X-Radiography of Textiles, Dress and Related Objects (Conservation and Museology)

X-RADIOGRAPHY OF TEXTILES, DRESS AND RELATED OBJECTS Butterworth-Heinemann Series in Conservation and Museology Serie...

Author: Sonia O'Connor | Mary Brooks

154 downloads 1047 Views 57MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

X-RADIOGRAPHY OF TEXTILES, DRESS AND RELATED OBJECTS

Butterworth-Heinemann Series in Conservation and Museology Series Editor:

Andrew Oddy Formely of the British Museum, London

Consultants:

Page Ayres Cowley Conservation Architect, New York David Bomford The J. Paul Getty Museum, Los Angeles John Fidler Simpson Gumpertz and Heger Inc, Los Angeles Velson Horie The British Library, London Sarah Staniforth The National Trust, Swindon Jeanne Marie Teutonico The Getty Conservation Institute, Los Angeles David Saunders The British Museum, London Architectural Tiles: Conservation and Restoration (Durbin) Chemical Principles of Textile Conservation (Tímár-Balázsy, Eastop) Conservation and Restoration of Ceramics (Buys, Oakley) Conservation of Building and Decorative Stone (Ashurst, Dime) Conservation of Furniture (Rivers, Umney) Conservation of Historic Buildings (Feilden) Conservation of Leather and Related Materials (Kite, Thomson) Conservation of Ruins (Ashurst) A History of Architectural Conservation ( Jokilehto) Lacquer: Technology and Conservation (Webb) The Museum Environment, 2nd edition (Thomson) Radiography of Cultural Materials, 2nd edition (Lang, Middleton) Tapestry Conservation: Principles and Practice (Lennard, Hayward) The Textile Conservator’s Manual, 2nd edition (Landi) Upholstery Conservation: Principles and Practice (Gill, Eastop) A Practical Guide to Costume Mounting (Flecker) Contemporary Theory of Conservation (Muñoz-Viñas) Digital Collections (Keene) Digital Heritage: Applying Digital Imaging to Cultural Heritage (MacDonald) Fragments of the World: Uses of Museum Collections (Keene) Historic Floors (Fawcett) Managing Conservation in Museums (Keene) Materials for Conservation (Horie) The National Trust Manual of Housekeeping Natural Materials: Sources, Properties and Uses (DeMouthe) Organic Chemistry of Museum Objects (Mills, White) Pigment Compendium: Dictionary (Eastaugh, Walsh, Siddall, Chaplin) Pigment Compendium: Optical Microscopy (Eastaugh, Walsh, Siddall, Chaplin) Pigment Compendium CD (Eastaugh, Walsh, Siddall, Chaplin) Restoration of Motion Picture Film (Read, Meyer) Risk Assessment for Object Conservation (Ashley-Smith) Structural Aspects of Building Conservation (Beckman, Bowles)

Published titles:

Related titles:

X-RADIOGRAPHY OF TEXTILES, DRESS AND RELATED OBJECTS

Sonia O’Connor

●

Mary M. Brooks

AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD PARIS • SAN DIEGO • SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Butterworth-Heinemann is an imprint of Elsevier

Butterworth-Heinemann is an imprint of Elsevier Linacre House, Jordan Hill, Oxford OX2 8DP, UK 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA First edition 2007 Copyright © 2007 Elsevier Ltd. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (44) (0) 1865 843830; fax (44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloguing in Publication Data A catalogue record for this book is available from the Library of Congress ISBN: 978-0-7506-6632-9

For information on all Butterworth-Heinemann publications visit our web site at http://books.elsevier.com

Typeset by Charon Tec Ltd (A Macmillan Company), Chennai, India www.charontec.com Printed and bound in Italy 07 08 09 10 11

10 9 8 7 6 5 4 3 2 1

This book is dedicated to Ed Newton MInstNDT MIAQP In grateful recognition of his constructive criticism and support.

