hair nets: their manufacture and use in Ancient Greece. American Journal of Archaeology 89, 411−18. Linscheid, P. 2011. Frühbyzantinische textile Kopfbedeckungen (Wiesbaden, Reichert Verlag). Niekamp, B. & Woś Jucker, A. 2008. Das Prunkkleid des Kurfürsten Moritz von Sachsen (1521–1553) in der Dresdner Rüstkammer. Riggisberger Berichte 16 (Riggisberg, Abegg-Stiftung). Roget, P.M. 2001. Roget’s International Thesaurus, 6th edition (New York, Harper Collins). Rudenko, S. 1970. Frozen Tombs of Siberia: The Pazyryk Burials of Iron Age Horsemen (Berkeley, University of California Press). Schuckelt, H. 2010. Türckische Cammer (Dresden, Staatliche Kunstsammlungen Dresden). Simpson, S.J. & Pankova, S. 2017. Scythians: Warriors of Ancient Siberia (London, British Museum, Thames & Hudson). Wilckens, L. von, 1991. Die textilien Künste. Von der Spätantike bis 1500 (Munich, C.H. Beck). Wünsche, R. 2008. Starke Frauen (Munich, Staatliche Antikensammlungen und Glyptothek). Zander-Seidel, J. 1990. Textiler Hausrat – Kleidung und Haustextilien in Nürnberg von 1500–1650 (Munich, Deutscher Kunstverlag).

Section III Braiding and lace making

Chapter 8 Braided strings and Turk’s head knots on European secular and religious textiles Joy Boutrup

Introduction Braided strings and Turk’s head knots are part of a larger group of textiles which today is called passementerie. They are often present together on medieval and renaissance purses, bags, bookmarks, seal strings as well as on many religious textiles. Both were made by the same artisans and often of the same materials. These parts of textile objects have until recently been studied very little and are often not analysed or even described in detail in the records of church collections and museums. This paper aims to point out the lack of information in these fields at present and to show some examples of analysis and description. The organisation of the professional makers and the trade has been researched to some extent mainly based on the guild records and regulations in the different countries in Europe (Braun-Ronsdorf 1956, 26−30; Heutte 1972, 12−35). The guilds have changed and merged over time in most countries so that these artisans have not always been in a separate guild but often part of other guilds. It is remarkable though that women were allowed in the guilds and often were masters of workshops themselves (Depping 1837, 382−86). Even outside the guilds many women were actively producing and selling passementerie in market places, and a certain more or less professional production probably took place in cloisters and monasteries all over Europe. There is no written evidence from the professional workshops as to what tools they used or which structures of braids or knots they produced. Knowledge of the structures is limited to studies of the surviving objects and the few written instructions from the early fifteenth century. These are all instructions for loop braiding written by non-professionals: a manuscript in the British Library, MS Harley 2320, and an East Anglian household book known as the Book of Secrets, which remains the property of the Tollemache family (Griffith 1997). There are no contemporary instructions for

132

Joy Boutrup

Turk’s head knots known at present. Loop braiding is depicted in two frescoes and one panel painting all of which portray domestic scenes. They are ‘The weaver fresco’, c. 1316 in the Haus zur Kunkel, Constance, Baden Wurtemberg, Germany; a weaving scene in the lower right corner of The Triumph of Minerva, Palazzo Schifanoia, Ferrara, Italy, c. 1467; and the History of the Virgin Mary, on an altarpiece by Nicolás Zaortiga in the church of St Mary, Zaragoza, Spain, dated 1465. The first written description of these types of crafts is found in Christoff Weigel’s Abbildung der gemeinnützlichen Hauptstände from 1698. An engraving by Weigel shows a button maker holding the button stuck on an awl in one hand and needle and thread in the other hand (Fig. 8.1). Other kinds of passementerie works are shown hanging on the wall, corded strings and fringes are being made in the foreground and background. It is clear from the text that several types of Turk’s head knots are produced and also that the loop braiding is used in an advanced state. Braids with motifs and letters are mentioned.

Loop braiding Loop braiding is one of several braiding techniques, but recent studies have shown that it was the main production method both in professional workshops and in amateur surroundings from the middle ages until the invention of braiding machines (Speiser & Boutrup 2009a & b; 2011; 2012). It is a fast braiding method especially well-suited for thin and slippery yarns, such as silk. No tools are needed and several people can work together by exchange of loops, one method of which is described in the manuscripts with instructions for loop braiding. Loop braiding allows for a large variation of braids such as flat, compact round, hollow round double-layered structures with exchange of layers (Speiser 2000; Griffith 1997; London, British Library MS Harley 2320). Some structures could only have been made by loop braiding while other structures could also have been made by other methods. However, certain mistakes in the braids indicate that loop braiding has been used. The first workable methods for the analyses and rendering of oblique interlaced structures were devised by Noémi Speiser in her book on braiding (Speiser 1983, 76−79, 217−19). The analyses consist of counting the ridges, their number and length of span, the number of elements or loops, a drawing of the cross section and then deriving the trackplan from these details. The trackplan is a flat projection of the three-dimensional track the threads are following in the braid. The structure can then be compared with the existing contemporary instructions and with other analysed braids. The following are four examples of analyses of braid structures found in Europe. Tasselli, Cathedral Museum, Trier, Germany Tasselli are small, round, embroidered adornments on bishops’ gloves. The Cathedral museum in Trier possesses one set of two and a single one from excavations in the cathedral (Fig. 8.2). All three are made the same way but with different motifs in the

8.  Braided strings and Turk’s head knots on European secular and religious textiles 133

Fig. 8.1. Silk button maker (Christoff Weigel 1698). (© Deutschefotothek.de)

centre. The centre is embroidered and the edge covered with a narrow silk braid c. 4 mm wide. In Speyer, Germany, is a similar and contemporary tassellus edged with two braids of the same structure (Müller-Christensen 1972, 1005, fig.  1583). The structure is complex although the making is very simple when using loops. Two people each with five loops cooperate in making this structure. The complexity

134

Joy Boutrup

comes from the fact that some loops are passed over during braiding, a method now called unorthodox. Further complexity derives from the way the loops are exchanged between the workers, now called unorthodox exchange (Speiser & Boutrup 2009a, 18−20). This method gives a twist to both loops involved in the exchange (Fig. 8.3). Braids of this type are often found on early textiles, for example most of the braids on relic purses in the Basilica of Tongeren, Belgium; in Valeria, Sion, Switzerland (Speiser & Boutrup 2009a, 27−32); and in St Servatius, Maastricht, The Netherlands (Stauffer 1991). The finds from the London excavations also show a large number of this type of braid with two, three and four people cooperating (Crowfoot et al. 1992, 138−41; Boutrup 2010, 20−29). Relic purse, Vannes Cathedral, fourteenth century Vannes cathedral in Brittany has a collection of early relic purses (Plantec 1991, 86−99). Purse no. 811 was originally of red silk lined with yellow silk but the red silk is only preserved at the edges (Fig. 8.4). The drawstrings are compact and more or less round, the cross section has been changed in some places due to pressure. There are eight ridges alternating S and Z. A similar structure has been identified by Noémi Speiser in the Victoria and Albert Museum on a sudarium, which was probably originally from Halberstadt cathedral (V&A 8279−1863); also in braids on the girdle of Philip of Schwabia (d. 1208) in Speyer (Boutrup 2011); and in a seal string (Boutrup 2008, 11−16). There are two braids of this structure in the finds from the London excavations (Boutrup 2010, 21). There are variations in the structure that can easily be caused by mistakes during braiding. These variations are all consistent with the use of loops and is caused by variations in the way the loops are taken (Fig. 8.5). The structures can be made by two people holding four or five loops each. There is no exchange of loops involved as the two people work alternately (Fig. 8.6). They are actually producing two braids interlaced across each other. Similar methods of crossing two structures are mentioned in the Tollemache and Harley manuscripts, although in connection with another structure (Speiser 2000, 64). These two early examples are representative of the fact that these types of braids are not mentioned in the fifteenth-century instructions and that no orthodox exchange is used. Until around 1400 there seem to be only braids made by one person or, when two people work together, either no exchange is involved or when an exchange is used then only the unorthodox exchange. The early fifteenth-century instructions give a precise instruction for the orthodox exchange and from this time new types of braids appear. The next two examples are from the fifteenth century and show the use of the instructions in the Tollemache and Harley manuscripts for orthodox exchange. Aurifrisium, fifteenth century The aurifrisium in Halberstadt cathedral is about 12  mm wide, single layered, flat with eight gold threads and eight of blue silk (Fig. 8.7). There are 14 ridges, 1 1 1 1 1 1 1 2 1 1 1 1 1 1, almost plain interlacing. The first impression is that it is braided

8.  Braided strings and Turk’s head knots on European secular and religious textiles 135

Fig. 8.2. Detail of a tassellus from the excavations in Trier cathedral, early twelfth century, T170,9, Cathedral Museum, Trier

Fig. 8.3. Diagram of the structure and the corresponding trackplan of the tassellus. (© J. Boutrup)

with open ends, but two mistakes along the braid indicates that it is made with loops. In loop braiding this structure is folded along the midline of the braid during braiding and the legs of the loops move in different layers. If by mistake a loop is taken crossed instead of open the two layers connect and the finished braid will

136

Joy Boutrup

not open and lie flat. The photo shows such a mistake to the left. The braid can be made by two people, one holding five loops and one holding three loops. The five loops (to the left in the drawing) make plain interlacing using the method called ‘Spanish’ and described in the seventeenth-century manuscript of Lady Bindloss held among the Standish Family Papers (Speiser & Boutrup 2009b, 10−12). The exchange

Fig. 8.4. Detail of a relic purse from Vannes cathedral, Brittany, fourteenth century, purse no. 811, Vannes cathedral

Fig. 8.5. Trackplans of some of the structures and side view of one of the structures found in the braid from the relic purse. (© J. Boutrup)

8.  Braided strings and Turk’s head knots on European secular and religious textiles 137

Fig. 8.6. Diagram to show how the hands are held during braiding. Two people stand beside each other slightly turned face to face. The two near hands are crossed

Fig. 8.7. Aurifrisium from Halberstadt cathedral, fifteenth century, obj. no. 186. The braid is c. 12 mm wide

138

Joy Boutrup

Fig. 8.8. The upper diagram shows the folded trackplan and the lower diagram shows the steps made during braiding. If a loop is taken crossed instead of open the two layers will connect and the braid will not be able to open properly. (© J. Boutrup)

is orthodox and the loop from the left goes over the loop from the right as shown in the diagram (Fig. 8.8). Corporal box The braid is stitched to a multi-coloured velvet on the lid of the corporal box. It is flat and has 12 ridges with the following spans 2 2 3 2 3 3 2 3 2 3 2 2 (Figs 8.9 & 8.10). In the Tollemache and Harley manuscripts are several instructions for double exchange in order to make vertical stripes in braids. However, the braid on the corporal box is one of the few examples of this so far recorded. The braid is certainly loop braided and it corresponds exactly with the instructions in the fifteenth-century manuscripts. It is also braided while folded along the midline and contains the same type of mistakes as the aurifrisium. It is made by three people, one holding five loops of silver, the second holding five loops of red silk and the third holding five loops of gold. The exchange is made twice between the adjacent workers approximately ten times when gold is in the middle of the braid (on the third person’s hands). The exchange is then only made once between the workers and the colours will move from hand to hand. This is done until gold is again on the third person’s hands (Fig. 8.11).

Turk’s head knots Turk’s head knots look like tubular braids and are made by interlacing one thread with itself (Fig. 8.12). Several knots of the same size or of different sizes can be interlaced

8.  Braided strings and Turk’s head knots on European secular and religious textiles 139

Fig. 8.9. Corporal box from the Gdansk Paramentenschatz, St Anne’s Museum, Lübeck, No. M373. The braid is c. 10 mm wide.

Fig. 8.10. Diagram of the braid on the corporal box. (© J. Boutrup)

into each other, parallel to each other, interlacing with each other in twill or in overall plain interlacing to make a compound knot. The knots used for tassels and buttons are made around a wooden form which can be a ball, lentil shaped, pear shaped or olive shaped, etc. The form has a hole vertically through the middle, which is used to hold it with an awl while making the knot. It also serves for fastening the knot and for hiding ends of braided strings or cords attached to the tassel. Turk’s head knots are also used for gliding rings on purse strings, horse gear, etc. To make the ring stiff and stable it is normally lined inside with a roll of parchment, although the smaller knots are often made without anything inside. A compound Turk’s head knot is characterised by the number of bights, the number of knots in the compound, the so-called passes, and the number of parts in

140

Joy Boutrup

Fig. 8.11. Trackplans of the two structures of the braid on the corporal box. The upper trackplan shows the structure with single exchange, the lower the structure with double exchange. (© J. Boutrup)

Fig. 8.12. Bights and parts in simple Turk’s head knots. (© J. Boutrup)

each knot together with the type and shape (Figs 8.13 & 8.14). Each knot is in plain interlacing in itself. The knots can lie parallel or can, in combination with the other knots, form a sort of twill or tabby. The number of knots in a compound knot with parallel knots is easy to count. In compound knots in twill the span of the longest float in the twill ridges indicates the number of knots or passes forming the compound

8.  Braided strings and Turk’s head knots on European secular and religious textiles 141

Fig. 8.13. Compound Turk’s head knot with parallel knots. (© J. Boutrup)

Fig. 8.14. Analysis of a compound knot with parallel knots. (© J. Boutrup)

knot. There are two types of twill compound knots found on the investigated objects. In one type all knots are of the same size, all the bights at the same border and all ridges of the same width. This type is today called a herringbone knot. It is mainly used for gliders as it is cylindrical. The other type has knots of different or the same size but the borders of the knots are shifted in relation to each other so that the

142

Joy Boutrup

Fig. 8.15. Compound knots on a fourteenth-century pannisellus, three passes parallel, four bights, three parts, obj. 140, Halberstadt cathedral

twill ridges have different widths. This type is called pineapple knots and adapts to different shapes (Hall 1996, 62−88) The twill compound knots are called ‘spikat’ in the earliest known description of Turk’s head knots (Weigel 1698, 586). The early knots seem to be simple with few bights and parts and mainly of the type with parallel knots (Figs  8.15 & 8.16). At the end of the sixteenth century the more complex and larger knots in twill begin to appear (Figs 8.17 & 8.18) and in the seventeenth-century buttons are made with compound knots in overall plain interlacing (Hart & North 1998, 114−15; Pietsch & Stolleis 2008, 188−91) (Fig. 8.19). It is relatively easy to count the number of ridges in each knot as long as two or more colours are used. It is difficult though if all knots are of the same material and colour then other, more circumstantial, methods have to be used (Figs 8.20 & 8.21).

Conclusion It is too early to say anything definitive about when the different types of braids and knots occur but from the research so far, there seem to be the tendencies mentioned above. Some uncertainty is due to the fact that the dating of many of the textiles

8.  Braided strings and Turk’s head knots on European secular and religious textiles 143

Fig. 8.16. Detail of a fourteenth-century relic purse with a glider in the form of a Turk’s head knot, seven passes, four bights, three parts, purse no. 48, Tongeren Basilica, Belgium

Fig. 8.17. Analysis of a simple pineapple knot. (© J. Boutrup)

is very vague; the dating of purses for example is normally based on the fabrics and thus not very precise as the fabric often had been used for other functions for a long time before finally being recycled as a purse. A general chronology of the techniques is also hampered by the fact that the development or the introduction of fashions varied in different parts of Europe. It is a surprise that many of the structures are not mentioned in the sources and that there are very few finds with oblique twining while more than half of the fifteenth-century instructions belong

144

Joy Boutrup

Fig. 8.18. Analysis of a complex pineapple knot. (© J. Boutrup)

Fig. 8.19. Tassels on the seal string of James VI of Scotland, 1590, six passes, six bights, seven parts, three times, gold and five parts two times, rose silk. Ratification of a marriage contract between princess Anne of Denmark and James VI of Scotland. (© Danish National Archives)

8.  Braided strings and Turk’s head knots on European secular and religious textiles 145

Fig. 8.20. Different pineapple knots in compound tassels on a hunting horn belt, 1580–1590. The belt is tablet woven and the strings are loop braided. obj. W2771. (© Germanisches Nationalmuseum, Nüremberg)

Fig. 8.21. Compound knot in plain interlacing such as it is often found on buttons. Three passes, five bights, four parts, two knots are red, one is white. (© J. Boutrup)

to that type (Speiser  2000,  68−77). However, there are still many collections to be studied for a complete picture of the types and their occurrence to emerge. Turk’s head knots have been out of fashion in passementerie for a long time. Most of the literature on passementerie, which stems from the nineteenth and twentieth centuries, has no instructions for the elaborate Turk’s head knots. Instructions for these knots are mainly to be found in books on rope and sailing (Smith 1960, 29)

146

Joy Boutrup

or in books on leatherwork for horse gear (Grant  1950; Hall 1996). Furthermore, methods for the analysis of compound one-coloured Turk’s head knots have not yet been finally developed.

Bibliography

Boutrup, J. 2008. Braided seal strings from 1590. Strands 15, 11−16. Boutrup, J. 2010. Analysis and interpretation of braids from the London excavations. Archaeological Textiles Newsletter 50, 20−29. Boutrup, J. 2011. Die Schnüre am Gürtel Philipps von Schwaben. In M. Herget (ed.), Des Kaisers Letze Kleider (Munich, Historisches Museum der Pfalz Speyer), 114–117. Braun-Ronsdorf, M. 1956. Die Taschenmacher Zünfte. CIBA Review 129, 26−60. Crowfoot, E., Pritchard, F. & Staniland, K. 1992. Textiles and Clothing c.1150−1450. Medieval Finds from Excavations in London: 4 (London, HMSO). Depping, G.D. 1837. Règlemens sur les Arts et Métiers de Paris rédigés au XIIIe siècle et connus sous le nom du Livre des Métiers d’Etienne Boileau. Collection des documents inedits sur l’histoire de France, Series 1 (Paris, Crapelet). Grant, B. 1950. Leather Braiding (Centreville, MD, Cornell Maritime Press). Griffith, J. 1997. The Tollemache Book of Secrets Facsimile (London, The Roxburghe Club). Hall, T. 1996. Introduction to Turk’s Head Knots (self-published). Hart, A. & North, S. 1998. Historical Fashion in Detail: The Seventeenth and Eighteenth Centuries (London, V&A Publications). Heutte, R. 1972. Le Livre de la Passementerie (Dourdan, H. Vial). Müller-Christensen, S. 1972. Die Graber im Konigschor. In H.E. Kubach & W. Haas, Der Dom zu Speyer (Munich, Deutscher Kunstverlag), 923−1024. Pietsch, J. & Stolleis, K. 2008. Kölner Patrizier- und Bürgerkleidung des 17. Jahrhunderts. Riggisberger Berichte 15 (Riggisberg, Abegg-Stiftung). Plantec, M. 1991. Les sachets à reliques de la cathédrale de Vannes. In B. Simon, L. Smets & A. Van Hoof (eds), Medieval Textiles, Particularly in the Meuse-Rhine Area (Sint-Truiden, Provinciaal Museum voor Religieuze Kunst), 86–99. Smith, H.G. 1960, The Marlin Spike Sailor (Camden, ME, McGraw-Hill). Speiser, N. 1983. The Manual of Braiding (Basle, self-published). Speiser, N. 2000. Old English Pattern Books for Loop Braiding (Basle, self-published). Speiser, N. & Boutrup, J. 2009a. European Loop Braiding, Part I (Leicester, Jennie Parry). Speiser, N. & Boutrup, J. 2009b. European Loop Braiding, Part II (Leicester, Jennie Parry). Speiser, N. & Boutrup, J. 2011. European Loop Braiding, Part III (Leicester, Jenny Parry) Speiser. N. & Boutrup, J. 2012. European Loop Braiding, Part IV (Leicester, Jenny Parry). Standish Family Papers, Wigan Record Office, Leigh, Lancashire (D/D St Bundle C2-1). Stauffer, A. 1991. Die mittelalterlichen Textilien von St Servatius in Maastricht. Riggisberger Berichte 8 (Riggisberg, Abegg-Stiftung). Weigel, C. 1698. Abbildung der gemeinnützlichen Haupstände (Dresden, Sächsische Landesbibliothek und Universtatsbibliothek).