This page intentionally left blank

Contents

Preface Acknowledgements Contributors

xiii xv xvii

Part 1: Textile X-radiography

1

1. Introduction Sonia O’Connor and Mary M. Brooks

3

The catalyst for this book The history of textile radiography Radiography at the Textile Conservation Centre Imaging textiles using mammography Developing practice Notes References 2. Principles of X-radiography Sonia O’Connor Introduction X-rays and the electromagnetic spectrum Properties of X-rays Production of X-rays Conventional radiography Optimising image quality Viewing film radiographs Working with film radiographs Storage and archive of radiographic films Summary Note References 3. High definition X-radiography of textiles: methods and approaches Sonia O’Connor Introduction Why textiles seem difficult to X-ray Low energy high definition radiography

3 5 7 8 10 10 11 12 12 12 13 13 15 16 21 21 22 22 22 22 23 23 23 24

Choosing X-ray equipment and facilities Practical approaches to textile radiography Determining correct exposure parameters Thin homogeneous textiles Layered and more complex textile objects Thicker textiles Heterogeneous textiles Mixed-media objects X-raying ‘special needs’ textiles Special radiographic techniques Notes References

25 30 35 39 39 40 40 41 43 50 56 56

4. Textile X-radiography and digital imaging Sonia O’Connor and Jason Maher

58

Introduction Digital versus analogue Components of a digital image Storage of digital images Digital image capture Direct and computed radiography CR and textile radiography Digital image processing Summary Notes References

58 58 59 62 64 67 68 69 71 73 73

5. Image interpretation Sonia O’Connor Introduction Negative images Interpretation basics Characteristic images Effect of exposure on image interpretation Image artefacts References

74 74 74 76 78 81 88 90 vii

viii

Contents

6. Assessing the risks of X-radiography to textiles Sonia O’Connor, with a contribution on DNA by Jason Maher Introduction Colour Dating DNA Organic textile fibres and dyes X-ray analysis Testing radiographed silk samples Summary Notes References 7. Radiation safety Graham Hart Introduction Myths and legends Justification, optimisation and limitation United Kingdom Ionising Radiations Regulations Radiation and risk Putting risk in perspective Practical radiation protection Radiation monitoring Conclusion Acronyms References

91

91 91 91 92 92 93 94 94 94 94 96 96 96 96 97 97 100 100 101 103 103 103

Part 2: Exploring the X-radiographic features of textile objects 105 Sonia O’Connor and Mary M. Brooks 8. Materials Fibres Fillings Supports, stays and substructures Miscellaneous materials Note References 9. Threads, fabrics and construction techniques Yarns and threads Cords and plaits Woven textile structures Non-woven structures Construction techniques Other construction methods and materials

107 107 109 119 123 125 125

126 126 126 126 136 140 145

Notes References 10. Surface decoration Painted and printed textiles Underdrawing Appliqué and embroidery Metal threads Unusual materials used for surface decoration Note References

149 149 150 150 153 153 154 156 159 159

11. Makers and making, degradation and repair Makers and making Degradation Use and wear Reuse, repair and conservation Summary Notes References

163 163 164 166 168 171 171 172

Part 3: Case studies

173

Introduction Mary M. Brooks and Sonia O’Connor 12. Evaluating X-radiography as a tool for examining upholstered furniture Kathryn Gill Introduction Practical challenges to the radiography of historic upholstered seat furniture Radiography for documentation: case study of an eighteenth century upholstered chair Radiography as a complement to photographic evidence: case study of the Seehof Suite Interpretation of the X-ray images What is not revealed by radiography: case study of the Audley End settee Investigation of a portable medical facility for object examination Conclusion Acknowledgements Notes Acronyms References