Chapter 9 A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle, East Tyrol Beatrix Nutz, Rachel Case and Carol James

Introduction In Lengberg Castle, East Tyrol, a perplexing textile fragment made of humble fabric but with exquisite detail was found beneath the floorboards among thousands of objects. It is of white linen cloth with a variety of decorative elements in linen thread. These decorative elements include needle lace, loop braids, and most strikingly, sprang. All three textile techniques were worked with an undyed linen thread. The complex tree of life pattern of the sprang suggests this is a noteworthy garment. It was meant to be seen. Since the original function of this textile could not be determined with certainty by merely investigating the extant garment, a reconstruction was attempted in order to produce a copy with which wearer trials could be conducted. Thanks to a grant from the Janet Arnold Foundation, new light can now be shed on this mysterious textile, and how it was worn. After frustrating failed attempts, a surprising solution was determined: the textile is headwear.

Archaeological context In the course of extensive reconstruction at Lengberg Castle (Municipality Nikolsdorf, East Tyrol, Austria) beginning in July 2008, archaeological investigations of several parts of the building were carried out under the direction of Harald Stadler (Institute for Archaeologies, University of Innsbruck). During the project a filled vault was detected in the south wing of the castle in room 2.07 on the second floor. The backfill was taken out by the construction company’s workers and stored for subsequent sieving, which took place in summer 2009. The fill consisted of dry material in different layers, among them organic matter such as twigs and straw, but also worked wood, leather (mainly shoes) and textiles. The building’s history as well as

148

Beatrix Nutz, Rachel Case and Carol James

investigations on construction techniques carried out by architectural historians of the Landeskonservatorat Tirol and the archaeological features suggest the finds date to the fifteenth century. This date has now been confirmed by the radiocarbon-dating of five textiles carried out at the ETH-Zurich. Lengberg is first mentioned in a document from 15 August 1190, in which a donation of Count Heinrich of Lechsgemünd to the monastery of Viktring in Carinthia was confirmed and Volcarth, Caloth and Otto de Legenberch, ministerialis to the House of Lechsgemünd, were named as witnesses.1 In 1207, Heinrich of Lechsgemünd sold the castle to Archbishop Eberhard of Salzburg. Until 1803, the castle was property of the archbishops of Salzburg who assigned its administration to various nobles (Hörmann 2003). Two administrative periods are relevant to the Lengberg textiles and therefore deserve closer attention. In 1419, the brothers Andre and Peter Mosheimer received the castle for lifetime care (‘Burghut’ or ‘Pflege’). At first, Peter Mosheimer and his wife Anna inhabited the castle. After Peter’s death in 1453 (or 1454), his brother Andre became lord of Lengberg. In turn, Andre’s son, also named Peter, was responsible for Lengberg from 1465 to 1480 when Virgil of Graben (d. 1507), one of the most powerful nobles of his time in the region of Upper Carinthia and East Tyrol, was assigned lifelong lordship of Lengberg (Plattner 2012, 39–40). Virgil of Graben had the old castle ‘Veste Lengenberch’ reconstructed into a new prestigious residence by adding an additional floor. It was during this modification that the vault between the roof of the first floor and the floorboards of the new second storey was filled with waste. This remodelling is mentioned by Paolo Santonino (Santonino 1947, 34–37) in his itinerary, where he also gives us a short description of the castle and mentions the consecration of the chapel by Pietro Carlo, bishop of Caorle (1472–1513), on 13 October 1485. It may be assumed that the vault was filled with waste during the addition of the second storey in order to level the floor. Taking into consideration that the reconstruction was probably finished by the time the chapel was consecrated, the majority of the finds (except for small pieces that later fell through gaps in or between the wooden floorboards) must predate October 1485. It is unlikely that the consecration took place while the construction was still underway, because the chapel and the room with the finds were situated in the same wing of the castle, the chapel was on the first floor and the room with the filled vault underneath was on the second in the fifteenth century. In addition, Paolo Santonino did not mention ongoing reconstruction work when he described Virgil of Graben’s rebuilding project.

Enigmatic linen fragments There are three fragments from Lengberg Castle featuring hole designs in patterns created using the thread manipulation technique of interlinking, most likely sprang (Nutz 2018). The largest sprang piece is quite elaborate and is still attached to a plain-weave, undyed linen fabric woven in Z-spun threads with a fabric density of fifteen threads per cm (38 threads per inch). With a warp count of 120 threads (=

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 149 60 loops), the sprang design is not entirely symmetrical. The person who made the sprang miscounted the warp threads and made the pattern slightly off-set, so the design is not centred. This asymmetry is almost indiscernible in the finished piece and was most likely an acceptable imperfection. In the light of some other finds from Lengberg and with it being impossible to ‘test-wear’ the fragile extant piece, the textile was initially thought to be a supportive undergarment (Nutz 2013a, 223). Only after several attempts at reconstruction and tryings-on did the original use become apparent. The sprang textile proved difficult to reconstruct as a supportive bust garment because the physical dimensions of the ‘bra’ did not easily match with female anatomy. It would have had fit a woman with a very narrow ribcage and extremely full breasts. The main panels of linen were too small, and the sprang did not stretch enough to accommodate breasts gracefully. There is no discernable method of fastening, and the placement of the still-tied knot, the location of the fingerloop braid and needle lace edging, and the roughly finished ‘top edge’ of the sprang indicated another possibility: headwear. But what type of headwear? Naming the garment is quite a challenge. Terms like Gebende, Rise, Kruseler, Hennin and the like blend historical and historicizing concepts to seemingly authentic names of precisely defined clothes (Zander-Seidel 2010, 37), and the Lengberg textile does not quite fit the descriptions of what are commonly known as a coif, cap or veil either. Therefore, the term ‘headwear’, which can be applied to any type of garment meant to be worn on the head, is used throughout this article for the Lengberg find.

The headwear from Lengberg The extant headwear from Lengberg (Fig. 9.1, pattern Fig. 9.2) consists of four pieces of plain-weave natural linen. The main side sections measure approximately 31 cm (12.2 inches) long, with the width tapering from 13 cm (5.11 inches) at the widest point near the mid-crown to 7.5 cm (2.9 inches) wide at the point connecting to the ties. One of the panels is ripped away nearest the needle lace, 3.5 cm (1.37 inches) from the fabric edge. These panels are connected with a large central section of patterned sprang, also in natural linen, measuring 27 cm (10.6 inches) in length, which makes up the centre panel of the headwear and creates an attractive effect from the forehead to the back of the head. The sprang pattern resembles a tree of life (Fig.  9.2 left), repeated one and a half times. The last seven rows approaching the centre (Collingwood 1974, 32) were worked by interlinking four warp threads (two pairs), where the sprang was cut and sewn with rough stitches to prevent unravelling, thus narrowing the fabric towards the back of the garment. An almost identical tree of life with two birds at the foot of the tree was embroidered in drawn thread work using blue and white linen thread on an antependium.2 This textile from the abbey of the Benedictines in Engelberg, Switzerland, is dated to the fourteenth century (Schmedding 1978, 221–22).

150

Beatrix Nutz, Rachel Case and Carol James

Fig. 9.1. Panelled sprang headwear from Lengberg Castle. A. Detail of sprang with needle lace and loop braided lace at the bottom. B. Sprang headwear as found with ties knotted. (Photos: Beatrix Nutz)

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 151

Fig. 9.2. Pattern of sprang headwear. The pattern was created according to the extant cap’s measurements. It will fit a person with a head measurement of about 57 cm (following the hairline) = 22.44 inches. (Drawing: Rachel Case & Beatrix Nutz)

The ties, which were still knotted together when the garment was recovered, are connected to the main body of the headwear in two different needle lace patterns (Fig.  9.3. B & C), with both the main piece and the ties using the selvedges of the fabric. The ties have a preserved length of 29.4 and 32.5  cm (11.5 and 12.8  inches) respectively, from the selvedge to which the needle lace is sewn to the torn and narrow ends, with a width of 5 and 6.35 cm (2 and 2.5 inches).

Fig. 9.3. Needle lace patterns from Lengberg Castle. (Photos and graphics: Beatrix Nutz)

152 Beatrix Nutz, Rachel Case and Carol James

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 153

Fig. 9.4. A. Extant sprang headwear from Lengberg Castle (left) and reconstruction/recreation (right and centre). B. Headwear with sprang? (left), Master of Uttenheim, Birth of the Virgin, altarpiece (© German National Museum, no. Gm1180) and recreation (right and centre). (Photos: Beatrix Nutz, Rachel Case & Andreas Blaickner; Recreations: Rachel Case & Carol James)

A simple decorative loop braid and needle lace (Fig. 9.3. A) is added to the edge of the main panels and the top (front) of the sprang. The somewhat mysterious placement of needle lace sewn to the selvedges of the fabric between the main part of the garment and the ties may indicate a limited fabric width, that the maker was working with small scraps of linen in order to reduce fabric use or that it was simply to add more decorative elements. A narrow hem (0.5 cm; 0.019 inch wide) along all the cut edges of the panels and ties is finished with whip stitches.

The reconstruction of the Lengberg headwear A reconstruction of the garment (Fig. 9.4. A) from Lengberg Castle was made using materials and methods that approximate those of the fifteenth century. Only with

154

Beatrix Nutz, Rachel Case and Carol James

Fig. 9.5. Women working on sprang frames. A. Fresco at Crngrob, Slovenia, 1460. From W.F. Schweitzer, Textile Making Methods Depicted in Frescos, CIBA Review 1966 (1), 11, fig. 10. B. Fresco at Sveti Primoz, Slovenia, 1504. (© Institut für Realienkunde – University of Salzburg). C. Woman sitting at a sprang frame, book illustration ‘The Planet Venus and her Children’, early sixteenth century. (© Bibliothèque Nationale Paris, Ms. Allemande 106, fol. 62r)

the help of the reconstruction that made trial wearings possible, could the original use be determined as headwear. Antique, hand-woven, natural linen from Germany with a fabric density of 17 threads per cm (43 threads per inch) was used for the side panels and ties, and hand-spun plied linen thread within the range of the thread diameters of the extant headwear was used for the sewing thread, needle lace and sprang. It is very difficult to find tightly spun and plied commercial linen thread with a spin/ply angle of c. 45°, so linen thread was commissioned from a spinner, who spun the flax on a spinning wheel and finished it by hand. The sprang pattern chart used in the reconstruction and created by Carol James is based on Beatrix Nutz’s drawing of the pattern.3 The sprang was created on a wooden floor loom, made to resemble and function as a fifteenth century sprang frame (Fig. 9.5). It was necessary to create a 65 cm (25.6 inches) warp in order to fit the 27 cm (10.6 inches) sprang length for the headwear. Sprang is created at one end of the warp, which results in a mirror image at the other end of it.4 No temporary shed sticks were used, but to achieve the very dense sprang needed each row was beaten with a weaving sword as it was completed (Fig. 9.6). The finished, non-stretched width of the sprang is approximately 8.5 cm (3.34 inches); the original width was difficult to

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 155

Fig. 9.6. A. Sprang in progress. B. finished sprang. (Recreation and photos: Rachel Case)

determine due to extensive damage. The central sprang panel is attached to the side panels with whip stitches. The raw edges of the fabric were turned under twice to the inside of the fabric, and sewn using whip stitches, creating the narrow hem. The needle lace between the main body of the headwear and the ties (Fig. 9.3. B & C) was created by ‘casting on’ the stitches along the main headpiece and working them from left to right, back and forth, turning the piece over for each row in order keep the stitches uniform in direction. Once the pattern was complete, the lace on each was whip stitched to the ties. The needle lace along the face edge of the headwear (Fig. 9.3. A) was a simple one-row pattern worked right to left and whip stitched to an almost flat five-loop fingerloop braid, which in cross-section has one rounded face and one flat face. The pattern for the braid is called a broad lace of five loops, and it is a quite simple and quickly made braid (Benns & Barrett 2005, 39; Speiser 2000, 49; Nutz 2014, 116). Wearing the reconstructed headdress not only revealed the beauty, but also the functionality of the decorative elements. The ties are crossed at the back, covering the somewhat unsightly rough sewing of the sprang end, brought around the front of the head, knotted and can be tucked under if the ties are kept long; otherwise they are tied with the ends left showing. In several of the images from the visual record, ties can be discerned as a bunched band around the crown of the head. The placement of the still-tied knot on the extant headwear indicates that it too was tied at the front of the head. The sprang panel has stretch, which can expand to accommodate hair, maybe worn in braids wrapped around the crown of the head. Wearing the headdress before and after adding the loop braid and needle lace edging revealed its practical use as it provides the grip needed along the hairline to keep the headwear from

156

Beatrix Nutz, Rachel Case and Carol James

shifting during wear. In addition, it frames the face attractively. The symmetrically worn and ripped areas on the main side panels of the extant textile would have been at the ears, which is a natural place for it to incur damage.

Needle lace and/or loop-braided lace edging on other garments from Lengberg Several fragments from Lengberg Castle use loop-braided lace and needle lace on the edges, and range from supportive bust garments to headwear and a shirt. The edging seems to be simply decorative in some cases, while providing grip and structure in others. The shirt edge (Fig. 9.7. A) (Nutz 2015, 97, fig. 4) is decorative, made along the neckline, and appears to serve no other purpose. It is done with knotted buttonhole loops (Emery 1994, 36), in what Therese Dillmont calls the thirty-third lace stitch (Dillmont 1890, 462). Only small sections of this lace are preserved, and it cannot be determined if it was ever finished off by sewing a braided lace to the ‘outer’ edge. The loop braid and needle lace edges on two fragments of another type of headwear, initially thought to be supportive bust garments too (Fig. 9.3. D & E), may have also provided necessary grip to hold them in place. The pattern of the decorative edge from Lengberg seen in Fig. 9.3. E can be found on headwear from the visual record as well (Fig. 9.8). There, the lace probably served the same purpose of providing grip as it did on the Lengberg headwear. Another garment of undetermined function which might have had a secondary use as an infant’s garment (Nutz 2015, 100, fig. 6) features a loop-braided lace only. Needle lace is also found as a decorative element on the skirted bra (Case et al. 2017, 169, fig. 1/3)(Fig. 9.7. B). Two fragments of needle lace are no longer attached to their garments. One fragment is still connected to a loop braid (Fig. 9.7. D) and done with simple looping, or buttonhole looping (Emery 1994, 31). Dillmont speaks of plain net stitch, or first lace stitch (Dillmont 1890, 450).The other fragment stands alone and is done in using the fourteenth lace stitch (Fig. 9.7. C) (Dillmont 1890, 455). While the needle lace from Lengberg is mainly worked onto a fabric edge or hem, and the braided lace sewn onto the needle lace, the lace in Fig. 9.7. D is worked directly on the edge of the braided lace.

The narrow sprang strips from Lengberg A narrow sprang fabric (Fig. 9.9. A, pattern chart Fig. 9.10. A) of which two fragments survive was originally worked with 32 warp threads (= 16 loops), now missing six to seven warp threads due to also being torn lengthwise. The average thread diameter of the S-plied 2-ply thread is 0.6 mm (0.02 inch). The smaller fragment, which shows the same pattern, is sewn onto a plain-woven linen fabric of Z-spun single thread with 13 to 15 threads per cm. The preserved length of the larger fragment is 15 cm (5.6 inches). The pattern has holes worked in five rows combined with holes of twice the length worked in seven rows of interlinking (Collingwood 1974, 148−49).

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 157

Fig. 9.7. Needle lace and loop braided lace from Lengberg Castle. (Photos and drawings: Beatrix Nutz)

The second sprang strip (Fig. 9.9. B, pattern chart Fig. 9.10. B) with a non-stretched width of 2.2 cm (0.86 inch) and a preserved length of 11 cm (4.3 inches) was made with a slightly finer plied thread 0.5 mm (0.019 inch) in diameter. It too was worked in a hole pattern with 48 warp threads (= 24 loops), the holes arranged in the shape of a flower again worked in seven rows. The exact purpose of these two narrow sprang fabrics cannot be determined, but considering that the large sprang is part of a headdress, and one of the two narrow sprang pieces was clearly sewn to a woven fabric, their stretchy properties could have been used in a coif or cap to accommodate various head sizes and quantities of hair.

158

Beatrix Nutz, Rachel Case and Carol James

Fig. 9.8. Needle lace and/or loop-braided laces as edging. A. Extant textile from Lengberg. B. Midwife. Master of the Crucifixion Triptych, Birth of the Virgin, private collection. (Photo: Beatrix Nutz. Painting reproduced in Otto Benesch, Der Meister des Krainburger Altars, Wiener, Jahrbuch für Kunstgeschichte VII (1930), 191, fig. 56)

Head coverings with narrow netted strips are depicted in the Landshuter Ringerbuch (see Fig. 9.13. L).

Other extant headwear Although the Lengberg headwear is a unique find, there are at least two other extant coifs or caps that feature decorative elements using thread manipulation and needlework. The most well-known example is probably the St Birgitta’s cap or huva from fourteenth century Sweden. The cap is made from two panels of very fine

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 159

Fig. 9.9. Narrow strips of sprang possibly for panelled sprang headwear (see Fig. 9.13. L). (Photo and drawings: Beatrix Nutz)

160

Beatrix Nutz, Rachel Case and Carol James

Fig. 9.10. Sprang pattern charts for narrow sprang strips from Lengberg Castle. (© Carol James)

plain-weave natural linen, with a continuous loop tie that wrapped around the head to secure it. The midline of the two halves of the cap is decorated with intricate open embroidery running from the forehead to the nape of the neck. There is a narrow embroidered strip of fabric along the face edge that continues

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 161 from the right side nape as a long tie torn off at the end. The tie on the left side, which was sewn on to the binding strip, was broken too, with only a short piece remaining. The line of embroidery is along the face edge of the strip only and does not continue onto the ties. At the back of each panel, tiny pleats allow for fullness in the cap to accommodate the hair. The conservators believed that the ties once formed one long narrow band and therefore the ends were sewn together. The cap might have been worn with the ties crossed at the back, brought around the top of the head, with the loop resting at the crown of the head. The tie is attached at the back, making it necessary to wrap it around the head to secure it. This method of tying the cap is seen in depictions of women wearing a similar piece of headwear, especially in the Maciejowski Bible (also called the Crusader Bible).5 The cap is quite small, measuring approximately 45 cm (17.5 inches) along the face edge, and would have fitted a small-framed woman or a girl (Dahl & Sturtewagen 2008, 99–129). The open embroidery of the cap of St Birgitta is reminiscent of the needle lace of the Lengberg headwear. The function of the open embroidery may simply be decorative. Another embellished coif from the Middle Ages is in a private collection. It is part of an assemblage of discarded garments and furnishing fabrics uncovered beneath the floorboards of an Italian castle. It has been determined that the textiles were discarded between 1470 and 1540, but the coif itself may be older. Detailed technical data on this coif, or of any of the other finds, have not been published to date, as investigation into the finds from this collection are still a work in progress.6 The Italian coif is made from fine, plain woven natural linen with a thread density of 18 threads per cm. The thread used to sew the coif and its embellishments appears to be of linen. It is constructed with three panels, each panel connected by decorative ‘lace’ very similar to St Birgitta’s cap. The coif ’s face edge also has a decorative element similar to that in the Lengberg pieces that also feature loop braids and needle lace.7 Remnants of braided ties are still attached to the front corners.8 The placement of the ties at the side of the face indicates that they would be tied under the chin or allowed to hang loosely, which appears to fit with depictions of male coifs from the fourteenth and fifteenth centuries in construction, but there is no evidence in pictorial depictions of decorative seams on men’s headwear as in this coif.

The three techniques The decorative needlework or thread manipulation techniques applied to the Lengberg textiles include looping (needle lace), loop manipulation (fingerloop) braiding and interlinking, most likely sprang (Emery 1994, 60−61, 68−70). All decorative elements are worked using undyed, Z-spun S-plied two-ply thread with diameters ranging from 0.6 to 0.9 mm (0.023 to 0.035 inch) and twist angles of 37° to 42° (Fig. 9.11).

162

Beatrix Nutz, Rachel Case and Carol James

Fig. 9.11. Microscope images of the plied yarn used for A. loop braided lace, B. needle lace and C. sprang. Dino-Lite Microscope with a magnification of 35×, 55× and 60×. (Photos: Beatrix Nutz)

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 163 Looping – needle lace With looping the fabric structure is built up by repeated interworking of a single continuous element, a thread, with itself (Emery 1994, 30). Needle lace or needle point is defined as ‘all forms of open work constructed stitch by stitch with a needle and thread’. In French, point a l’aiguille, point coupé, vélin. In Italian, punto in aria, trina ad ago. In Flemish, Naaldkant. In German, Nadelspitze (Levey 1983, 122). In contrast to reticella lace, a cutwork in which threads are pulled from linen fabric to make a ‘grid’ on which the pattern is stitched using buttonhole stich, needle lace requires no woven foundation but is sewn ‘in the air’ using various buttonhole or lace stitches (Dillmont 1890, 450–65). Needle lace is either built up stitch by stitch on an outline structure of thick threads tacked along the lines of a design drawn out on a parchment pattern or the stitches are merely looped into the stitch above. The simple buttonhole stitch can be varied by giving it one or more twists or by knotting it (Levey 1983, 1–2). Needle lace is thought to have begun in the late fifteenth century and developed over the course of the sixteenth century. It is debated from whence the punto in aria, first mentioned by Giovanni Antonio Tagliente in 1527 in his pattern book Opera nuova che insegna alle donne a cusire (Tagliente 1527) can be considered fully-fledged needle lace (Levey 1983, 6–7). The needle lace found at Lengberg Castle shows that the beginning of needle lace must date back further. One of the textiles with needle lace has been radiocarbon-dated to the first half of the fifteenth century (ETH−40538, 510±35 BP, δ13C -22.9±1.1, 86.1% probability: 1390 to 1450). While this needle lace is far from the elaborate and complicated needle lace of later times, it adheres to the definition of needle lace, that it is worked with a needle and a single thread in a succession of buttonhole stitches. There is no thread used as a ‘grid’ for the lace stitches. It cannot be determined if a temporary parchment backing to work the lace was used at Lengberg, but here the patterns are fairly simple and can probably be worked without a pattern base. Loop manipulation – fingerloop braiding The loop manipulation braiding technique (or fingerloop braiding) is a handbraiding technique requiring no assistance from tools except for an optional beater. The working ends of the threads are paired and form loops rather than being separate as in braiding with open ends (plaiting). One end of the bunched braiding threads is fixed on a support, and each loop is mounted on a finger of the hands (finger-held method). Alternatively one may slip the sequentially arranged loops around the hands (hand-held method). To make a braid, you exchange the loops, one at a time, between two hands following a prescription (Kinoshita n.d.). In Europe, holding the loops with your fingers is the only method proven, whereby depending on the number of loops, a person is working alone or together with one or more partners. The method allows fast work and is especially suitable for thin, fine material and preferable to the more time-consuming method of

164

Beatrix Nutz, Rachel Case and Carol James

plaiting with open ends. In the European Middle Ages and the Renaissance loop braiding was the most important production method for braided laces and bands and has been exercised by both professional craftsmen as well as in monasteries and private homes (Boutrup 2010, 2011). In Europe this technique goes back to at least the twelfth century. A ribbon made of silk thread, worked with seven loops from London (Crowfoot et al. 1992, 138) and a ribbon as part of a tassel on a liturgical glove from Speyer Cathedral (Speiser & Boutrup 2012, 40) are among the oldest surviving examples. The oldest depiction of loop braiding known to date can be found in the Haus zur Kunkel in Constance in south Germany containing several frescoes dated to before 1306 (Wunderlich 1996). The weaver fresco shows 21 scenes of women engaged in textile production, each scene with a sentence in medieval German above relating to the work in progress. In one scene two workers, a girl dressed in red and a young woman dressed in green, are occupied with loop braiding. The first instructions for the braiding of laces written in English and German date to the fifteenth century, for example the manuscript MS Harley 2320 in the British Library and the Haus- und Arzneibuch – Donaueschingen 793 (Badische Landesbibliothek Karlsruhe 1445/1470). But loop braiding could date back even further as an Iron Age find from the salt mines at Hallstatt shows (Grömer et al. 2015). However, many laces can be braided with loops or open ends, the result looking (almost) the same for both methods. A sure proof that a lace has been braided using loops is given when the loops have not been cut off after finishing the lace and can still be seen. This is the case with several laces from Lengberg (Nutz 2014, 122), one even having been sewn onto a garment, probably a headdress. The lace was simply tied off with a knot and the loops left visible (Fig. 9.12).

Fig. 9.12. Needle lace and loop braided lace with loops sewn onto a fabric edge. (Photo: Beatrix Nutz)

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 165 Sprang – interlinking Sprang can be done by interlinking, oblique interlacing and oblique twining of one set of elements with the two-way duplicating method on a frame with the threads fixed at both ends (Emery 1994, 60, 69). This two-way duplicating method creates a midline when both ‘ways’ meet, thus testifying to the technique used. Yet, since the workpiece can be cut in half before the middle is reached – something that is done quite frequently as examples show (Rutlin et al. 1964, 298−300, figs 6−10) – this is not a fool-proof method of identifying sprang. Damaged extant pieces like the finds from Lengberg complicate matters even more. Is the midline missing because it never had one, has the piece been cut in half or has the midline been torn off? Therefore it is important to note that although these textiles were most likely produced using the interlinked sprang method (manipulation of stretched threads on a frame), it is possible to create similar structures using other methods such as interlinking using a needle or plaiting with open ends. Without evidence of a midline and the reversal of the twist direction, it is almost impossible to discern how interlinking is created. However, in using the sprang method to recreate the textiles from Lengberg exact replicas were produced, strongly suggesting that they were most likely originally made using the sprang technique. The word ‘sprang’ that we use today to name this technique is of Scandinavian origin, originally meaning an openwork textile. In Norway they also consider firfletting to be a variety of sprang, but decided it needed a name of its own (Rutlin et al. 1964, 301–302). Firfletting involves manipulating warp threads without the use of a loom, but it looks very different in structure compared to sprang. It makes a similar looking ‘lace’, but does not twist the same way as sprang as it is three to four threads braided together, and therefore rather bears similarities to early bobbin lace (Nutz 2019, 42). In an inventory from a guild in Hamburg (northern Germany) spranck is mentioned, which may also be related to our modern word (Collingwood 1974, 34, 44). Peter Collingwood lists a number of medieval written records from northern Europe from 1328 to 1578, but no central or southern European sources. So what about southern Germany and Austria? To date no written records relating to sprang could be detected for this region, but it may be because we do not know what this type of work was called in the fifteenth century. In a book published in 1800 in Leipzig the technique is referred to as Rahmenstricken (= frame knitting) or Greifstricken (= pick knitting) (Netto & Lehmann 1800, 32). Should it indeed be that the word stricken (knitting) was used to describe sprang fabrics as well as (in modern terms) actually knitted ones, we will not be able to distinguish between the two. By the twentieth century the technique was all but forgotten in most central European countries, but even in some of the countries with an unbroken tradition of sprang production, there is no distinct name for the technique. In Slavic languages such as Czech, sprang is also called knitting (= pletení) on a frame (= rámu) or loom (= krosien) or knitting without needles (= pletení bez jehlic) (Kindlová & Tomková 2001; Mihalić 2013; Kielbusová 2014). In French publications (Paulis 1930, 1933) and on museum websites

166

Beatrix Nutz, Rachel Case and Carol James

(e.g. the Cinquantenaire Museum in Belgium) the technique of sprang is named le point de dentelle which translates to ‘lace stitch’. On some Neolithic pots from Rietzmeck, Kreis Rosslau, Germany, there are impressions of fabrics interpreted as interlinked in structure, yet there is no evidence that they were made by the sprang method (Schlabow 1960; Collingwood 1974, 37). The oldest extant European sprang textiles date back to c. 1400 BC, sprang hairnets from Bronze Age burial sites in Denmark (Collingwood 1974, 38). Sprang fabrics dating to the fifteenth century are a wool fragment (42.5 × 35.6 cm) dating to c. 1450 to 1500 from London, today housed in the Metropolitan Museum of Art in New York (Accession Number: 28.197) and a linen table cloth (320 cm × 75 cm) worked with a circular warp from central Switzerland dated 1400 to 1450 in the National Museum in Zürich (Inv. No. LM-22119) (Schneider 1975, 73−74, 232−33). Both fabrics are in hole design, the table cloth with leaf motif and diamond patterns with stylised lions, eagles, trees and geometric figures. The textile from London is one half of a sprang fabric, cut in the middle to form a fringe. Pictorial sources of the fifteenth and sixteenth centuries also show, that working interlinking on a frame was done in late medieval and early modern times. On the fresco Sveta nedelja (Holy Sunday) in the Annunciation Church in Crngrob in Slovenia dating to 1460 a woman is sitting in front of a free-standing frame on which threads are stretched vertically and her hands are in a working position common for sprang (Fig. 9.5. A). Another, more detailed church fresco in the church Sveti Primož (St Primus and Felician) at Kamnik, also in Slovenia, dating to 1504, depicts the Virgin sitting in front of a sprang frame (Fig. 9.5. B) (Schweitzer 1966, 10–17). From the beginning of the sixteenth century is the book illustration ‘The Planet Venus and her Children’ in a manuscript from the Upper Rhine (region between Basel, Switzerland and Bingen, Germany). The woman sitting at the sprang frame seems to be working on an open fabric (Fig. 9.5. C), maybe a hairnet (Drinkler 2010, 19).

Pictorial evidence of netted or sprang headwear To date no other extant example of this garment with sprang, needle lace and loop braided lace that was found at Lengberg Castle is known, but several pictorial examples substantiate its interpretation as headwear. The earliest image found depicting someone wearing headwear resembling the sprang garment is in an Austrian manuscript from 1435 (Fig. 9.13. A) which shows a woman, probably a nurse, attending a deathbed. This fits with the radiocarbon-dating of the sprang headwear to 1390 to 1450. The altarpiece by the Master of Uttenheim, Birth of the Virgin, dating to the second half of the fifteenth century, features perhaps the best image of a headdress with a central sprang or net panel. Based on the image of this headdress, worn by the midwife in the bottom left corner of the painting, Carol James created a pattern chart for a central panel made in the technique of sprang and recreated the headwear (Fig. 9.4. B).

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 167

Fig. 9.13. Late medieval and early modern images of sprang or net headwear. A − F Women. G − L Men. (Graphics: Beatrix Nutz)

168

Beatrix Nutz, Rachel Case and Carol James

Most of the depictions of this type of headwear are men, but there are several examples of women wearing headwear with panels of fabric and sprang or netting (see appendix & Fig. 9.13). Many seem to be workers or labourers, not the nobility, but are working class and perhaps middle class. These people, both male and female, are servants, clerks, cooks, nurses or midwives, executioners, athletes (wrestlers and fencers who may be nobles) or doctors. One exception has been found so far in the Thun-Hohenstein Album – a man in armour who is wearing this type of headwear under his helmet. Pierre Terjanian wrote: ‘Both the armour and the helmet with buffe in this drawing may be dated on stylistic grounds to c. 1480. … He may conceivably be, as Leitner has suggested, Archduke Sigismund of Tyrol (1427–1496)’ (Leitner 1888, VI; Terjanian 2013, 346–47). This may be a garment worn to keep hair out of the face, but more elaborate than a coif or kerchief, a kind of formal headwear. It is interesting to note that many examples of the headwear worn by women in the imagery have a smaller profile than those worn by men. In the estate account of Lazarus Holzschuher of 1545 a guldin Hauben mit sampt dem wulst (a golden cap/coif with the wulst) is listed among the other possessions. Jutta Zander-Seidel wrote that this points to men also having worn a wulst under their coifs, similar to woman (Zander-Seidel 1990, 230). Although the extant headwear is made completely of undyed linen fabric and thread, some pictorial examples of this panelled garment seem to indicate that they could have been made with other textiles such as colorful silk or wool in green, red, yellow or black fabric. The central panel could be sprang, but could also be netting made by a different technique or another textile.9 The panelled headwear seems to be unique to Germany and Austria in the fifteenth and early sixteenth century as to date no contemporary images of this type of headwear could be found in sources originating in other areas of Europe. As long as no other extant garment like the one from Lengberg is found, it is necessary to rely on Austrian and German artwork of the period between c. 1435 and 1520 for comparisons.

Written sources on sprang headwear No written sources have been found on sprang headwear to date, although there are documents that mention women’s caps in inventories, wills and laws during the Middle Ages (Dahl & Sturtewagen 2008). There are two reasons mainly responsible for the lack of written sources on sprang headwear. First and foremost, we do not know what sprang was called in the fifteenth century in German speaking areas (where the pictorial sources known to date originate) or even if there was a specific term for it (see above ‘Three Techniques – Sprang’). The second reason is that garments and their construction are rarely described in detail in written records, and the modern German words for different types of headwear are rather imprecise. The German term Haube probably applies best to the Lengberg headwear, but even that word is ambiguous. The dictionary of the Brothers Grimm defines Haube as: male headgear of different shape and for different social ranks. Generally, a men’s fashion in the Middle Ages

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 169 and later.10 In another lexicon, a Haube (French: coiffe) is ‘a light, roundish headgear, depending on fashion and rank of very different styles, at the end of the Middle Ages very baroque shape … especially in the costume of married women.’11 The definition for Haube in the current version of the Duden is: ‘Middle High German: hūbe, Old High German: hūba. 1. Headgear made of soft or starched, often pleated fabric, adapted to the head [covering the ears] for a female person. 2. (southern Germany, Austria) [wool] cap. 3. (History) Headgear of a Warrior; helmet.’12 That Haube was indeed used for male and female headgear in Tyrol in the fifteenth century is documented in inventories. Haube, in historical records sometimes also spelled Hawbe, as protective headwear was listed in 1446 as being in the armoury of Pergine Castle, Trentino-Alto Adige/South Tyrol (Zingerle 1909, 84)13 and in 1478 ein Truchl von Pain mit acht Hawbn (a chest made of bone with eight coifs/caps) was part of the dowry of Paola Gonzaga when she married Count Leonhard of Görz who resided in Lienz, East Tyrol. As can be seen, German-speaking people do not agree on the gender of the wearers much less on a distinct shape of the garment. To make matters worse, sometimes, the terms Haubttuch, Haupttuch, Hawbtuch appear in wills and inventories. And while haupt can be translated as ‘head’ and haupttuch as ‘headcloth’ and sounds promising, it is better not to jump to conclusions. Haupt can mean ‘main’ too and is used as such to refer to altar or table cloths, or even towels. The estate of Hans of Wehrburg (South Tyrol) lists in 1420 ‘ain tischlachn, ain genat hanttuch oder haupttuch, VI hanttucher, ain lang tischtuch’ (Zingerle 1909, 173) (one table sheet, one sewn hand towel or main cloth, six hand towels, one long table cloth). The dowry of Paola Gonzaga contained ‘zway Haubttucher mit gulden Leisten. Item zway Haubttucher mit weißn seidn Fransn. Item zway Haubttucher mit guldn Fransen. Item zway Haubttucher ausgenait a la morescha’ (Kollreider 1952, 146) (two main cloths with golden borders, two main cloths with white silk fringes, two main cloths with gold fringes, two main cloths sewn a la morescha).14 The fringes very strongly suggest that these were table cloths as caps or coifs did not have fringes. The only obvious conclusion gained from written records is that women must have worn a type of headwear named Haube (or hūbe, hūba, haube, hawbe, coiffes) and further research might reveal useful data.

Conclusion This unique find from Lengberg Castle provides important new insight into late medieval and early modern headwear and the development of textile techniques on which future research can be based. Although no other extant examples of this type of headwear with a sprang insert have been found to date, it can be concluded that it did actually exist due to the many examples in the visual record of the fifteenth and sixteenth centuries in Germany and Austria. Both men and women are depicted wearing panelled sprang or netted headwear that very closely resemble the extant Lengberg sprang garment. The size of the headwear from Lengberg indicates that it was worn by someone with a small head, most likely a woman, perhaps even a girl.