175 175 175 176 176 178 180 182 183 183 183 184 184

Contents ix

13. The use of X-radiography in the Textile Conservation Laboratory, Opificio delle Pietre Dure, Florence: methodological, technical and research approaches towards a non-invasive investigative technique 185 Susanna Conti and Alfredo Aldrovandi Introduction: concepts and issues 185 Selecting appropriate analytical approaches 186 Radiography applied to textiles: technical issues 186 Radiography of large textiles 186 The use of radiography at the Opificio delle Pietre Dure 187 Pilot study of the use of radiography in textile conservation: case study of a chasuble 188 Case study: a Chinese screen 190 Case study: a fifteenth century velvet fragment from a nineteenth century collection 193 Case study: a mitre 194 Case study: dressed statue 196 Case study: wax sculpture Dormitio Virgini (‘The Death of the Virgin’) 199 Discussion 201 Acknowledgements 201 Notes 201 References 201 14. The role of X-radiography in the documentation and investigation of an eighteenth century multi-layered stomacher Gabriella Barbieri Introduction The Nether Wallop cache The practice of concealment: a contextual framework The stomacher Rationale for research General aims of project Specific objectives of X-ray examination Methodology Interpretation of the radiographic images Materials and construction Patterns of use Patterns of degradation Conclusion Acknowledgements Notes

203 203 203 203 204 205 206 206 206 206 207 208 209 209 210 210

Acronyms References 15. Hidden Structures: the use of X-radiography in the Fashion Gallery at Snibston Discovery Park, Leicestershire Clare Bowyer Introduction The Fashion Gallery, Snibston Discovery Park Hidden Structures Selecting objects for radiographic display images Chosen objects and radiographs Feedback and evaluation

210 210

212 212 212 212 213 214 216

16. X-radiography of a knitted silk stocking with metal thread embroidery Sonia O’Connor, Mary M. Brooks and Josie Sheppard

217

Introduction The stocking Condition Radiography What the radiography revealed Summary Acknowledgement Note Reference

217 217 218 219 221 224 224 224 224

17. A chalice veil rediscovered Sonia O’Connor and Mary M. Brooks

225

Introduction Description Condition Evidence from radiography Conclusion Acknowledgement Notes References

225 225 226 226 230 230 230 230

18. The use of X-radiography in the analysis and conservation documentation of a set of seventeenth century hanging wall pockets 231 Mary M. Brooks and Sonia O’Connor Introduction The hanging wall pockets

231 231

x

Contents

Materials and construction Condition before treatment Radiography Information obtained from the radiography Embroidery techniques Damage and repair Conclusion Acknowledgements Note Reference

231 231 231 233 235 235 236 236 236 236

19. ‘In needle works there doth great knowledge rest’: the contribution of X-radiography to the understanding of seventeenth century English embroideries from the Ashmolean Museum of Art and Archaeology, Oxford 237 Mary M. Brooks and Sonia O’Connor Introduction Radiography techniques The contribution of radiography to understanding materials, condition and construction Summary Acknowledgements Notes References 20. X-radiography of dolls and toys Mary M. Brooks, Sonia O’Connor and Josie Sheppard Introduction Materials and manufacture of European dolls: a brief overview The value of radiography for curation and conservation Summary of radiography methods Information from radiography Summary Acknowledgements Notes References 21. X-radiography of teddy bears and other textile artefacts at the Victoria & Albert Museum Marion Kite Introduction Early radiography of textiles and dolls at the Victoria & Albert Museum Radiography of teddy bears

237 237 238 247 247 247 247 249

249 249 256 258 261 265 265 265 265

266 266 266 266

Taking and interpreting radiographic images Stuffings, squeakers and structures Construction and assembly methods Threads and fabrics Conclusion Acknowledgements Notes References

22. X-radiography of patchwork and quilts

268 268 269 272 272 272 272 272

273

Mary M. Brooks, Sonia O’Connor and Josie Sheppard Introduction Quilting and patchwork: a brief overview The value of radiography for curation and conservation Special requirements for radiography of quilts and coverlets Information from radiography Benefits of radiography Acknowledgements Note References

273 274 274 275 275 284 287 287 287

23. Revealing the layers: The X-radiography of eighteenth century shoes at Hampshire County Council Museums and Archives Service 288 Sarah Howard and Robert Holmes Introduction Radiography at HCCMAS Films and processing Selection of shoes for radiography Construction of heels Stitching General construction Conclusion Notes References 24. The contribution of X-radiography to the conservation and study of textile/leather composite archaeological footwear recovered from the Norwegian Arctic Elizabeth E. Peacock Introduction History of Russian Pomor hunting activities on Svalbard

288 288 289 289 289 290 293 293 293 293

294 294 294

Contents xi

The burial environment on West Spitsbergen, Svalbard 295 The Russekeila site 296 The artefacts and their recovery 296 The footwear recovered at Russekeila 296 Radiography of the footwear 296 Results and implications 298 Conservation strategy and implementation 299 Russian Pomor textile/leather composite archaeological footwear revisited 299 Conclusion 300 Notes 301 References 301 25. Controlled lifting and X-radiography of gold threads from ancient archaeological textiles 302 Elizabeth Barham Introduction The Spitalfields Roman sarcophagus textile finds The Prittlewell Anglo-Saxon chamber-grave textile finds Conclusions Acknowledgements Notes Reference 26. X-radiography of ethnographic objects at the Horniman Museum Louise Bacon Introduction Radiography equipment and methods used at the Horniman Museum Textile elements in ethnographic artefacts The conservation benefits of radiography for ethnographic artefacts with textile components: two case studies The ethics of radiography Conclusion Acknowledgements Notes References 27. The use of X-radiography in the conservation treatment and reinterpretation of an incomplete musette Sylvie François Introduction Musette

302 302 304 306 306 306 306 307 307 307 308 308 311 311 312 312 312

The Horniman Museum musette Treatment proposal and the role of radiography Radiographic procedures Interpreting the radiographs Impact of radiography on the treatment and interpretation of the musette Acknowledgements Notes References 28. X-radiographic examination of a historic mannequin on display in Edinburgh Castle, Scotland David Starley and Fiona Cahill Introduction History of the mannequins Description Background to the projects Purposes of radiography Radiographic procedure and equipment Interpretation of the radiographs Evidence for dating Summary Acknowledgements Note References 29. X-radiography of Rembrandt’s paintings on canvas Ernst van de Wetering Précis by Mary M. Brooks and Sonia O’Connor

314 314

315 315 316 318 318 318 318

319 319 319 320 320 320 320 322 323 323 324 324 324 325

Editors’ note 325 Rembrandt’s oil paintings on canvas 325 Objectives of research into canvas supports 325 Radiographs as a means of studying canvas 325 Research methods and results 326 Characteristics of canvases by, or attributed to, Rembrandt 327 Conclusion 327 Acknowledgements 327 References 328 Index

314

314

329

This page intentionally left blank

Preface Mary M. Brooks and Sonia O’Connor

This book presents results from a research project exploring the potential of X-radiography as a tool for the characterisation, condition assessment and study of ancient, historic and contemporary textiles and hence to aid conservation decision making and curatorial studies. The project formed part of the research supported by the Arts and Humanities Research Council (AHRC)1 Research Centre for Textile Conservation and Textile Studies. The Research Centre was established in July 2002 by the Textile Conservation Centre (TCC), University of Southampton, in conjunction with the Department of Archaeological Sciences, University of Bradford, and School of Art History and Archaeology and the Whitworth Art Gallery, University of Manchester. The Research Centre’s goal is to improve the conservation and interpretation of historic textiles by enhancing knowledge and understanding. To this end, research is structured into four themes: Textile Materials, Modern Materials, Textiles and Text and Worldly Goods. This radiography project is part of the Textile Materials theme and is a collaborative project between the University of Southampton and the University of Bradford. The first questions that are asked about radiography of textiles are: ● ● ●

Is it safe for me and safe for the artefact? Will it show me anything I would want to know? Can I get it done easily?

This book addresses these questions and presents conservators, curators and others interested in understanding textile artefacts and their histories with a comprehensive view of the role radiography has to play in the study and conservation of textiles.