170

Beatrix Nutz, Rachel Case and Carol James

By Anna, the wife of Peter Mosheimer? Or maybe Andre Mosheimer’s wife of whom we have no record? Or someone else, maybe a servant? Although it is a functional garment, its delicate embellishments are subtly beautiful and graceful in their simplicity of material, yet deceptively complicated in the techniques used to create them. Maybe the headwear was worn as an everyday garment by someone of high standing in the castle (Nutz 2013b). To which of the female inhabitants of Lengberg Castle it belonged to though will forever remain a mystery. Notes

1 Ministerialis in the Holy Roman Empire were an upper class of originally unfree servants in the court, administrative and military service. They were entrusted by their landlord with a special function, such as the management of a court or of various possessions, such as castles. 2 An antependium (from Latin ante- and pendēre ‘to hang before’), also known as a parament or hanging, is a decorative piece, usually of textile, that can adorn a Christian altar. 3 For the pattern please contact Carol James at http://www.spranglady.com/contact-me.html. 4 For more information about sprang and its methods of construction, see Collingwood 1974. 5 The Crusader Bible, Old Testament Miniatures with Latin, Persian, and Judeo-Persian inscriptions France, Paris, 1240s. https://www.themorgan.org/collection/Crusader-Bible/thumbs. 6 Amica Sundstöm and Maria Neijman https://historicaltextiles.org/tidskapseln-the-timecapsule-la-capsula-del-tempo/ / [Accessed 04 December 2019]. 7 A loop-braided lace sewn onto the coif edge in an as yet undetermined technique. Personal communication with Amica Sundstöm and Maria Neijman, 7 March 2018. 8 Amica Sundström and Maria Neijman, Coif/ Cuffia. [online] historical textiles. Available at: https://historicaltextiles.org/coif-cuffia/ [Accessed 04 December 2019]. 9 E.g. in Hartmann Schedel, Georg Alt and Michael Wolgemut, Das buch der Cronicken vnd gedechtnus wirdigern geschichte[n], vo[n] anbegyn[n] d[er] werlt bis auf dise vnßere zeit Nürmberg, 1493 GW M40796 Folio NP. http://daten.digitale-sammlungen.de/%7Edb/0005/bsb00059084/images/ index.html?seite=232. Kesselmartyrium des Hl. Johannes Evangelist, altarpiece, 1485−1495, Master of the Fridolfing Altar, Johanneskirche, Fridolfing, Bavaria. http://tarvos.imareal.oeaw.ac.at/server/images/7001613.JPG. Hl. Margareta werden die Brüste abgeschnitten, altarpiece, 1490 −1500, Peterkirche, St Lambrecht, Styria, Austria. 10 Deutsches Wörterbuch von Jacob Grimm und Wilhelm Grimm, http://www.woerterbuchnetz. de/DWB?lemma=haube [Accessed 31.05.2018]. 11 Meyers Großes Konversationslexikon, http://www.woerterbuchnetz.de/Meyers?lemma=haube [Accessed 31.05.2018]. 12 https://www.duden.de/rechtschreibung/Haube [Accessed 31.05.2018]. 13 …siben hawben, ain ledigs geheng an ain'hauben, fünf hauben one geheng, zway englisch hauben mit geheng, zway helmlein, vier eysenhüte… (seven caps, one single aventail on a cap, five caps without aventails, two English caps with aventails, two helmets, four iron hats,…). 14 GROPPO MORESCHI. This term appears in a number of sixteenth-century pattern books, including Taglienti's Opera nuova of 1527. Although it may refer to Macramé, none of the patterns appear suitable for this technique. It is more likely to refer to designs incorporating Moresque arabesques to be worked in knotted embroidery stitches (Levey 1983, 121–22). 15 There were usually multiple prints of woodcuts, especially those used as book illustrations. The references to these images are exemplary and the manuscripts, or books, can be found digitised online on the websites of the respective libraries. Most Austrian images can be found in the database IMAREAL, https://realonline.imareal.sbg.ac.at/

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 171

Bibliography

Badische Landesbibliothek Karlsruhe, 1445/1470. Haus-und Arzneibuch – Donaueschingen 793. Unteres Inntal (Mühldorf). Available from: http://digital.blb-karlsruhe.de/urn/urn:nbn:de:bsz:31-37968 [Accessed 07 June 2021]. Benns, E. & Barrett, G. 2005. Tak v bowes Departed: A 15th Century Braiding Manual Examined (London, Soper Lane). Boutrup, J. 2010. Analysis and interpretation of braids from the London excavations. Archaeological Textiles Newsletter 50, 20–29. Boutrup, J. 2011. Die Schnüre am Gürtel Philipps von Schwaben. In M. Herget (ed.), Des Kaisers letzte Kleider: Neue Forschungen zu den organischen Funden aus den Herrschergräbern im Dom zu Speyer; [Begleitbuch zur Ausstellung ‘Des Kaisers letzte Kleider – Rettung der organischen Funde aus den Kaiserund Königsgräbern im Dom zu Speyer’ im Historischen Museum der Pfalz, Speyer, 10. April bis 30. Oktober 2011] (Munich, Ed. Minerva), 114–17. Case, R., McNealy, M. & Nutz, B. 2017. The Lengberg finds – remnants of a lost 15th century tailoring revolution. In M. Bravermanová, H. Březinová & J. Malcolm-Davies (eds), Archaeological Textiles - Links Between Past and Present: NESAT XIII (Liberec & Prague, Technical University of Liberec, Faculty of Textile Engineering/ Institute of Archaeology of the CAS, Prague), 167–76. Collingwood, P. 1974. The Techniques of Sprang: Plaiting on Stretched Threads (London, Faber and Faber). Crowfoot, E., Pritchard, F. & Staniland, K. 1992. Textiles and Clothing: c. 1150−c. 1450 (London, H.M.S.O.). Dahl, C.L. & Sturtewagen, I. 2008. The Cap of St. Birgitta. Medieval Clothing and Textiles 4, 99–142. Dillmont, T. de, 1890. Encyclopedia of Needlework (Newcastle-on-Tyne, Thomas Hunter). Drinkler, D. 2010. Eng anliegende Bekleidung in Antike und Renaissance. Zeitschrift für Kunsttechnologie und Konservierung 24 (1), 5–35. Emery, I. 1994. The Primary Structures of Fabrics: An Illustrated Classification (London, Thames and Hudson). Grömer, K., Kania, K. & Boutrup, J. 2015. Iron Age finger-loop braiding: finds from the Hallstatt Salt Mine. Archaeological Textiles Review 57, 39–46. Hörmann, M. 2003. Lengberg. In W. Beimrohr, M. Hörmann-Weingartner & O. Trapp (eds), Tiroler Burgenbuch, vol. 9, Pustertal (Bozen, Innsbruck, Athesia,Tyrolia-Verl), 545–56. Kielbusová, M. 2014. Krosienky: Pletení bez jehlic (Prague, Grada). Kindlová, Z. & Tomková, M. 2001. Pletení na krosienkách (Prague, Vzdělávací spolek uměleckých řemesel: Paličkování). Kinoshita, M. n.d. [online]. What is the Loop-Manipulation Braiding Technique? Available from: https://loopbraiding.wordpress.com/ [Accessed 07 June 2021]. Kollreider, M. 1952. Madonna Paola Gonzaga und ihr Brautschatz. In R. Klebelsberg (ed.), Lienzer Buch.: Beiträge zur Heimatkunde von Lienz und Umgebung (Innsbruck, Wagner), 137–48. Leitner, Q. von, 1888. Artistisches Quellenmaterial aus der gräfl. Thun-Hohenstein‘schen Fideicommiss-Bibliothek in Tetschen. Jahrbuch der kunsthistorischen Sammlungen des allerhöchsten Kaiserhauses 7, I–VI. Levey, S.M. 1983. Lace: A History (London, Victoria and Albert Museum). Mihalić, J. 2013. Novi život drevnog umijeća: Obnova pletenja na lučkom okviru u selu Trg kraj Ozlja (Karlovac, Gradski muzej). Netto, J.F. & Lehmann, F.L. 1800. Die Kunst Zu Stricken in ihrem ganzen Umfange. Oder: Vollständige und gründliche Anweisung alle sowohl gewöhnliche als künstliche Arten von Strickerei nach Zeichnungen zu verfertigen.: Mit dreissig illuminirten und schwarzen Kupfertafeln (Leipzig, Voss). Nutz, B. 2013a. Bras in the 15th century? A preliminary report. In C. Nübold & J. Banck-Burgess (eds), NESAT XI: The North European Symposium for Archaeological Textiles XI: 10–13 May 2011 in Esslingen am Neckar (Rahden, Westf, VML Vlg Marie Leidorf), 221–25.

172

Beatrix Nutz, Rachel Case and Carol James

Nutz, B. 2013b. Dressed to the nines. Kleidung zur Identifikation des sozialen Standes. In N. Hofer, T. Kühtreiber & C. Theune (eds), Mittelalterarchäologie in Österreich  – Eine Bilanz: Beiträge der Tagung in Innsbruck und Hall in Tirol, 2. bis 6. Oktober 2012 (Vienna, Österreichische Gesellschaft für Mittelalterarchäologie), 294–302. Nutz, B. 2014. DRGENS SN WIR VS NVT SCHAME – No shame in braiding. 15th century fingerloop braids from Lengberg Castle. Estonian Journal of Archaeology 18 (2), 116–34. Nutz, B. 2015. Hemdchen, Kleidchen, Schühchen: Säuglings- und Kinderbekleidung des 15. Jahrhunderts aus Schloss Lengberg, Osttirol. In A. Paetz-Schieck & U.-C. Bergemann (eds), Das Bild vom Kind im Spiegel seiner Kleidung: Von prähistorischer Zeit bis zur Gegenwart (Regensburg, Verlag Schnell & Steiner), 94–109. Nutz, B. 2018. Linen sprang from Lengberg Castle. In H. Hopkins & K. Kania (eds), Ancient Textiles, Modern Science II (Oxford, Oxbow Books), 60–72. Nutz, B. 2019. Nets – knots – lace: early 16th century headdresses from East Tyrol. Archaeological Textiles Review 60, 24–49. Paulis, L. 1930. Le Point de Dentelle. Bulletin des Musées Royaux d’Art et d’Histoire 2 (6), 165–66. Paulis, L. 1933. Le Point de Dentelle: A propos d’une gravure d’après Martin Devos. Bulletin des Musées Royaux d’Art et d’Histoire 5 (1), 19–21. Plattner, P. 2012. Schriftfunde aus Schloss Lengberg in Osttirol (Diplomarbeit, Universität Innsbruck). Rutlin, S., Brun, G. & Bringsvær, G. 1964. Den store håndarbeidsboken (Oslo, Dovre Mønster). Santonino, P. 1947. Die Reisetagebücher des Paolo Santonino, 1485-1487. Aus dem Latein. übertr. von Rudolf Egger. 13th ed. (Klagenfurt, Kleinmayr). Schlabow, K. 1960. Abdrücke von Textilien an Tongefassen der Jungsteinzeit. Jahresschrift für mitteldeutsche Vorgeschichte 44, 51–56. Schmedding, B. 1978. Mittelalterliche Textilien in Kirchen und Klöstern der Schweiz: Katalog (Bern, Stämpfli). Schneider, J. 1975. Textilien: Katalog der Sammlung des Schweizerischen Landesmuseums Zürich; ausgewählte Stücke (Zürich, Verlag Berichthaus). Schweitzer, W.F. 1966. Textile making methods depicted in frescos. CIBA Review 1966 (1), 10–17. Speiser, N. 2000. Old English Pattern Books for Loop Braiding: A Monograph Critically Comparing English Instructions from the 15th and the 17th Century (Arboldswil, Selbstverl. des Autors). Speiser, N. & Boutrup, J. 2012. European Loop Braiding Part IV. Track Plans as a Tool for Analysis Applications of Loop-Manipulation Braids (Leicester, Jennie Parry). Tagliente, G.A. 1527. Opera nuova che insegna alle donne a cusire, a racammare & a disegnar a ciascuno, et la ditta opera sara di molta utilita ad ogni artista, per esser il disegno ad ognuno necessario, la qual e intitolata esempio di recammi (Venice). Terjanian, P. 2013. The Art of the Armorer in Late Medieval and Renaissance Augsburg: The Rediscovery of the Thun Sketchbooks (Part 1). Jahrbuch des kunsthistorischen Museums Wien 13/14, 298–395. Wunderlich, W. 1996. Weibsbilder al fresco: Kulturgeschichtlicher Hintergrund und literarische Tradition der Wandbilder im Konstanzer Haus ‘Zur Kunkel’ (Constance, Stadler). Zander-Seidel, J. 1990. Textiler Hausrat: Kleidung und Haustextilien in Nürnberg von 1500–1650 (Munich, Dt. Kunstverlag). Zander-Seidel, J. 2010. ‘Haubendämmerung’. Frauenkopfbedeckungen zwischen Spätmittelalter und Früher Neuzeit. In R. Schorta & R.C. Schwinges (eds), Fashion and Clothing in Late Medieval Europe – Mode und Kleidung im Europa des späten Mittelalters (Basel, Schwabe), 37–43. Zingerle, O. von, 1909. Mittelalterliche Inventare aus Tirol und Vorarlberg. Mit Sacherklärungen (Innsbruck, Verlag der Wagner’schen Universitätsbuchhandlung).

9.  A unique survival: A woman’s fifteenth-century headdress from Lengberg Castle 173

Appendix. Images of headwear with sprang or net inserts in the fifteenth and early sixteenth centuries15 Date

Person represented

1435

Nurse (?)

‘Death of Joseph’, Biblia pauperum, fol. 18r, Pierpoint Morgan Library, MS M.230

Fig. 9.13. A

Andromache

‘Hector and Andromache in bed’, Guido de Columnis, Historia destructionis Troiae, fol. 151r, Austrian National Library cod. 2773

Fig. 9.13. B

1445−1450

Source

Fig.

1460/80

Midwife

Master of Uttenheim, Birth of the Virgin, altarpiece, German National Museum no. Gm1180

Fig. 9.13. C

1469−1480

Midwife

Master of the Schottenaltar, Birth of the Virgin, altarpiece, Schottenstift, Vienna

Fig. 9.13. F

1470−1475

Executioner

‘The burning of Jerome of Prague’, Ulrich of Richental, Council of Constance, fol. 84r, Austrian National Library cod. 3044

Fig. 9.13. G

Midwife

Master of the Crucifixion Triptych, Birth of the Virgin, altarpiece, private collection

Fig. 9.13. D

1470s−early 1490s c.1480 (c.1532)

1490 1490s−c.1509

Man with armour and sword (knight) Cook Wrestlers

Thun-Hohenstein Album (Thun’sche Skizzenbuch), Augsberg, Umĕleckoprů myslové Museum v Praze, GK 11572-B, fol. n38r Kuchenmeisterey, woodcut, Wellcome Collection

Fig. 9.13. H

Fig. 9.13. K & L Hans Wurm, Das ist ain hybsch ring byechlin (Landshut Wrestling Manual), Johann Sittig, Augsburg, fol. 1v, 2r, 6r, 7r, 8r, 11r, Bavarian State Library BSB-Ink R-190- GW 5

1491

Servant serving wine

Stephan Fridolin, Schatzbehalter, fol. 51v, woodcut, University Library Heidelberg, GW 10329

1491

Executioners

Stephan Fridolin, Schatzbehalter, fol. 34r, 47v, 52r, woodcuts, University Library Heidelberg, GW 10329

1491

Soldier (?)

Stephan Fridolin, Schatzbehalter, fol. 125v, woodcut, University Library Heidelberg, GW 10329

1493

Executioners

Fig. 9.13. I

Hartmann Schedel, Das buch der Cronicken vnd gedechtnus wirdigem geschichte[n], fol. 104r, 104v, woodcuts, Nuremberg, Bavarian State Library BSB-Ink S-197 – GW M40796 (Continued)

174

Beatrix Nutz, Rachel Case and Carol James

Date

Person represented

1493

Servant with feather fan

1495

Midwife

1496/1497

Executioner

1496/1498

Torturer

1497

1497

1500

Source

Master of Mariapfarr, Birth of the Virgin, altarpiece, parish church Salzburg

Fig. 9.13. E

Albrecht Dürer, The Martyrdom of St Catherine, woodcut, Metropolitan Museum of Art, New York, no. 19.73.154 Albrecht Dürer, The Martyrdom of St John, woodcut, National Gallery of Art, Washington DC, no. 1941.3.5

Doctors or Hieronymous Brunschwig, Das ist das Buch medical students der Cirugia, Hantwirckung der Wund Artzney, woodcut, Augsburg, Bavarian State Library BSB-Ink B-931 – GW 5594 Patient

Fig.

Hartmann Schedel, Das buch der Cronicken vnd gedechtnus wirdigem geschichte[n], fol. 94r, woodcut, Nuremberg, Bavarian State Library BSB-Ink S-197 – GW M40796

Hieronymous Brunschwig, Das ist das Buch der Cirugia, Hantwirckung der Wund Artzney, woodcut, Augsburg, Bavarian State Library BSB-Ink B-931 – GW 5594

Moresque dancer Sculpted relief, Golden Roof, Innsbruck

1510

Wrestlers

Hye in disem buchlin findt man die recht kunst vnd art des Ringens, fol. 1v, 1r, 2r, 3r, 4r, 5r, 10v, 11r, Bavarian State Library BSB-Ink R-191

1512

Wrestlers and fencers

Albrecht Dürer, Oplosisaskalia sive Armorvum Trachtandorvm Meditatio Alberti Dvreri [Weapon Training, or Albrecht Dürer’s Meditation on the Handling of Weapons], Albertina, Vienna, MS 26-232, fol. 41r, 42r, 48r, 49r, 66v, 77r, 79r, 85r, 87r

c.1512

Fencer

Fencing book, Nuremberg (?), State Library Berlin Libr. pict. A 83, fol. 58r

1516

Cook

Küchenmeisterei, woodcut, Johann Knobloch (publisher), Strasbourg, Bavarian State Library Res/4 Oecon. 209 y

Fig. 9.13. J

Chapter 10 From narrow four-strand plaits to openwork bobbin-made braids and edgings Lena Dahrén

Introduction The starting point of this study is a group of extant bobbin-made braids and edgings produced in the late sixteenth and early seventeenth century. Some were reused on church vestments produced in the early seventeenth century, others remain in their original position on fashionable garments. I discovered these artefacts while doing research for my doctoral thesis in Textile History, Uppsala University (Dahren 2010). By undertaking technical analyses of these extant early modern braids and edgings, I have gained knowledge and an understanding of the bobbin-technique used by the craftsmen and women of the time. The intention is not to reproduce the patterns in order to make braids in the original style but to try out the technique to understand how the material, tools and technique interact to give a specific result. All bobbin-made braids and edgings mentioned in this article are made from metal thread, that is a narrow, thin metal strip wound S-wise around a core of fibre, normally silk. Usually the metal is silver, silver-gilt or an alloy with copper (Fig. 10.1). To date no material analysis has been carried out on the gold braids, however, similar gold threads used for woven fabric of comparable date have been analysed and shown to be silver-gilt. I have occasionally seen that the gold-coloured thread is green from verdigris, which suggests that the metal contains copper. For simplicity here I will refer to these as gold threads.

Terminology Interpreting the terminology of early bobbin-made work is problematic and has been discussed before by, for example, dress and textile historians Janet Arnold and Santina Levey in relation to early modern English inventory accounts (Levey 1983, 15; Arnold 1988, 219–27, 351, 366, 368). The situation they describe

176

Lena Dahrén

Fig. 10.1. The metal thread is a flat strip of metal, wound around a core of silk

is similar in Sweden. In the Swedish accounts the term posament is well known from the 1540s and according to the accounts of the Swedish royal wardrobe large amounts of gold posament were used for royal dress in the 1560s especially for the coronation of Erik XIV. However, the tool for production of some of these posaments was mentioned as a ‘loom’ (vävstol) or as ‘instruments’, which were made by a woodturner (Räntekammarböcker, vol. 39:2, fol. 127r). This might imply that it was a new technique and as such did not have a vocabulary of its own. The Swedish ambassadors, who visited London in the mid-sixteenth century, brought large amounts of posament to Sweden and something which in the accounts is called ‘genomsichtlige posament’ (Klädkammaren IIA, 26:9, fol. 24r). Directly translated into English it would be ‘passement which can be seen through’, or perhaps more correctly ‘openwork passement’. The Swedish word spets (German = spitze, English = lace, French = dentelle), which is used today, is not found in the Swedish accounts until the early seventeenth century. When used earlier it is in the meaning of an iron point for a sword or lance. What today is called bobbin-made lace is referred to in the early modern accounts with the term bård, which I understand is equivalent to the English word border or braid, and udd, which is edging or pointed edging. There is also the word knytning, which in English would approximate to knotting. The term knytning is still used for the traditional bobbin-made edgings associated with traditional costume in the region of Dalecarlia in central Sweden. The Swedish textile scholar, Agnes Branting, stated in 1907 that the designation for the bobbin technique in different countries followed the term used for the tools that

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

177

the threads are wound onto (Branting 1907, 110). This statement is supported by the English term bobbin lace, the French dentelle aux fuseaux, the German Klöppeln and the Swedish term knyppling. In all these cases the word for the tool – a bobbin – is involved. Following Agnes Branting’s suggestion, and inspired by Santina Levey, who use the term ‘braid-like lace’ for the patterns in Nüw Modelbuch allerley Gattungen Däntelschnür and Le Pompe (Levey & Payne 1983, 12), I have chosen to use the term bobbin-made braids and edgings or bobbin-made work when discussing the work made from fourstranded plaits. To promote a clearer international understanding there is a need for further archive study into the terminology of early modern bobbin-made work in different European countries and a discussion concerning the specific terms used.

The technique It is not possible to tell from its appearance whether a four-strand plait has been made by manipulating the threads with the fingers or by working with the threads wound onto bobbins. However, it is easier to make a long plait if the threads are wound onto bobbins than to handle long threads that are likely to become entangled while working them. It would also have been discovered that plaiting was easier to manage if the plaits were anchored to a firm surface such as a hard-stuffed pillow, to stabilise the working. The very early narrow bobbin-made braids would therefore seem to mark the technical transition from braid making, simple plaiting and other kinds of bobbin-made work which all were part of the passementier’s production (Diderot & d’Alembert 1967, 126−38; Levey 1983, 5; Shepherd 2009, 1). There was probably no separation between the different techniques used when it came to the production of a decorative border as the only aim for choosing a specific technique was achieving the desired result. This hypothesis is supported later in this study by two unusual borders from Hedvig and Hjärtlanda churches where a combination of techniques can be seen. It should therefore be remembered that this discussion of terminology was not a problem for the producers. However, when studying textiles from written words in the contemporary archives and describing extant braids, it is useful to have relevant definitions of the different types for clarity.

Early pattern books for bobbin-made work and recent reconstructions There are three pattern books currently known for early bobbin-made work – Nüw Modelbuch allerley Gattungen Däntelschnür, Le Pompe and Le Pompe. Libro Secondo, all of which were published as facsimiles in the 1980s. The cover of Nüw Modelbuch shows two women sitting by windows working threads wound onto bobbins on sloping pillows (Fig. 10.2). There is no pricking under the bobbin-made border for the women to follow, which suggests that they are working ‘freehand’ – which means that you work without a pricking as guidance for the work and with very few pins to support the work, normally only at the sides. This

178

Lena Dahrén

is probably the way bobbin-made braids were worked at that time. All the patterns in Nüw Modelbuch are geometrical and woodcut in white on a black background (Fig. 10.3). Above each pattern the name of the design and the number of threads needed is given in Roman numerals. There are no instructions on how to make the designs (Nüw Modelbuch 1561/Burkhard 1986).

Fig. 10.2. Nüw Modelbuch printed in Zurich in 1561, Inv. No. 10519. (© Billedsamlingen, Designmuseum Denmark)

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

179

Fig. 10.3. Nüw Modelbuch with geometrical designs woodcut in white on a black background

Le Pompe and Le Pompe. Libro Secondo, are two different pattern books with both geometrical and flowing designs woodcut in black on a white background, published by the Sessa brothers in Venice in several editions (Fig. 10.4). The 1559 edition of Le Pompe, was made as a facsimile together with parts of Le Pompe. Libro Secondo printed in 1560 (Levey & Payne 1983). The facsimile publications have additional parts with reconstructions of the designs including prickings and working-diagrams as guidance on how the bobbinmade braids and edgings can be made. The reconstructions are made from the woodcuts of the designs in the pattern books as in the 1980s there were no known extant braids of the designs to study. As everyone can understand, it is very hard to make technical reconstructions from a woodcut without any extant items to study. However, as Santina Levey writes in her comprehensive introduction to the facsimile of Le Pompe, very few bobbin-made pieces have survived from this time mostly because the valuable gold and silver threads have been burnt out and melted down for recycling into other precious metallic items, including coins (Levey & Payne 1983, 12). Interest in early modern bobbin-made work has increased over the last decades. There is an international research group of lace-makers, founded within the international organisation for bobbin and needle lace (OIDFA), who specialise in the study of the early so-called ‘freehand-lace’. They are working their interpretations of the pattern book designs as well as analysing and reconstructing surviving bobbinmade work in museum collections.

180

Lena Dahrén

Fig. 10.4. The designs in Le Pompe are woodcut in black on a white background

The facsimiles of the early modern pattern books have been a source of information and endless discussions for this group. Lace-makers appreciate that there are small but important technical differences that they have to be aware of while reconstructing early modern work because the freehand way of working is different from what most of them are now used to, but the techniques are important to understand because they affect the appearance of the finished work. A lot has been learnt during the almost 40 years that have passed since the two facsimiles with their interpretations of early bobbin-made work were published. Many wise words have been said by, for instance, Santina Levey about the difficulty of reconstructing the techniques as the woodcut patterns only give the lace-maker an idea of how to make the design (Levey 1983, 10). Obviously, this requires not only surviving items to be available for analysis, but also a lace-maker with the relevant knowledge and practical experience who is able to distinguish the different techniques and understand how the material and stitches chosen for the braid contribute to the finished work. In Sweden, there is a relatively large amount of extant bobbin-made braids and edgings, produced in the period 1550−1640, which were donated to small parish churches in the seventeenth century and reused for church vestments. This makes it possible to analyse the original braids and edgings to reproduce the technique as it was used in the early modern period and to compare with the designs in the early pattern books. Approximately 100 extant braids and pointed edgings of silver and gold thread in different designs from the period 1550−1640 were analysed and documented as part of my doctoral thesis (Dahrén 2010). I have made 10 to 15 samples of them and drawn

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

181

technical details from many more. The samples are made from my personal experience of making freehand bobbin lace. This knowledge was gained during the many years I worked as bobbin-lace consultant within the Swedish Handicraft Association and where I had the privilege of collaborating with bobbin lace-makers to whom the tradition of making freehand bobbin lace had been handed down through generations. My special interest is traditional peasant bobbin lace which is made from handspun linen thread and used for traditional interior textiles or clothing. That is freehand lace as it has been produced for instance in the region of Dalecarlia, which is situated in the centre of Sweden. The laces from that area are made using a similar technique and designs to some of the preserved bobbin-made braids of gold and silver from the sixteenth and early seventeenth centuries and some of the designs can even be found in the sixteenth-century pattern books. Plaited techniques can be hard to distinguish from each other. Mistakes can easily be made, including the assumption that the technique is bobbin made. In Uppsala cathedral there is an openwork braided insertion made from coloured silk and gold thread on a sudary (a linen cloth to prevent direct contact between the bishop’s hand and the crosier) dated 1422, which previously was assumed to be bobbin made (Fig. 10.5) (Estham 2010, 261–62). If so, it would have been one of the oldest existing examples of the technique in the world and it is mentioned in several general surveys of bobbin and needle lace. In the late 1990s the textile scholars, Noémi Speiser and Milton Sonday, were invited to study the insertion and were able, through technical

Fig. 10.5. Loop-braided insertion of coloured silk and gold, Uppsala cathedral. (© Gabriel Hildebrand)

182

Lena Dahrén

analysis, to determine that it was not bobbin-made but loop-braided (Speiser & Boutrup 2011, 33–39). It was recognised as a design known as ‘Katherine wheel’, which is an openwork loop braid. From the technical drawings and detailed pictures which Speiser published in the report, it is obvious that the insertion had never been studied by anyone who had mastered the technique, as the plaits on each side of the centre plait are made of three – not four – strands, the latter being the normal number of threads for plaits in bobbin-made work (Fig. 10.6) (Speiser & Boutrup 2011, 48, 54–55). Braids of geometric diamond-shaped mesh are seen quite often in early modern paintings. The earliest known plaited work with picots is depicted in a painting dated 1460 made by the renowned artist of the Gonzaga court in Mantua, Andrea Mantegna (Fig. 10.7. A). In the right-hand corner, the painting shows the high-priest Simeon in a yellow mantle trimmed with gold diamond mesh braids made of plaits decorated with pairs of picots on opposite sides of the plaits (Fig. 10.7. B). A similar type of braid to this, but less dense, appears on an altar-screen, made in North Germany 1500−1525 in honour of Saint Ninianus (National Museum, Copenhagen, inv. no. CDXXXVI). The Saint is depicted as an archbishop wearing costly vestments of cloth of gold with golden braids on both the mitre and cope. An extant bobbin-made border of silver and gold four-strand plaits of the same type as that in Mantegna’s painting is preserved on a burial dress worn by the six-year-old Gräfin (Countess) Katharina zur Lippe (1594−1600), who was buried in the Augustiner Kirche, Blomberg, Germany in 1600 (Fig. 10.8). The dress is now in the collections of the Detmold Landesmuseum, Germany (Arnold 1985, 42; Bringemeirer 1985, 215).

Fig. 10.6. Diagram of loop-braided insertion in Uppsala cathedral, after Noémi Speiser

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

183

a

b

Fig. 10.7. A. Andrea Mantegna, Presentation, c. 1460, oil on canvas, Staatliche Museum Berlin. B. Detail of right-hand corner. (© pbk Photo Agency)

184

Lena Dahrén

Fig. 10.8. Braid of interlacing silver and gold plaits. Dress of Gräfin Katharina zur Lippe

The technique described from extant bobbin-made work Bobbin-made work as discussed in this article results from two movements: ‘twist’ and ‘cross’. The threads are worked two and two, that is in pairs, with one pair in each hand. Twist is performed within each pair by moving the right bobbin over the left (Fig. 10.9. A). Cross is made between the pairs by moving the right bobbin in the left-hand pair over the left bobbin in the right-hand pair (Fig. 10.9. B). By repeating twist and cross twice the threads within each pair of threads go straight through each other (Fig. 10.9. C). Repeating the twist and cross three times results in a short plait (Fig. 10.9. D). This last stitch is often used in the early modern bobbin-made freehand lace and gives stability to the work, as the threads from each pair are divided so the stitch is secured and holds the threads in place. Designs from a straight plait A minimum of four threads is normally needed to do the kind of bobbin-made work under discussion. This means that a bobbin-made braid is usually made with an even number of threads. There are exceptions to this rule. The most common exception is if there is a contour thread (gimp) in the work. However, there are some surviving

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings a

b

c

185

d

Fig. 10.9. A. Twist. B. Twist, cross. C. Twist, cross, twist, cross. The threads within each pair go straight through each other. D. Twist, cross, twist, cross, twist, cross. The threads within the pair are divided and secure the stitch

bobbin-made edgings with points made of plaits with three threads (Dahrén 2010, 96−97). By combining the two movements, twist and cross, in different ways, endless variations can be obtained and even small changes in technique and additions of threads can lead to changes of design, character and function. A plait made from four single threads is rather insignificant. Made with double threads the plait becomes heavier and has more visual impact as the paired metal threads reflect more light than a plait made from single threads. Such a double thread four-strand plait trims the edge of a chasuble from Södra Vånga church (Fig. 10.10. A & B). Slight changes to the design and character of the four-strand plait can be made by adding picots, contour or heavier threads, or spangles. For example, by adding a flat metal strip a simple, narrow plait immediately looks more elaborate as the flat metal strip reflects more light than the wrapped metal thread (Fig. 10.11. A & B). By making pairs of picots on opposite sides of such a plait, small holes appear in the centre of the plait and the design changes from a straight, closed plait into a narrow plait with an openwork pattern (Fig. 10.12. A & B). The size of the picot is determined by the diameter of the pin it is passed around – and it seems that the makers used rather thick pins because the picots tend to be noticeable. In early modern bobbin-made work dating to the sixteenth century, picots normally consist of just one thread passed around the supportive pin, as can be seen in diagram Fig. 10.12. B. The practice of making a picot by passing two threads around a pin seems to appear at the beginning of the seventeenth century and is used in the production of pointed edgings from about the 1620s. However, it was also used in straight braids, for example a braid on

186

Lena Dahrén

Fig. 10.10. A & B. Plait and diagram of plait of parallel threads. Södra Vånga church

Fig. 10.11. A. A four-strand plait (right), width 2 mm, with a flat metal strip passing through. Royal Armoury, Stockholm, (3376)31142. B. Diagram

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

a

187

b

Fig. 10.12. A. Detail of plait from the Douce bible, Bodleian Library. (© The Bodleian Library). B. Diagram of the plait. Only one thread is passed round the pin, to form the picot

a chasuble at Mariestads cathedral (Dahren 2010, 203). A reason for this change may be that the earlier bobbin-made braids appear to have used a heavier thread that is stiffer and reacts in a different way to the thinner material used for the seventeenthcentury pointed edgings. The difference between a straight braid and a pointed edging can result from a small technical change, for example making picots on just one side of the plait. A pattern for such an edging is to be found in Nüw Modelbuch. It is called ‘Spitzle mit IIII’. The Roman numeral ‘IIII’ means that the design should be made with four threads (Fig. 10.13. A). Such a narrow silver edging is used to outline an embroidered crucifix attached to the back of a chasuble in Algutstorp church and gives a suggestion of how the design in the pattern book can be made (Fig. 10.13. B). The narrow edging with picots on one side can be varied by placing the picots further apart (Fig. 10.14. A & B). Zigzag patterned edging The designs discussed so far have been composed of variations of a single straight plait. The straight plait can become a zigzag edging by making a short plait of three crossings length and then adding a picot on the right-hand side, then creating another short plait and adding another picot on the left-hand side. This kind of zigzag edging was, as with the earlier simple plaits, used in the sixteenth and early seventeenth centuries as decoration attached to the surface of, for instance, dark velvet. It was also used to divide the surface and to frame embroidered work (Fig. 10.15. A & B).

188

Lena Dahrén

a

b

Fig. 10.13. A. Design from Nüw Modelbuch of a plait with picots on one side only. B. Bobbin-made plait with picots on one side, Algutstorp church

a

b

Fig. 10.14. A & B. Diagrams of plaits made by four threads with picots on one side

If two such zigzag edgings are mirrored and joined with a short plait, an openwork pattern results (Fig. 10.16. A & B). This braid is known with small variations almost everywhere bobbin-made work has been made and it is still made today; in Sweden it is called ‘Udd och stad’ or ‘K och O’ and in Denmark it is known as ‘Eternelle’. The same kind of openwork braid becomes almost unrecognizable in its character by adding two coarser metal threads as contour threads forming an endless loop or twist around each other (Fig. 10.17. A & B).

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

189

b

a

Fig. 10.15. A. A picot on each side of a plait gives a zigzag edging. B. This edging frames an embroidery. (© Staatliche Museum Berlin, Kunstgewerbemuseum and Christine Waidenschlager)

a

b

Fig. 10.16. A. Two zigzag edgings joined to a braid, Aspö church. B. Diagram

A short plait with the length of three crossings is often used in the early modern bobbin-made braids and edgings to join separate elements. It is a stitch which gives stability to the work, facilitating the making of the plaited pattern, and preventing the grid from becoming distorted and thus maintaining the right-angles of the design. It is assumed that the early bobbin-made braids were made freehand without the aid of an underlying pattern and with pins as support only at the sides. The third twist and cross in the stitch means that the two threads from the elements that are to be joined do not pass straight through each other but lock the threads giving stability to the work. This stitch is still used in the region Dalecarlia, central Sweden, where freehand lace is traditionally made. The stability that this stitch gives to the work is so vital to the so-called ‘freehand’ work that it helps to confirm whether a piece was made freehand,

190

Lena Dahrén a

b

Fig. 10.17. A. Bobbin-made braid with contour threads. Bredaryd church. B. Diagram. The short plait in the centre secures the join and gives stability to the braid

in other words without the support of an underlying pattern. The disappearance of this stitch is an indication of when supportive patterns started to be used for bobbinmade work (Dahrén 2010, 70−71). Plaits adorned with picots By passing one thread from the plait around a pin, a picot appears on the side of the plait. When there are picots on both sides of a plait they are usually made by letting the same thread or pair of threads go from one side to the other and making picots by passing the outermost thread around a pin on both sides. This gives a small asymmetry which is hardly noticeable (Fig. 10.18. A–C). If several picots are placed on both sides of a plait directly after each other they form what the eye perceives as a small woven square made from four warp threads and one single weft thread (Fig. 10.19. A & B). This is an illusion as the square is made of a four-strand plait with picots on both sides. It works as follows: if the two left-hand threads of the plait function as vertical warp threads, the two right-hand threads go straight through the two vertical threads as weavers. When preparing the turning back, the two horizontal threads are twisted once and the outermost thread is passed around a pin and forms a picot, thus just one thread forms the picot. Then another twist is made with the same two threads and these then work back through the vertical two threads again. The turn is repeated on the other side by making another picot; that is the two threads are twisted once, then the outermost thread is passed around a pin and the picot is closed with another twist of the two threads (Fig. 10.19. C).

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

a

b

191

c

Fig. 10.18. A. Four-strand plait with picots on both sides made by the same thread. B. Plait where a pair of threads goes through the plait and the outer thread form the picots. C. Detail of a gold fourstrand plait with picots similar to Fig. 10.18 B

a

b

c

Fig. 10.19. A. Several picots on each side in a row give an impression of a small square. Hedvig church. B. Braid made of plaits with squares in a diagonal grid, Ytterjärna church. C. Diagram of how this type of ‘square’ is made

192

Lena Dahrén

a b

Fig. 10.20. A. Plaits joined by letting the threads pass straight through. B. Diagram

Diamond-shaped grids formed of two plaits passing straight through each other As already shown in the Mantegna painting, geometrical diamond-shaped patterns can be formed by joining plaits which meet on a diagonal angle and pass straight through each other with double threads (Fig. 10.20. A & B). In contemporary bobbin lace terminology this crossing is called a windmill. I prefer to call it cross, twist, cross or a linen-stitch, as it works like tabby weave. In braids with a diamond-shaped grid this is a common way to join the plaits and to avoid a mixture of different coloured material, such as gold and silver. By using different colours of metal thread, that is plaits made from gold and silver, a secondary pattern results within the grid as the gold plaits pass through the silver plaits and vice versa (as described earlier). An example of this can be seen in the gold and silver border on the burial costume of Gräfin Katharina zur Lippe (Fig. 10.8). The same type of colour effect can be noticed in the border of silver and gold thread, attached to a chasuble, from Ytterjärna church (Fig. 10.19. B). Joining three braids The patterns just described are made of regular diamond-shaped grids. A braid from Hedvig church, Norrköping, has a hexagonal pattern, which encloses a decorative feature (Fig. 10.21. A−C). The braid consists of six plaits of which two plaits are made of gold thread and four are of silver thread. The four silver plaits make a hexagon, which encloses a decorative design formed by the two gold plaits in the centre, and vice versa. This is achieved by three plaits joining and passing through each other on each side of the braid. The join is made by allowing two threads to go up and two down in a regular way as in tabby, which gives decorative interlaced triangles of double threads that enrich the braid. A similar type of braid from Fogdö church also includes a hexagonal pattern (Fig. 10.22). Here four gold plaits produce the hexagon, which encloses a decorative element made of two silver plaits that meet in the centre. In the Fogdö braid, more care is put into the regular spacing of the outer points on each side of the braid than

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

193

a

b

c

Fig. 10.21. A. Braid of gold and silver thread, Hedvig church. B. Detail of three plaits joining and passing through each other. C. Diagram

Fig. 10.22. Braid of gold and silver thread, Fogdö church

194

Lena Dahrén

Fig. 10.23. Braid of gold and silver thread, Hjortsberga church. (© Gabriel Hildebrand)

in the Hedvig braid which means that there is less effort put into the regularity of the triangular form made by the joining plaits. In the Fogdö braid, threads simply pass through each other into their new working position in order to avoid any mixing of colours. Consequently, the join in Fogdö is not as regular and decorative as the triangles are in the braid from Hedvig church. Another braid, this time from Hjortsberga church, also has a hexagonal pattern of golden thread, which encloses a diamond-shaped meeting of two plaits in the middle (Fig. 10.23). The pattern is repeated but with silver threads enclosing the gold join. Both sides of the braid are adorned with picots made from short plaits, a picot and then again, a short plait. The gold and silver threads pass straight through each other so the materials do not mix. So far it is the same method as in the braids from Fogdö and Hedvig churches. The difference from the earlier two braids is that this braid combines different techniques. There are the full plaits forming the central pattern. However, the outer sides of the hexagon and the points on both sides are made by splitting the plaits into pairs of threads. All joins are made by short plaits. Braids in three parallel bands Another variation of this hexagonal pattern is seen in a braid from Aspö church, Sweden (Fig. 10.24). The central section is again made of silver and gold plaits, which alternately form the pattern. This is however another type of braid, or group of braids, which I have defined as braids made in three parallel bands. The central part is a