We have deliberately taken a broad view of textiles and dress – ranging from single layered fragments to complex three-dimensional mixed-media artefacts. The first part of this book covers the principles of radiography and explores the techniques that are best suited to the taking of high quality images of textile artefacts at no risk to the object or personnel. How the resulting radiographs can be best examined and interpreted is explained and developing technologies, which increase the range and depth of information obtainable through radiography, are also introduced. The aim of the second part has been to demonstrate the extraordinary contribution that radiographic investigation can make to our understanding of textiles and to encourage further research and exploration. The concluding case histories in the third part, written by colleagues with experience in the radiography of textiles, demonstrate the added value that radiography can provide in very diverse circumstances. We hope this book will be the catalyst for the wider use and exploration of radiography in relation to the study and conservation of textiles and that the examples and case studies presented will excite and inspire conservators to feel confident both to undertake radiography of textiles themselves and those unfamiliar with radiography to brief others on appropriate techniques when commissioning textile radiographs. Researching the systematic application of the established technique of radiography to textiles has been a fascinating journey. Our goal throughout has been to communicate what we have learnt to the wider conservation and textile communities through presentations, publications and teaching. Many people have generously shared their experiences, given us guidance

Note 1.

On 1 April 2005, the Arts & Humanities Research Board (AHRB) became the Arts & Humanities Research Council (AHRC). xiii

xiv

Preface

and suggested avenues for further exploration. It has been our privilege to draw together work from practitioners in this area and we are very grateful both to those who made us aware of such work and also to our colleagues for participating in the book and sharing their knowledge. We are grateful to all those individuals and institutions who have allowed us to

photograph and radiograph objects from their collections and given permission for their reproduction. Unless otherwise specified, objects are from private collections. We thank all the case studies authors for their input and acknowledge that any errors or omissions are our own.

Acknowledgements

We also thank colleagues in the AHRC Research Centre for Textile Conservation and Textile Studies especially Dr Maria Hayward, Director, and Dinah Eastop, Associate Director, Dr Paul Wyeth (all of the University of Southampton) and Dr Carl Heron (University of Bradford) for their unstinting support. Chris Bennett provided invaluable administrative support. Nell Hoare MBE, Director of the Textile Conservation Centre, gave both encouragement and much appreciated practical help. Thanks are also due to Professor Mark Pollard, now of Oxford University. We owe particular thanks to Dr Maria Hayward and Dr Paul Wyeth who undertook a peer review of the whole book and gave perceptive and constructive feedback and to Ed Newton, for his review of the technical material relating to radiographic theory and practice. Our special thanks go to Jason Maher for his huge contribution to the illustrative material for this book and for sharing his expertise in digital imaging. We extend our thanks to colleagues and friends for their help in tracking down information, allowing us to radiograph objects in both personal and institutional collections and giving us other valuable help: Mary Ballard, Smithsonian Institution, Washington Peter Baucham, Argos Inspection, Washington, Tyne & Wear Tom Bilson, Courtauld Institute of Art, London, for information about the conservation of two dolls M. Annette Brooks for loaning pearls Phyllis Brown, Sheffield, for loaning her lace collection Jo Buckberry, University of Bradford Vivienne Chapman, Conservation Centre, National Museums Liverpool Andy Chopping, Photography, Museum of London Archaeological Service for comparison of methods for digitising X-ray images

Nicola Christie, Paintings Conservation, Conservation Centre, National Museums Liverpool for access to X-ray equipment and information Mathew Collins, BioArch, University of York, York, for information relating to damage of proteins and DNA David Crombie, Paintings Conservation, Conservation Centre, National Museums Liverpool, for technical information Dr Ian Croudace, National Oceanography Centre, University of Southampton, for access to radiography facilities Joyce Dawson, PhD student, Textile Conservation Centre, University of Southampton Dinah Eastop, for information about and access to artefacts from the AHRC Deliberately Concealed Garments Project Vanessa Fell, English Heritage Elia Flores, Radiology Programme, Blinn College, Texas, for generously sharing information on her use of mammography facilities Chrissy Freeth, University of Bradford Dr Paul Garside, Textile Conservation Centre, University of Southampton, for undertaking tests to evaluate the effect of X-rays on silk Mike Gentry, Johnson Space Centre, Houston Ruth Gilbert, PhD student, Textile Conservation Centre, University of Southampton Lynn Grant, textile conservator Dr Sonia Guillen, Director of the Centro Malliqui, Bioanthropology Foundation of Peru Louise Hampson, York Minster Archives Allan Hall, Department of Archaeology, University of York, for identification of plant remains Michael Halliwell, Textile Conservation Centre, University of Southampton, for providing photographs from the TCC image archive Liesbeth M. Helmus, Centraal Museum, Utretcht