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

195

Fig. 10.24. Braid of three parallel bands of gold and silver thread, Aspö church

hexagon made of plaits adorned with picots along the outer side. The same motif is repeated and interlaced first in gold threads and then in silver. On both sides of the central section there is a narrow braid with two twisted contour threads of the type illustrated in Fig. 10.17. These three parallel braids are made at the same time even though the threads in the three parts are not always interlaced. This can be seen in the structure of the join, where the threads are laid around each other and returning to their original pathway. It would not be possible to make such a pleasing feature of the join if metal threads were to be drawn through each other afterwards. Comparing the four braids (Figs 10.21−10.24) with the hexagonal motif it may be argued that the different ways to solve the joining of the threads and other technical details (such as the edgings) can be related to different workshops or to different production dates of the braids. However, the basis for such a reasonable assumption is, so far, too small. As a hypothesis I would argue that the Fogdö braid is older than the Hedvig braid on account of the pointed edgings, where the three picots along the edging in Hedvig church seem to be a later feature of a more ornamental kind, which was common in early seventeenth-century bobbin-made pointed edgings. Braids made from three parallel parts with a central pattern of plaits formed into loops Approximately ten extant braids have been identified composed of three parallel bands and they are all preserved in secondary use attached to liturgical vestments

196

Lena Dahrén

such as chasubles or antependia (Dahrén 2010, 198−216). However, these kinds of braids can be seen in portraits of royalty and high nobility depicted from the late sixteenth century, which suggests that they were a new introduction during the second half of the sixteenth century. Among these extant braids there are two that differ technically from the others in a significant way: their central part consists of loops formed by plaits laid and crossed in a remarkable technique that is unknown in modern bobbin-lace making. This type of design with curved loops is represented in the pattern 3A in the facsimile of Le Pompe and Le Pompe. Libro Secondo (Fig. 10.25. A & B) and there is a reconstruction made of the pattern with diagram and pricking in the facsimile (Levey & Payne 1983, 92−93, 123). However, the techniques used to reconstruct the woodcut design rely on more modern techniques, for example sewing, that have not been observed in the extant early modern braids I have studied. Examining the extant braid preserved in secondary use on a chasuble in Hedvig church thus gave an opportunity to analyse the technique and to present an alternative possible solution (Fig. 10.26. A & B). The preserved early modern plaits are made of a relatively heavy metal thread. To make a plait of the character depicted in the pattern book demands thread of the assumed thickness shown in the woodcut and forming the plait into a durable loop a

b

Fig. 10.25. A. Design from Le Pompe. Libro Secondo with loop-shaped forms. B. Detail of the loop

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

197

a

b

Fig. 10.26. A. Braid of gold and silver with central design of plaited loops, Hedvig church. B. Detail of the loop braid

requires material that is stiff enough to keep its curved form unsupported. Support provided by pins during production (as suggested in Levey & Payne 198373, 92) is of no use if the material is too weak and the loop will not maintain its shape once the pins are removed. The thickness of the thread and tension of the work go together. The stiffness of a heavy metal or thick linen thread would give the form and character of the loop in a manner which a thin linen or cotton thread would not. A technical solution to the puzzling loop is provided by the extant bobbin-made braid from Hedvig church. After analysing this extant braid my suggestion for the production of such a loop is to make the plait with a relatively heavy metal (or linen) thread to the required length. Divide the four threads of the plait (one thread up, one down, one up, one down) so a tabby shaft appears and, while holding the bobbins of the plait, pull the plait backwards through the shaft with the help of a pin so a loop appears. Before securing the plaited loop in its curved form, place two threads from the closest short plait from the outer braid around the looped plait, in preparation for later making a stitch to join the looped plait with the outer border as described below (Fig. 10.27. A). Secure the loop by continuing to work the plait on the other side of itself. Fasten the loop on the outer side to the parallel border by using the two threads earlier placed around the curved plait to make a short plait (Fig. 10.27. B). Another braid of this remarkable type belongs to Hjärtlanda church (Fig. 10.28). Here the central section is made by two plaits, which make an ornamental knot. These

198 a

Lena Dahrén b

Fig. 10.27. A. Diagram showing the division of the threads of the left-hand plait around the loop plait and the threads of the looped plait itself. B. Diagram of the secured loop

Fig. 10.28. Braid with pattern of plaits formed into decorative knots. Hjärtlanda church

gold plaits were probably made separately in advance and rolled onto bobbins or some kind of similar tool to be workable and the knots were similarly made in advance and secured at the sides in the same way as suggested in the previous example. As mentioned above, these two bobbin-made braids are made in a totally different way from traditional techniques as practised today. Such a method of joining the plaits into the braid has, as far as I know, not been observed earlier in extant bobbin-made work. Hence, they are unique examples showing a combination of bobbin-made work

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

199

and passementerie and were probably carried out by the same maker. A braid with similar knots to those in the second braid, from Hjärtlanda church, is held in the Iklé collection, Switzerland. It has been analysed and a reconstruction of the technique was made by lace historian, Rosemary Shepherd (Fig. 10.29) (Shepherd 2009, 50). She solved the problem of making the knots by threading one of the two gimps onto a tapestry needle and passing the thread through the knot. However, she points out that the solution works only when making short braids and I agree with her conclusion that this was probably not the way that these knots were produced in the 1500s. I would suggest that they are more likely to have been made following the method set out above for the braid from Hjärtlanda church. How to make small variations Bobbin-made patterns can be created and varied by the Fig. 10.29. Reconstruction addition of a few threads and making small technical of an ornamental knot changes. From the study of extant narrow bobbin-made called ‘Double Coin Knot’. plaits it is possible to show how narrow, straight braids (© Rosemary Shepherd) can become pointed openwork edgings with a minimum of changes in the technique and addition of threads. For example, a straight four-strand plait can easily be transformed into a zigzag pattern by making picots alternately to the right- and left-hand side of the plait. Such a zigzag edging is preserved as a trimming of an embroidered handkerchief from the late sixteenth or early seventeenth century in Aspö church (Fig. 10.30. A & B). The use of the zigzag design as an edging is emphasised by attaching a spangle to the picots on only one side. As an edging this design is however unstable as the zigzag form tends to stretch. By adding two threads on one side to act as a selvedge or foot-side, the zigzag plait is stabilised and its use as an edging is improved as the point is secured by the selvedge threads (Fig. 10.31. A & B). Such an edging is preserved on a silk ribbon on a glove belonging to the Royal Armoury in Stockholm, and the design is found in Nüw Modelbuch, ‘Spitzle mit VIII’, p. Xr. However the extant edging is made of six threads not eight as suggested in the pattern book. This indicates that the pattern book suggests that the selvedges are to be made with a plait and not the twisted threads of the extant edging. By the further addition of two more threads, this time on the other side, the work loses its character as an edging and becomes again a straight braid consisting of a four-strand plait moving in zigzag between two straight lines, i.e. two twisted selvedge threads, one on each side of the plait. Such a pattern is preserved on a chalice-veil

200

Lena Dahrén

a

b

Fig. 10.30. A. Zigzag edging, Aspö church. B. Diagram of the edging

a

b

Fig. 10.31. A. Plait with a twisted footside, Royal Armoury, Stockholm, (3389 d, e) 21053-54. B. Design from Nüw Modelbuch

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

201

from Skokloster church and is included in the Nüw Modelbuch, named ‘Model mit XII’, p. XIIIv (Fig. 10.32. A & B). Again, the extant braid is made of eight threads not twelve as suggested in the pattern book. This indicates that the pattern was intended to have plaits on both sides of the braid, not the two twisted threads of the extant braid. The last and most significant change to the braid is purely technical and does not require any addition of threads – just knowledge of the technique. It is again a four-strand plait zigzagging between two foot-sides (Fig. 10.33. A & B). The change appears if the plait in the centre, instead of turning at one side, goes straight through the selvedge and continues as a four-strand plait on the other side of the selvedge. The plait turns in a point with three picots and returns, and then goes again straight through the selvedge and back towards the foot-side on the opposite side. This change converts the straight braid to a distinctive edging accented by three picots. Such a pattern is also included in Nüw Modelbuch named ‘Krönle mit XII’ p. IXr., and can be seen on a chalice-veil from Barva church where eight rather than twelve threads are used.

Conclusion Different methods of plaiting have been around for a very long time – and can be identified in dress depicted in sculpture of c. 800 to 600 BC from Mesopotamia (Assyria) and among the textile fragments found in the archaeological excavations at Birka, Sweden, dating to the ninth and tenth centuries AD. As the earliest extant bobbinmade braids are normally plaited grids or straight four-strand plaits with picots, it would appear that bobbin-made work arose from plaiting techniques. When the material used was gold and silver, the simple monotonous plaiting technique made the precious material stand out and sparkle. Bobbin-made braids appear to have been a new development in the late fifteenth or early sixteenth century coinciding with a growing demand for gold and silver braids among the upper classes. This was probably due to the fashion of attaching trimmings to the surface of dress fabrics made from silk or velvet. By developing the plaiting technique to form new patterns, the demand for seethrough glittering open-work designs could be met. The result of this development is what today is designated ‘bobbin-made lace’ however, as explained earlier, due to the terminological difficulties and the fact that there was no specific term for this work in the sixteenth century, I would prefer to call it ‘bobbin-made work’: that is ‘bobbin-made braids or edgings’. Among the earliest bobbin-made work are very narrow four-strand plaits which were used to create geometric grids or striped effects on fabric and to frame precious embroideries. For this purpose, an embroiderer could instead have used a continuous cord of plied metal thread or something similar as the intention was to further enrich the embroidered outlines that framed the work. The plait was not important in its own right but a secondary feature. This is probably one reason why there was no distinctive terminology for the earliest bobbin-made work.

202

Lena Dahrén

a

b

Fig. 10.32. A. Braid, Skokloster church. B. Design from Nüw Modelbuch

a

b

Fig. 10.33. A. Edging, Barva church. (© History museum Stockholm). B. Design from Nüw Modelbuch

The production of different kind of plaits was among the techniques practised by guild-regulated artisans called the passementiers (Diderot & d’Alembert 1967, 126–38). When the demand arose for decorative braided openwork made of gold, silver and silk thread, it was easy for the skilled artisan to vary the technique of four-thread

10.  From narrow four-strand plaits to openwork bobbin-made braids and edgings

203

plaiting and new decorative designs developed. It seems that the simplicity of the technique was its special virtue. Gold, silver and silk are shown to advantage when worked as bobbin-made braids. The relatively heavy metal threads needed for the technique were exposed to the light in such a way that they glittered and emphasised the appearance and the social status of the wearer.

Acknowledgements Thanks for financial support to Agnes Geijer’s Foundation for Nordic textile research, Berit Wallensbergs Foundation, Gyllenstiernska Krapperup Foundation and Foundation for textile research in memory of Else-Marie Zennström. All images are copyright of the author unless otherwise noted. The designs from the pattern books are in the Public Domain.

Archives National Archive, [Riksarkivet] Stockholm, Räntekammarböcker Royal Palace Archive [Slottsarkivet], Stockholm, Klädkammaren

Museums of studied textile artefacts Bodleian Library, University of Oxford: Douce Bib Eng. 1583 b.1. Collection Schnütgen, Cologne: Inv. No. 1878,1295 Borås textilmuseum: Södra Vånga B M 8868 Designmuseet Denmark, Copenhagen, Billedsamlingen: Nüw Modelbuch allerley Gattungen Däntelschnür Inv. No. 10519 Royal Armoury, Stockholm: Sveptröja (3376)31142; Silk ribbon (3389 d, e) 21053–54 Smålands museum, Växjö: Hjärtlanda No. D7 (L1597) Swedish History Museum [Historiska museet], Stockholm: Hjortsberga SHM 15662; Skokloster SHM 31797:8; Barva SHM 23128:3

Bibliography

Arnold, J. 1985. Patterns of Fashion. The Cut and Construction of Clothes for Men and Women c. 1560−1620 (London, Macmillan). Arnold, J. 1988. Queen Elizabeth’s Wardrobe Unlock’d (Leeds, Maney). Branting, A. 1907. Knytning, knyppling och språngning. In B. Salin (ed.), Fataburen (Stockholm, Nordiska museet), 105−118. Bringemeirer, M. 1985. Kinderkleid der Katharina Grafin zu Lippe von 1600. In M. Bringemeirer (ed.), Mode und Tracht: Beitrage zur geistesgeschichtlichen und volkskundlichen Kleidungsforschung (Munster, F. Coppenrath), 218–20. Burkhard, C. 1986. Faszinierendes Klöppeln nach Mustern des ältesten Klöppelbucher deutscher Sprache (1561) (Bern, Paul Haupt). Dahrén, L. 2010. Med kant av guld och silver: En studie av knypplade bårder och uddar av metall 1550−1640 (Uppsala University).

204

Lena Dahrén

Diderot, D. & d’Alembert, J. le R. 1967. Encyclopédie ou dictionnaire raisonné des sciences des arts et des métiers, 1751–1780. Vol. 12 (Stuttgart Bad Cannstatt, Frommann). Estham, I. 2010. Textilsamlingen från 1200-talet till 1860-talet. In H. Bengtsson (ed.), Uppsala domkyrka V, inredning och inventarier (Uppsala, Upplandsmuseet), 213−341. Levey, S.M. 1983. Lace: A History (London, Victoria and Albert Museum). Levey, S. & Payne, P. 1983. Le Pompe, 1559. Patterns for Venetian Bobbin Lace (Bedford, Ruth Bean). Shepherd, R. 2009. An Early Lace Workbook: Bobbin Lace Techniques before the Baroque (Leura, New South Wales, Lazy Daisy). Speiser, N. & Boutrup, J. 2011. European Loop Braiding. Part III: Loop Braiding in Swedish Bridgettine Tradition (Leicester, Jennie Parry).

Section IV Spinning

Chapter 11 The story of twist: Handspinning as a medieval craft Katrin Kania

Introduction While I never had the opportunity of meeting Peter Collingwood, his books and research have been invaluable to me, especially for my own ventures into tablet weaving and sprang. His attention to detail in the many textile techniques he wrote about, together with his always practical approach, are two things I truly admire. As my tribute to his work, I have taken a closer look at spinning with handspindles, specifically the technique used in medieval Western and Central Europe for spinning wool. Spinning is one of the first steps in most textile production processes. Yet spinning with handspindles in connection with the study of archaeological textiles is rarely studied in depth, and the influence of the yarn properties on the resulting fabrics is often underestimated. This paper aims to give detailed understanding through a look at spinning from a craft perspective, examining the variations in technique, the differences between modern ‘revival’ spinning and historical spinning according to our sources and the influence of the tool versus the influence of the spinner. Theoretical basics are combined with technical considerations and personal experience to show and explain the interplay of fibre, distaff, spindle and spinner.

Modern versus historical handspinning Spinning fibres into yarn is the first step for most textile techniques. Handspinning is a process that includes a wide range of varying techniques, and deserves a close look at its many details. Yet, due to the industrialisation of textile production, in most of Europe there no longer is a living tradition of handspinning. Recent years have seen

208

Katrin Kania

a revival of spinning with both the wheel and the spindle, as evidenced by a variety of how-to literature (Baines & Messent 1984; Field 1986; Boeger 2008; Franquemont 2009; Austin 2018). Regarding the process of spinning with a handspindle, it is often assumed that this did not change over the whole of human history. Anthropological data, however, shows that many different techniques for using handspindles existed and still exist today, using a range of tools that may appear similar, although they are used in different ways. An overview of different forms of contemporary handspindles and distaffs can be found on the pages about spindle typology of the University of Innsbruck (University of Innsbruck 2021). The assumption of one single unchanged handspindle spinning technique throughout history thus does not hold up. Similarly, comparison between archaeological finds of both finished textiles and the tools used to produce them shows that not only are historical spindles different from modern handspindles, the product is different as well. The yarns produced by handspinning today usually conform to the modern tastes of spinners and knitters. The modern use of the handspindle may differ greatly from the methods used in prehistory and history. This is evidenced by comparison of the tools, visual evidence of their use from historical images and comparing surviving examples of historical yarn with yarns produced with the typical modern spinning method. Historical techniques lead to significantly different outcomes than the typical result of spinning with modern handspindles using modern techniques.

Yarn influences The yarn produced by spinning will strongly influence the characteristics and quality of a fabric made from it, and can result in major differences between fabrics. The yarns used will influence drape, handle and resilience of a fabric, and consequently of any item produced from the fabric. Historical spinning techniques will produce yarns that differ from typical modern yarns, and fabrics produced from them will, in general, be very different from fabrics available today. It is thus helpful to understand the spinning process in detail when examining, comparing and researching historical textiles. Spun yarns are the building block of textile crafts, as they are the basic material for almost all other textile techniques. The most prominent of these is weaving fabrics. It might be considered the most important as well, as the amount of yarn necessary for a fabric is significantly greater than that needed for techniques such as, for instance, bandweaving, knitting, netting, nålbinding, sewing, embroidery and braiding, which also all depend on yarn as their basis. As with all other crafts, the material has a strong influence on the properties of the finished product. There is not only a significant difference between the fibres derived from plants and those from different animal origins. Different breeds of sheep also produce wool with markedly different characteristics. Wool preparation and the resulting yarn properties such as fibre alignment within the yarn, the yarn diameter and

11.  The story of twist: Handspinning as a medieval craft

209

its regularity, and the amount of twist inserted into the fibres will also have an influence. Especially the importance of the amount of twist is often underestimated today.

Basic principle of spinning Turning fibres into yarn is done by attenuating fibres into a band and adding twist. This is the basic principle of the process, regardless of the tool or machine used. The twist locks the fibres in position through pressing the fibre surfaces together. The friction between these surfaces gives the yarn its stability against tension. Hence fibre surface structure is one of the factors determining how much twist is needed to achieve a stable yarn. The smoother surfaces of vegetable fibres will require more pressure – more twist – to lock together than wool fibres of similar length with their characteristic scaly surface. Fibre surface structure, however, is only one of three factors determining the amount of twist needed. The other two factors are fibre length and the amount of fibres in the drafting (see Fig. 11.1). Both have an influence on the total surface area in the yarn, and thus on friction. While flax, hemp, or nettle fibres also have a

Fig. 11.1. Diagram of the three factors determining the necessary amount of twist to make a stable yarn. (© Katrin Kania)

210

Katrin Kania

Fig. 11.2. Microscope images of fibre surfaces, showing the differences between wool and flax fibres. Contemporary Karakul sheep and flax fibre. (© Antoinette Rast-Eicher, ArcheoTex)

smoother surface than wool, the much greater length of the individual fibre, if wellprocessed, compensates for this. Figure 11.2 shows the scaly surface of wool fibres and, for comparison, the surface of flax fibres. Similarly, when more fibres are twisted together, less twist is needed for a stable yarn.

Twist direction Twist direction is no significant factor in yarn stability. Direction of twist is usually described as ‘S’ or ‘Z’ according to the slant of the diagonal in the letter matching the slant of fibres on the yarn surface. While yarn stability does not depend on the twist direction, products made from the yarns will have different properties depending on both the amount of twist and its direction. For instance, fabrics woven with yarns spun in the same direction in warp and weft will look and behave differently than fabrics woven with yarns spun in different directions in warp and weft. Twist direction also has an influence on optical effects. Examples are the spin-patterned fabrics from the Hallstatt salt mines, such as Hallstatt Textile 217, dating to the Bronze Age, or Hallstatt Textile 4, dating to the Iron Age (Grömer et al. 2013, 275, 322). For these fine, high-quality fabrics, both S- and Z-spun yarns were used. The different reflection of light from these yarns in the woven fabric results in a subtle difference in perceived shade and shine, so that the fabrics appear to be striped or chequered.

Measuring twist amounts While the description of twist direction is unambiguous and simple, describing other characteristics of a given yarn is not as easy. The amount of twist inserted can be described by several different methods. The modern textile industries use a formula

11.  The story of twist: Handspinning as a medieval craft

211

to calculate the ‘twist multiplier’, which is directly connected to the twist angle (Lawrence 2003, 15–17). For counting the actual twist inserted in an existing yarn, laboratory apparatus is available that allows to un-twist a piece of yarn of a given length in a controlled way while counting the twists, then calculating the twists per length unit (usually inch or metre). Modern handspinners use a similar operation for manually counting the twists per inch. This destructive approach of un-twisting is obviously not suited for use with historical fabrics or archaeological textiles, and thus another method of measuring and describing the amount of twist is needed. This is the so-called ‘twist angle’. The twist angle measures the slant of the fibres on the yarn surface in relation to the long axis of the yarn. Measuring can be done with the help of a card or translucent piece of plastic, marked with appropriately slanted lines, or by taking a macro photograph and measuring the angle digitally (both shown in Fig. 11.3). In all cases, measuring the twist angle is a non-invasive, non-destructive procedure. However, like many other measurements of textiles, measuring the twist angle is not a precise operation due to several factors. Twist angle, much like yarn thickness, will vary along a given piece of spinning. This is true even for the very regular modern machinespun yarns, and much more so for handspun yarns. A higher twist angle will be found in thinner parts of the yarn, as these offer less resistance to being twisted. Any yarn with noticeable variations in thickness will thus have a greater variation of twist angles than a more even yarn. In addition, a handspinner can easily introduce

Fig. 11.3. Twist angle measurement with help of a measuring card (top) or with help of an image analysis programme (bottom). Left side shows a low-twist yarn, right side a high-twist yarn. (© Katrin Kania)

212

Katrin Kania

more or less twist in evenly spun yarn sections, which results in much more twist angle variation than in machine-spun yarns of even thickness. While twist can travel along the spun yarn, it does not even out completely. The measurement process itself also contributes to the imprecision of the twist angle as a yarn descriptor. Even under high magnification and with the aid of digital imaging and digital measuring tools, the exact angle of the fibres on the yarn surface remains hard to measure: the visible lengths of the fibres are short, and the yarn in archaeological textiles is usually not lying straight. As a final point, measuring the twist angle to get an indication of how highly twisted a piece of yarn is in relation to its thickness relies on the basic assumption that the fibres in their untwisted state lie perfectly parallel, and that twist is evenly entering the fibres. Fibre alignment in practice is never that perfect, and individual fibres may be wrapped around the surface of the yarn at an angle that does not correctly reflect the actual amount of twist in the yarn. As the stability of a yarn depends on the amount of fibres as well as the amount of twist, a measurement of twists per length has to be accompanied by an indication of the yarn thickness to be useful. In contrast, the twist angle gives a direct indication of the amount of twist in relation to the yarn thickness. A twist angle of 10° indicates a very soft-spun yarn, and a twist angle of about 30° would be classified as a high-twist yarn today. This makes twist angle a valuable measurement in spite of all the caveats listed above, and it would be highly desirable to have approximate twist angles listed as a part of the standard description of historical and archaeological textiles.

Twist in modern and historical handspinning The author organised a handspinning experiment in 2009 which took place during the first European Textile Forum in the Openlucht Museum Eindhoven, The Netherlands. The aim of the experiment was to collect data about the influence of fibre, handspindle weight and rotational properties and the spinning person on the resulting yarn. Fourteen handspinners participated in the experiment, spinning with five different spindles and two differently prepared wool types (Kania 2013). According to data from this experiment as well as informal collection of data from modern handspinners, most prefer to spin a relatively low-twist yarn, with twist angles commonly between 10° and 20°. Surviving archaeological wool textiles, however, commonly have a much higher twist angle, usually around 30° or higher. This is true both for very early finds such as those from Hallstatt, where the Bronze Age yarns have a twist angle of 30–40° and up to 50°, and the majority of the Iron Age yarns have a 45° twist angle (Grömer et al. 2013, 59); as well as for later medieval finds such as those from Herjolfsnæs, where most yarns have a twist angle of 40° or 45° (Østergård 2004, 233–52). The lower twist of modern handspinners’ work corresponds to the fairly lowtwist yarns that are machine-spun and available commercially. Using the formula given by Carl A. Lawrence for the twist multiplier in industrial yarn production

11.  The story of twist: Handspinning as a medieval craft

213

and his data for typical twist multiplier and the density of wool, it is possible to calculate the range of twist angles for wool yarns. This range for warp yarns with the twist multiplier of 2050–2400 lies between 14.7° and 17.1°, with weft yarns spun to an angle of 12.7° to 17.1° (twist multiplier 1750–2050) (Lawrence 2003). Using the same formula, the twist multiplier for wool yarn with a 45° twist angle would be 7794, more than three times that of the high end given for modern warp yarns. Most fabrics today are consequently woven from yarns that have a much lower twist than historical yarns. As a result, we are all used to these lower-twist yarns, which makes it unsurprising that modern spinners tend to reproduce the textiles they are familiar with. This happens even though yarns with higher twist might be preferable due to their higher durability. Figure 11.4 shows the properties of hightwist yarns compared to low-twist yarns. For a handspindle spinner, it would pose no great challenge to add a little more twist to the yarn. So why are most modern spinners producing soft-spun yarn on the handspindle? Apart from the tendency to reproduce the familiar feel of yarns, most modern handspun yarns are produced for the enjoyment of the spinning process. This usually means production in relatively small amounts, and the finished handspun yarn is usually plied and then used for knitting. A prized quality of knitting yarns today is their softness, which excludes high-twist yarns. Modern spinning wheels also favour soft-spun yarn through the relation of their winding and twisting actions, as opposed to the handspindle which allows the spinner to spin yarns with any amount of twist desired.

Fig. 11.4. Influences of twist on yarn and fabric properties. (© Katrin Kania)

214

Katrin Kania

For historical yarns, we can assume that durability was probably a prized quality, as textile manufacture took up a considerable amount of time. Expending that much time, energy and effort for a product that wears out quickly would be uneconomic. Another factor that differentiates historical spinning, done for the production of fabrics, from most cases of modern handspinning is the amount of yarn produced. A sweater or a large lace shawl will need about 800 to 1800 m of yarn, depending on size and gauge. This is a small amount of yarn as compared to the requirements for weaving a bolt of cloth. In a piece of fabric 5 m long and 1 m wide, with 15 threads per cm, about 15,000 m of yarn are used. This is not counting the extra yarn needed at the start and end of the piece, and the shrinkage after taking the piece off the loom, so at least 10% more yarn should be estimated for weaving such a bolt of fabric. A piece this size would be enough for a woman’s dress or a tunic and hose for a man – 15 km of yarn is thus not an unusual figure, but rather a middling to low one in medieval spinning production.

Medieval-style handspindle characteristics The tool used for production of yarns in such a massive scale looks deceptively simple, but fulfills the double purpose of most spinning tools: helping to add twist to the fibres and providing storage for the spun yarn. Though it is possible to spin yarn on just a stick or a hooked stick, the typical tool for production is the handspindle, consisting of a spindle stick and a spindle whorl. Judging from archaeological evidence, many if not all handspindle finds from all ages share some characteristic traits, as the example in Fig. 11.5 shows.

Fig. 11.5. An example of a wooden spindle from Pfakofen near Regensburg, Germany, sixth century. It is c. 25.5 cm long, tapers towards both ends and has a notch at one end. (© Historisches Museum Regensburg, photo Michael Preischl)

11.  The story of twist: Handspinning as a medieval craft

215

A characteristic trait for the spindle stick is having the thickest part of the spindle at about two thirds of the length, tapering towards both ends. Spindles are usually round in cross-section and are sometimes provided with a hook or notch at one or both ends, though most have plain ends. The tapered form serves several purposes. The taper in the longer part, resulting in a long slim upper shaft, has two effects. Firstly, the distance between the long axis of the spindle and the attachment point of the spindle to the yarn when spinning influences the rotation of the spindle: the larger this distance, the more wobbly the spindle will run, so a thin end improves the  rotation of the spindle. Secondly, the slim spindle shaft has a positive effect on the rotational speed. When flicking a stick or dowel into rotation between the fingers,  there is only a limited range of motion, and a limited area (the width of the finger or fingers) that can be used to set the stick in motion. A thinner stick is much easier to set into motion in this way than a thicker one, and it will have higher rotational speeds. As this speed determines how much time it will take to insert the required amount of twist into the yarn, it is directly connected to the maximum speed attainable during production. On the shorter (usually the lower) part of the spindle from thickest part to end, the taper allows for securely fitting spindle whorls with a range of bore sizes to the spindle. Medieval spindles thus do not need to have their spindle whorls attached permanently. The removable whorl has several advantages. One of them is the possibility to adjust for changing rotational properties of the spindle. As the yarn builds up on the spindle stick, the spindle does not turn as easily. At this point, the whorl can be replaced with a lighter whorl, returning the spindle to better rotational properties again. The same set of whorls can be used on several spindles. When several spindle sticks are available, it is possible to store spun yarns on the sticks until needed; it is also possible to use the yarn as a weft yarn directly from the spindle, using the stick as a shuttle (Marled Mader pers. comm., May 2020). However, Fig. 11.6. Comparison the removable whorl means that winding yarn onto the of a typical modern spindle has to be done differently from the method most h a n d s p i n d l e ( l e f t ) and a medieval-style modern spinners employ, when they let the yarn build spindle (right). Note the up in cone-form directly on top of the whorl. Figure 11.6 differences especially shows a comparison of a common type of modern at the spindle tip, the handspindle with one in the medieval style, including the different way of winding method of winding yarn. If this modern method is used on the finished thread, and with a medieval-style spindle, the pressure of the yarn the different shape of the cone will eventually push the whorl off the shaft and the whorl. (© Katrin Kania)

216

Katrin Kania

cone collapses. Instead of using the spindle whorl as a support for the stored yarn, the winding is done around the middle of the spindle stick, keeping the bottom of the yarn package always at some distance to the whorl. A stable package is achieved by an upward and downward motion of the yarn while winding, resulting in a crisscrossing winding pattern building upwards and downwards on the spindle shaft from its starting point. This results in an ‘egg’ or ‘rugby ball’ shaped form for the yarn package. If the yarn package is kept completely above or below the thickest part of the spindle stick, the package can even be pulled off the stick and will still remain intact.

Influence of spindle weights on yarns The spindle whorl itself, especially its rotational properties and its influence on the spun yarn, has been the topic of many debates in recent years. While there are physical and bio-dynamical limits on the thinness of a yarn spun on a heavy spindle or the thickness of the yarn spun on a light spindle, this influence might be much overestimated in common assumption. The spinning experiment from 2009 had the fourteen spinners from different backgrounds use spindles of 5 g, 15 g and 52 g weight, while having the same moment of inertia (MI), meaning the mass distribution of the spindle in relation to the axis of rotation. This was achieved by calculating the dimensions necessary for the individual spindle whorl types, then forming the whorls accordingly. The MI determines the rotational properties and thus how the spindle feels to the spinner when flicking it into motion, and how long and how fast it will turn. Table 11.1 shows resulting yarns in terms of tex spun (tex being the measurement of yarn weight per 1000 m of length) with the two types of wool used, combed Merino and carded Bergschaf wool. Listed are minimum and maximum grist (yarn thickness, described as the relation between weight and length of the yarn) as well as the median (half of the data points lie above this number, and half below) and the average (all grists added together and divided by the number of data points). A large difference between median and average shows that there are outliers in the data. While the 52 g spindle is 10 times the weight of the 5 g spindle, the grist spun on that spindle across all spinners is only around 1.7 times the grist spun with the 5 g spindle, and the maximum grist spun on the 52 g spindle is only a third larger than the maximum grist spun on the 5 g spindle. Tables 11.2 and 11.3 show the relation of weights and grists across spindles with the 15 g spindle and the 5 g spindle as reference. Technically, there is only one definitive influence of the spindle weight on the spun yarn: when spinning suspended, that is with the weight of the spindle hanging from the yarn already spun, the spindle weight will determine a minimum supported weight or minimum supported tension of the yarn produced. For yarns that are used for weaving or other textile techniques that put a controlled amount of stress or tension on the yarn, it is crucial that the yarns do not snap during use. If threads break in the loom, it means much more trouble and loss of time than a yarn breaking during spinning which can be mended easily. Spindle (whorl) weight can thus be

217

11.  The story of twist: Handspinning as a medieval craft Table 11.1. Yarn grist measured in tex across different spindles Spindle weight

Average grist overall

Median grist overall

Min. grist Merino

Average grist Merino

5 g 15 g 52 g

142 tex 165 tex 237 tex

139 tex 143 tex 233 tex

55 tex 67 tex 98 tex

129 tex 138 tex 186 tex

Median grist Merino

Max. grist Merino

Min. grist Bergschaf 126 tex 241 tex 63 tex 125 tex 354 tex 114 tex 172 tex 287 tex 192 tex

Median grist Bergschaf 146 tex 169 tex 286 tex

Max. Average grist grist BergBergschaf schaf 155 tex 308 tex 193 tex 289 tex 229 tex 392 tex

Table 11.2. Relation of spindle weights and grists spun, reference spindle 15 g Spindle weight

Relations compared to 15 g spindle

Relation grist average overall

Relation grist median overall

Relation grist maximum overall

Relation grist minimum overall

5 g

33

86

98

87

82

15 g

100

100

100

100

100

52 g

347

144

163

111

145

Table 11.3. Relation of spindle weights and grist spun, 5 g and 52 g spindle Spindle weight

Relations compared to 5 g spindle

Relation grist average overall

Relation grist median overall

Relation grist maximum overall

Relation grist minimum overall

5 g

100

100

100

100

100

52 g

1040

167

166

127

176

chosen on purpose to ensure yarn of a certain stability, using the tool weight for in-built quality control. In the spinning experiment results, there was no influence of the MI discernible in the spun yarns. The spinners did report in their questionnaires, however, that the spindles turned very differently, either slower (for spindles with a high MI) or faster (for those with a low MI). As the spindles all had a similar weight, in this instance the effect was solely connected to the change in MI. Most modern spindles, when flicked, will turn much more slowly than spindles in the medieval style. In that case the effect is probably due both to a difference in MI, with most modern spindle whorls having a significantly higher MI than medieval forms, and a thicker top part of the spindle stick in modern spindles, where the flicking motion takes place.

Methods of handspinning Just as there are seemingly subtle but actually influential differences between modern handspindles and the medieval spindle finds, there are significant differences between the spinning styles that can be used to make yarn. When looking at ethnographic data

218

Katrin Kania

from the last fifty years, there are still many different styles, techniques and tools, depending on the tradition of a given region, the fibre material and the purpose of the spun yarn. Two spinning techniques are relevant to medieval European spinning, which will be examined a little closer in the following passages. A third technique will be described for comparison, as it is typical for modern Western handspinning. Handspinning terminology is just as unfixed or contradictory at times as terminology in many other areas of textile work, especially regarding historical textiles. In the following, three terms will be used to refer to these styles, indicating the position of the spindle in relation to the spinner: ‘short suspended’ and ‘spinning in the hand’ for the two medieval styles, and ‘long suspended’ for the modern Western technique. Most people who have an interest in handspinning are familiar with the modern Western technique of spinning long suspended, shown in Fig. 11.7. When searching for spinning instructions on the internet, many pages and many YouTube videos

Fig. 11.7. Typical body and hand position of spinners using handspindles long suspended, both with and without a distaff. Note that body and hand position are very similar even when a distaff is used. Photo taken during the spinning experiment 2009. (© Katrin Kania)

11.  The story of twist: Handspinning as a medieval craft

219

will explain how to spin in this technique. For spinning long suspended, the fibre is typically held in one hand. The spindle is attached to the already spun yarn and hangs fully suspended. Both hands will stay close together during the spinning process, drafting fibres downward to form the yarn, while the spindle moves downward away from the fibre supply and the hands. The sequence of action is flicking the spindle into motion, then drafting. Right after flicking it into motion, the spindle rotates at a high or very high speed. As it transfers twist to the fibres, the spindle slows down and finally stops, and it becomes necessary to set the spindle into motion again. For this, the lower hand (not holding the fibre) moves down to flick the spindle again. When the spindle reaches the floor, the yarn has to be wound around the spindle stick. After winding, the spindle is secured to the yarn again, and the spinning process continues. The body and hand position of this spinning style are characteristic, and they are significantly different from the body and hand position shown on medieval illustrations of spinners (see Fig. 11.8). The first obvious difference is the presence of a distaff on medieval spinning pictures, while most modern spinners work without a distaff. The fibres are fixed to the distaff, which is a stick long enough to be tucked under the belt, under the arm or held between the knees. Alternatively it can be attached to a stand or a foot so that it stands on its own, but the version held by the spinner seems to have been much more common. Apart from being shown on every spinning image, the importance of the distaff for historical handspinning is also indicated by the fact that the poetic name for women is not the ‘spindle gender’, but the ‘distaff gender’, and that spinning is sometimes indicated as ‘spinning on the rock’. In German, the distaff is still called Rocken or Rockenstab, and spinning conventions even in the age of the spinning wheel were called either Spinnstube (spinning chamber) Fig. 11.8. Typical body and or Rockenstube (distaff chamber). The distaff is usually held on the left side of the hand position of spinners using body, the left hand is shown close to the distaff handspindles short suspended or in hand. Detail of a miniature showing and the yarn is formed between the left and the Eve spinning, Speculum humanae right hand. The right hand travels downward and salvationis, attributed to the Master towards the right side of the body, keeping the of the Dark Eyes, England, 1485−1509, already-spun yarn of the currently spun length British Library MS Harley 2838, fol. 5. tensioned diagonally across the body. (Public Domain picture graciously For spinning in the hand with the distaff on provided by Catalogue of Illuminated the left, the spindle is held in the right hand Manuscripts, British Library)

220

Katrin Kania

and continually turned between the fingers of this hand. The spinning action is much less intermittent than the action for spinning long suspended, as the right hand inserts twist at a continuous rate. The right hand starts close to the distaff, moving away in a downward and sideways motion, ending with the arm more or less fully extended downwards and slightly to the side of the body. This movement downwards and across is the drafting movement, while the left hand controls the fibre flow at the distaff to ensure an even thickness of yarn. As with spinning long suspended, the yarn has to be wound around the spindle at intervals, in this case at the latest when the right arm is fully extended down and to the side of the spinner’s body. For winding on the yarn, practical experience has shown that it is best when the left hand stays close to the distaff in its spinning position. Pinching the area between the unspun fibre and the yarn with the fingers prevents twist travelling up into the fibre supply and inadvertently tearing the already spun yarn from the drafting area when winding on. The pinch of the left hand makes it possible to keep the yarn under enough tension for easy winding. The hand holding the spindle does all the other actions required for winding on, that is unwinding the yarn spiralling towards the tip of the spindle shaft, then winding the spun length onto the storage area, forming a cop, while gradually moving towards the distaff. This part of the process is finished by winding the last length of spun yarn into a spiral towards the spindle tip in preparation for spinning the next length of yarn. The basic action and movement for spinning in hand and spinning short suspended are very similar to each other. As with spinning in hand, one hand always stays close to the distaff and controls the fibre flow, while the other hand starts close to the distaff and moves downwards and across the body with the drafting movement. When spinning in the short suspended manner, however, the spindle is not held in the hand, but hanging just below it. The yarn is fixed to the top of the spindle either with the help of a notch or using a half-hitch. Most medieval spindle finds have a plain tip, so a half-hitch would have been needed for spinning short suspended. When the suspended spindle is flicked into motion, it will turn freely and with high speed. Repeating the flicking in short intervals means that the spindle continuously rotates at a high speed. After the hand and arm drafting the fibres are fully extended, a little more length of yarn can be spun than with spinning in the hand: the spindle can travel slightly away from the right hand, hanging from the lengthening yarn. Several medieval images show the spinner at this stage, with the spindle at a distance from the extended right hand. Winding the yarn onto the spindle is done in the same way as when spinning in hand. If an extra bit of length has been spun, this can be taken up by moving the left hand to move the spindle back up towards the right hand, or by winding some of the yarn around the fingers of the left hand. The right hand then removes the half-hitch, winds the yarn onto the spindle, and finally places a new half-hitch in preparation for the next drafting and spinning process.