xv

xvi

Acknowledgements

Robert Janaway, Department of Archaeological Sciences, University of Bradford, for access to the University of Bradford’s teaching collection Allan Janus, Archives Reference Desk, National Air and Space Museum, Smithsonian Institution, Washington Jan Jansen, Experimental Zoology Group, University of Wageningen Andrew Jones, University of Bradford, for identification of fish scales Julie Jones, then of York Archaeological Trust, for information about radiographs Susan La Niece, British Museum, London Janet Lang for her support in this project Alison Lister, textile conservator, and Colleen Lister for loaning an ethnographic textile Ross McEwing, Wildlife DNA Services Ltd, Bangor, for information relating to damage of proteins and DNA Phil Morris, GE Inspection Technologies, Coventry, for technical advice and access to CR and DR Richard Mould for valuable contacts for radiography Mark Norman, Head of Conservation, The Ashmolean Museum of Art and Archaeology, Oxford Professor Terry O’Connor, Department of Archaeology, University of York, for information relating to damage of proteins and DNA Keith Oliver, Paper Conservation, Conservation Centre, National Museums Liverpool, for information relating to possible risk of damage to paper through radiography Rosemary Payne for loaning an ethnographic textile Alan Phenix, University of Northumbria, now at the Getty Conservation Institute, Los Angeles, for radiography and introducing us to Professor van de Wetering and his work on the radiography of painted canvases Eve Reay, Argos Inspection, Washington, Tyne & Wear Amy Ross, Johnson Space Center, National Aeronautics Space Administration, Houston Cordelia Rogerson, Textile Conservation Centre, University of Southampton, for access to samples and sharing thoughts on painted textiles

Amber Rowe, Head of Conservation Services, Textile Conservation Centre, University of Southampton, for suggesting textiles for radiography Vuka Roussakis, American Museum of Natural History, New York Hanna Szczepanowska, Smithsonian National Air and Space Museum, Washington Josie Sheppard, York Castle Museum, York Museums Trust, for generously allowing access to many fascinating artefacts Peter Sinclair, GE Inspection Technologies, Coventry, for technical advice and access to CR and DR James Spriggs, then of York Archaeological Trust, for loan of a chalice veil Rebecca Smith, textile conservator, for sharing information about Tog Susan Stanton, Conservation, The Ashmolean Museum of Art and Archaeology, Oxford Dr Arnulf von Ulmann, Germanisches Nationalmuseum, Nürnberg Penelope Walton Rogers, The Anglo-Saxon Laboratory, York Jacqui Watson, English Heritage, Fort Cumberland Howard Wellman, MAC Laboratory, Maryland Dr Catherine Whistler, Senior Assistant Keeper, Department of Western Art, The Ashmolean Museum of Art and Archaeology, Oxford Jane Williams Butt, Radiography, Bradford Teaching Hospitals, for access to CT scanners Amanda Young, Early Manned Spaceflight/ Astronaut Equipment, Division of Space History, National Air and Space Museum, Smithsonian Institution, Washington Thanks are also due to Sarah Vanstone, Debbie Clark, Marion Stockton and Stephani Allison of Elsevier for their input and help. Finally, we would like to thank our families and friends for their patient love and tolerance during the writing of this book. Mary M. Brooks and Sonia O’Connor