11.  The story of twist: Handspinning as a medieval craft

221

Change of predominant spinning method The reason for a change in the predominant method for handspindle spinning in the Western European area can most probably be traced back to various craft revivals, including those in the 1960s and 1970s, when a number of textile crafts, among them spinning, were rediscovered and found great success as a pastime and hobby. Lacking a continuous tradition of handspinning, a technique closer to that used in the Andes or in other areas of the world was imported. In demonstrations explaining historical handspinning today, the long suspended method is commonly used, and the difference in body position and the lack of a distaff is usually not questioned. Sometimes a distaff is used to store the fibre, but the technique is still long suspended (see Fig. 11.7). As this method is for most of us the only method ever seen, and the method taught to beginning handspinners, the association of handspinning with long suspended body position and sequence of movement is now firmly ingrained. Using a spindle to spin seems such a basic technique and so simple in principle that it does not readily occur to people that there can be a multitude of methods and tool variations. This is reinforced by explanations about and demonstrations of handspinning in museums, living history presentations and similar cultural events, where the long suspended technique is presented as that used in the Middle Ages. All these circumstances together contribute to a blind spot regarding the differences between modern and historical handspinning, as the author herself regrets to have had for many years.

Spinning wheels Another common misconception in popular knowledge about spinning is the development and use of the modern treadle-driven flyer spinning wheel. This type of wheel is often assumed to date back to medieval times, again reinforced by the use of these wheel types in ‘living history’ settings without sufficient explanation of its more recent date. There is a form of spinning wheel dating back to approximately the middle or end of the twelfth century, basically formed by a spindle turned sideways and driven by a large wheel, known as the Great Wheel. For using this wheel type, one hand turns the wheel, while the other hand holds the fibre supply and drafts the yarn out against the spindle. It is similar in principle to the hand-driven charkha wheels used in cotton spinning. The technique of one-handed drafting against the spindle tip relies on the controlled insertion of twist into the fibres. Since the twist always settles in the thinner parts of the drafted band first, locking them together, it is possible to draft out an even yarn. Gentle, controlled tension on the yarn helps to attenuate the yarn if thicker spots have occurred; a good balance between the twist insertion speed and the drafting speed enable the spinner to draft out an even amount of fibres.

222

Katrin Kania

This technique is especially suitable for spinning short and slippery fibres such as cotton fibres, and this wheel form was probably developed for cotton spinning and travelled to Europe in the later twelfth century, where it was then also used for spinning wool (Mazzaoui 1981, 78). Since wool fibres are much longer and much less slippery than cotton fibres, however, a special preparation of the fibres is necessary to enable them to be spun with the Great Wheel. For spinning an even, high-quality worsted yarn as it is done with the handspindle in both in-hand spinning and long or short suspended spinning as explained above, combing the wool as shown in Fig. 11.9 prepares the fibre in a way well suited to this kind of spinning. Wool combs or wool comb tines are known from archaeological excavations, though they are relatively rare (Goodall 1980, 34, fig. 40; Walton Rogers 1997, 1727–31). There are also a number of illuminations showing wool combs being used. For combing, a pair of combs with long tines is used; one of them with the tines pointing upwards, the second one with the tines pointing sideways. In medieval depictions, the comb with the tines pointing upwards is usually fixed in a stand, while the other comb is held by the worker (Cardon 1999, 178–80, figs 54, 55 & 57). Fibres are loaded onto the hand-held comb, which are then transferred to the other comb by drawing them through its tines repeatedly. In this process, longer fibres are

Fig. 11.9. Woman combing wool with two wool combs. Detail from the Smithfield Decretals, c. 1330−40, British Library MS Royal 10 E IV, fol. 138r. (Public Domain picture graciously provided by Catalogue of Illuminated Manuscripts, British Library)

11.  The story of twist: Handspinning as a medieval craft

223

gradually transferred to the second comb, while the shorter fibres, second cuts and noils remain in the first comb to be discarded, and remaining vegetable matter falls out of the fibre mass. The combing, transferring fibres from one comb to the other, is repeated as many times as necessary to achieve clean, evenly combed fibres. They

Fig. 11.10. Spinning on the Great Wheel. The image shows one woman spinning on the wheel, while a second woman is carding fibre in preparation. The finished rolls of carded fibre are collected in the basket. Detail from the Luttrell Psalter, c. 1330s. (© The British Library Board, Add. MS 42130, fol.193)

Fig. 11.11. The author’s reconstruction of a Great Wheel based on fourteenth-century manuscript illuminations. Overall height c. 200 cm, diameter of wheel c. 140 cm. (© Katrin Kania)

224

Katrin Kania

are then pulled from the comb to form a ‘top’. These lengths of combed fibre can then be used to dress a distaff, keeping them parallel to the distaff stick. Drafting is done next in the same direction, resulting in yarn that can be described as ‘true worsted’. Medieval images of wheel spinning show the spinner turning the wheel with the right hand, while the left hand holds the fibre supply and drafts. This means there is no possibility for fine control of the twist insertion as occurs when spinning short suspended or in the hand. To draft an even yarn, the spinner has to time the drafting motion of the left hand with the twist insertion provided by the right hand. This onehanded spinning technique works best with a fibre preparation that allows the wool fibres to easily slip along each other even when some twist has already been inserted. It is possible to use combed top to ‘spin from the fold’, this process takes apart the combed top to form short pieces, which are then folded up to spin from the middle of the folded piece. This results in a yarn that is not true worsted. Whether this possibility was used at all in medieval times is not clear; there is no firm evidence for it according to visual sources. All the images the author has seen that illustrate wool preparation in context with a spinning wheel of the Great Wheel type show wool processing with the help of cards and forming the wool into a rolag. An example of this combination can be seen in Fig. 11.10. For carding, two cards are needed; these are wooden paddles equipped with many small wires, slightly hooked into the direction of the handle. Modern cards are covered with a thin rubber or cloth-and-rubber mat that holds the carding wires, while old cards had a leather cover holding the wires in place. A small amount of fibres is placed on one of the cards, and the second card lightly brushes over the first one, the wires just touching but never interlocking. Once the fibres are aligned to the satisfaction of the person carding, they are taken off the cards and rolled up, with the axis of the roll parallel to the front edge of the card. There are several methods that can be used for carding, using different motions and sequences, an in-depth description of which is outside the scope of this article. All methods align the fibres mostly parallel to each other and perpendicular to the front edge of the cards. They have in common that whatever is placed on the card will end up in the resultant rolag. No fibre sorting takes place when carding, as opposed to combing, so there is no loss of fibre mass. However, this means that any short fibres or debris will stay in the final preparation, resulting in lumpy yarn containing vegetable matter. For a good quality of yarn and for even spinning, pre-sorting of fibre is thus necessary, and this sorting has to be done carefully and thoroughly before the fibre is placed onto the cards. Oiling the (previously washed and dried) wool before carding also helps with even spinning, as the lubricant facilitates even drafting especially if residual lanolin is in the fibres. There are several manuscript depictions of one person spinning at the Great Wheel while another person is carding the fibres in preparation. The best known feature in the Smithfield Decretals, c. 1330s and a French translation of Giovanni Boccacio’s De mulieribus claris, referred to as Des Clères et Nobles Femmes, c. 1470. The illumination

11.  The story of twist: Handspinning as a medieval craft

225

from the Luttrell Psalter, reproduced in Fig. 11.10, shows not only a spinner and a carder, but also the finished rolags placed into a basket, which is positioned between the two women. From personal experience in working with a reconstructed wheel of this type (see Fig. 11.11), the author can say that it is a very fast and efficient way of producing yarn. The resulting yarns, if spun to a similar thickness and a similar twist angle, are not dissimilar to those spun on a handspindle, though they are a little less resilient, may look a little more hairy and tend to be less even. The sequence of action is setting the wheel into motion with the right hand and simultaneously starting the drafting of the fibres from the rolag with the left hand. The right hand keeps turning the wheel and the left hand keeps drafting while the spinner walks a few steps backwards to spin a longer piece of yarn in one go. Once the limit of the yarn length has been reached, the left hand pinches the yarn to prevent twist from entering the undrafted fibres and the right hand turns the wheel until enough twist has been transferred to the yarn. Then the right hand stops the wheel, turns it backwards enough to unwind the spiral of yarn running from the cop to the tip of the spindle and starts turning the wheel in the spinning direction again to wind up. The spinner walks toward the spindle during the winding process. When only a little bit of yarn is left, it is led towards the spindle tip, again forming a spiral around the spindle running towards the spindle tip; the left hand releases the yarn previously pinched off and starts the drafting motion again. As with the handspindle, it is desirable to insert a large amount of twist as quickly as possible into the yarn for efficient spinning. As there is no gearbox for speed-changing in these early wheel types, a very large wheel in comparison to the drive pulley on the spindle is necessary. This results in the large wheels shown in the medieval illuminations. Starting and stopping the wheel again and again in this sequence requires an appreciable effort even with a lightweight wheel that is running smoothly. The author’s reconstruction of such a wheel is made from ash wood and weighs about 2.4 kg. Such a wheel requires more materials and much more effort to build than a handspindle, which can be whittled from a stick and equipped with almost any object as a whorl, provided this object can be bored through to fit it onto the spindle stick. While a spindle wheel does not take extraordinary skills to build, equipment for working wood and bending wood is definitely needed, as well as a sound understanding of how the wheel works. So what was the reason for the introduction of this rather unwieldy tool? The spindle wheel allows faster production of yarns as compared to the handspindle. Even when not highly practised on the wheel, the author achieves about double the output of yarn compared to spinning short suspended and almost triple the output compared to spinning long suspended, using the same fibres and spinning to about the same thickness and twist angle. The differences between the resulting yarns were relatively small, and the wheel-spun yarn would be stable enough for weaving.

226

Katrin Kania

Spinning speeds The amount of time required for spinning was an important issue for medieval textile production. Four factors influence the amount of time necessary to spin a given amount of yarn: 1) the materials and their preparation, including the quality of their preparation; 2) the tools used, including the issue of how well an individual crafter can work with a specific tool, and the tool’s suitability for the materials; 3) the thickness of the yarn, with fine and even yarn requiring more twists per length and thus more time; and 4) the craftsperson himself or herself, as individual skill, preferences and day-to-day condition will also influence productivity. While these influences can be traced largely through the experiences shared by modern spinners, there is limited evidence on how much time was actually spent on spinning in the Middle Ages. Written evidence is sparse and in consequence it is extremely difficult to gauge spinning worktimes, apart from estimates based on the time taken by one or two modern handspinners to complete a certain task. Unfortunately, this procedure often only multiplies the uncertainties. There are a few estimates of spinning times, some of which come from older sources. Recent speed spinning tests were carried out by Karina Grömer, who calculates that 80 m per hour would be a possible output for a good professional spinner (Grömer pers. comm. 2019). Spinning tests at the Centre for Textile Research (CTR) in Copenhagen yielded an output of 42.9 m per hour, averaged over 42 spinning tests undertaken by two spinners (Mårtensson et al. 2006), while the spinning experiment from 2009 cited above showed a large range of spinning speeds with a maximum output of 72.7 m per hour and an overall average of 34 m per hour. In contrast to these figures, Almut Bohnsack in her book, Spinnen und Weben, estimates the average output of a 15 g handspindle to be about 120 m per hour. This estimate is a calculation, based on measuring the rotational frequency of the spindle suspended from two fine threads during 6 seconds (Bohnsack 2002, 52–55). There is still insufficient data about spinning speeds to be useful for an extended analysis, as they omit information about the yarn thickness, twist angle or number of twists per length and the exact specifications of the spinning tool. In the case of handspindles, spinning technique and the weight and form of spindle and spindle whorl are also of interest, and are rarely, if ever, stated in records of spinning trials.

Time estimates for textile crafts The question of how much work time was required to complete a certain task is a frequent one for those researching historical craft procedures. Time consumed is an indication of the value of a piece of work. Even though prices for materials and prices paid per hour of labour differ significantly between the Middle Ages, Early Modern era and today, there is always only a limited amount of time available for work, especially intricate work, during a day. This means that even though wages

11.  The story of twist: Handspinning as a medieval craft

227

may be hard to estimate or trace back, an item that needs only a few hours in the making will be cheaper than an item using similar materials but needing a high number of hours. Several hundreds or even several thousands of hours can easily be spent making and decorating a textile such as an embroidered coronation mantle or an altar-frontal. For instance, an account of payments made in 1271 lists wages for four women working on an altar-frontal for the high altar in Westminster Abbey for three and three-quarter years (Staniland 1991, 9). Providing a time estimate for the manufacture of a given item is, at best, an educated guess. Changing techniques, raw materials and tools can lead to significant differences in work time required. This impact on work time may not be obvious without trials of the different techniques, tools or materials. In addition, the goals of the craftsperson may have changed. For instance, modern hand-woven fabrics are prized for their slight imperfections and unevenness, which sets them apart from most machine-woven fabrics. Coarse, loosely woven textures are also quicker to produce and thus more affordable to the modern customer. In contrast, medieval fabrics were esteemed for their fineness and evenness. The fineness of the fabric also has a huge impact on the production time, since more time is required for spinning, for setting up the loom, and for weaving. A square metre of cloth, woven with five threads per cm in warp and weft, contains approximately 1000 m of yarn. Five hundred wefts have to be brought in to weave the length of 1 m. A fabric with 25 threads per cm in warp and weft contains about 5000 m of yarn, and 2500 picks have to be made for weaving the 1 m of length – a factor of five which will also mean at least a factor five in time required, if not more, as finer work often needs a little more care and thus a little more time. Number games regarding historical craft work times can quickly come to a point where one cannot be sure of any number, especially if estimates for time needed are based on non-professionals working or on older records without sufficient contextual information. On the other hand, numbers have something like an aura of reliability surrounding them, making even rough estimates seem more factual than they might be. Is it then responsible to use them for publications, or for informing the wider public? It is definitely helpful to have and give ‘ballpark figures’ for certain tasks, and giving indications of time invested helps with the evaluation of tasks and their role in daily life. This is a good thing both for research and for education. However, it should be made as clear as possible that these figures are only rough estimates based on samples of individual craftsperson's performances. When using documented times from modern crafters, background information regarding tools, materials, technique and the amount of practice or proficiency are necessary to evaluate the information given and to enable comparison with other recordings of time needed. With more of these properly documented records, it may be possible eventually to gain a better understanding of both the time spent on doing a certain task and on the amount of variation between individual spinner’s performances.

228

Katrin Kania

Conclusion Spinning is both simple – as any variation of it turns fibres into yarn with the help of twist – and complicated, as there are many different techniques and tools. As yarn forms the basis for most other textile techniques, the process of spinning deserves attention in research, especially since yarn properties have a large influence on the resulting textile. Giving accurate descriptions of yarns by themselves or in surviving historical textiles is, however, not an easy task. Similarly, it is difficult to get information about the time necessary for making textiles in medieval times. Differences between the commonly used modern style of spinning with a handspindle (long suspended) and the medieval style (short suspended or in hand, both using a distaff) and differences between the yarns commonly spun by handspinners today and by medieval handspinners add to the general problems of estimating work times in historical circumstances. Altogether, research on historical spinning still has many open questions. Hopefully, this fascinating and complex topic will undergo more research in the future.

Acknowledgements This paper was made much richer by feedback from and discussions about spinning with many others, among them Isis Sturtewagen, Karina Grömer and the participants of the experiment undertaken in 2009. My thanks for help with the formulae to calculate twist multipliers and angles is due to Karl-Friedrich Pfeiffer.

Bibliography

Austin, P. 2018. Hand Spinning: Essential Technical and Creative Skills (Ramsbury, Crowood Press). Baines, P. & Messent, J. 1984. Spindle Spinning (Tunbridge Wells, Search Press). Boeger, L. 2008. Intertwined: The Art of Handspun Yarn, Modern Patterns, and Creative Spinning (Gloucester, MA, Quarry). Bohnsack, A. 2002. Spinnen und Weben. Entwicklung von Technik und Arbeit im Textilgewerbe (Bramsche, Rasch Verlag). Cardon, D. 1999. La Draperie au Moyen Age (Paris, CNRS Editions). Field, A. 1986. The Ashford Book of Spinning (London, Dryad). Franquemont, A. 2009. Respect the Spindle (Loveland, CO, Interweave). Goodall, I.H. 1980. Ironwork in Medieval Britain: An Archaeological Study (Cardiff, University College). Grömer, K., Kern, A., Reschreiter, H. & Rösel-Mautendorfer, H. 2013. Textiles from Hallstatt. Weaving Culture in Bronze Age and Iron Age Salt Mines. Textilien aus Hallstatt. Gewebte Kultur aus dem bronzeund eisenzeitlichen Salzbergwerk. Archaeolingua 29 (Budapest, Archaeolingua). Kania, K. 2013. The spinning experiment – influences on yarn in spinning with a handspindle. In H.J. Hopkins (ed.), Ancient Textiles – Modern Science (Oxford, Oxbow Books), 11–29. Lawrence, C.A. 2003. Fundamentals of Spun Yarn Technology (Boca Raton, Florida). Mårtensson, L., Anderson, E., Nosch, M.-L. & Batzer, A. 2006. Technical Report Experimental Archaeology Part 1, 2005–2006. Tools and Textiles – Texts and Contexts Research Programme (Copenhagen, Centre for Textile Research).

11.  The story of twist: Handspinning as a medieval craft

229

Mazzaoui, M.F. 1981. The Italian Cotton Industry in the Later Middle Ages 1100–1600 (Cambridge, Cambridge University Press). Østergård, E. 2004. Woven into the Earth. Textiles from Norse Greenland (Aarhus, Aarhus University Press). Staniland, K. 1991. Medieval Craftsmen: Embroiderers (London, British Museum Press). University of Innsbruck 2021. https://www.uibk.ac.at/urgeschichte/projekte_forschung/abt/ spindeltypologie/spindeltypen.html.en Walton Rogers. P., 1997. Textile Production at 16–22 Coppergate. The Archaeology of York, The Small Finds 17/11 (York, Council for British Archaeology).

Illuminated manuscript references

Smithfield Decretals, c. 1330s: British Library Royal MS 10 E IV fol. 147v, digitised version at http://www.bl.uk/manuscripts/FullDisplay.aspx?ref=Royal_MS_10_E_IV. Des Clères et Nobles Femmes, c. 1470, French translation of Giovanni Boccacio’s De mulieribus claris: The New York Public Library, Spencer Collection MS 33, fol. 23v, digitised version at https:// digitalcollections.nypl.org/items/510d47da-ebc8-a3d9-e040-e00a18064a99.