Contributors

Alfredo Aldrovandi obtained his degree in Physics at the University of Modena, Italy. He has worked at the Opificio delle Pietre Dure e Laboratori di Restauro (OPD) in Florence since 1983. As Head of the Physics Laboratory at the OPD, his research focuses on the development and application of diagnostic, non-invasive analysis for works of art. He also teaches physics on the OPD’s four-year conservation training course, the Scuola di Alta Formazione (SAF). Among his many activities, he collaborates with several institutions in the field of Fine Arts Conservation, taking part in research projects, conferences and courses. Louise Bacon has been Head of Collections, Conservation and Care at the Horniman Museum in South East London since 1986. An objects conservator trained at the Institute of Archaeology (University College London), she has been an Associate Member of the Museums Association since 1984 and completed her doctorate in Archaeometallurgy in 2003. She has more than 30 years of experience of museums and heritage organisations both in Britain and abroad where radiography has always played an important role in her work. Gabriella Barbieri originally trained as a linguist, gaining a BA (Hons) in French and Italian from Bristol University in 1989, and a diploma in Technical and Specialised Translation from the University of Westminster in 1990. She subsequently followed a career in technical and legal translation with the UK’s largest independent translation company. In 2000 Gabriella decided to pursue her life-long interests in textiles and costume and retrained at the Textile Conservation Centre, University of Southampton, and was awarded an MA in Textile Conservation. After graduation, Gabriella worked on a number of projects at the Textile Conservation Centre before taking up a position at a private conservation studio in London which specialised in the conservation of early Middle Eastern textiles and carpets. Since February

2005, she has been working at the Bowes Museum, County Durham, England, where she is involved primarily in the conservation and display of the Bowes Founders’ Collection, which includes a unique collection of embroidered and tapestrywoven seat covers, 16th and 17th century European lace and 15th to 18th century ecclesiastical textiles. Liz Barham has worked as a conservator for Museum of London Specialist Services since 1999, where the conservation team provides contract conservation services to archaeological units on fresh archaeological finds from sites across the UK, as well as practical conservation services and preventive conservation advice on wide-ranging historic and archaeological artefact collections for museums, archaeological archives, private institutions and individuals. She has an MA in the Conservation of Historic Objects (Archaeology) from the University of Durham and an Honours degree in Classical Civilisation from the University of Warwick. Clare Bowyer was a Project Assistant for the Fashion Gallery at Snibston Discovery Park and is now at Kettering Museum. Prior to this, she worked as an exhibition researcher at Charnwood Museum, Loughborough and as cur-ator for the costume web pages at the Gallery of Costume, Manchester. As well as volunteering in museums and galleries, she has worked for the Community Fund, English Heritage and Liberty, London. She read History of Fine and Decorative Arts at the University of Leeds before gaining an MA in the History of Dress at the Courtauld Institute, London. Mary M. Brooks MA, DMS, DipTexCons, FIIC, ACR, HEAM took her BA at Cambridge University. After working in the book world, she moved into management consultancy and then developed her long-term interest in textiles by taking the Postgraduate Diploma in Textile Conservation. She has since worked as a curator and conservator in the USA and Europe, including xvii

xviii

Contributors

developing an award-winning exhibition on conservation. She is now Reader at the Textile Conservation Centre, University of Southampton and has a special interest in the contribution that object-based research and conservation approaches can make to the wider interpretation and presentation of cultural artefacts. Fiona Cahill graduated from De Montfort University in 2002 with a degree in Restoration and Conservation. Since then she has worked for an architectural conservation company and was a project conservator for the National Trust in Liverpool. She joined the Royal Armouries as a conservator in 2004. Susanna Conti received her art diplomas in Painting and in Weaving in 1972 and 1974 respectively. In 1975 she qualified as an art teacher in the secondary school where she taught for one year. Subsequently, she trained in tapestry conservation and obtained her diploma in Textile Conservation from the prestigious Opificio delle Pietre Dure e Laboratori di Restauro (OPD) in Florence in 1980. She won her position as Capo Tecnico of the Textile Conservation Laboratory at OPD as a result of open national competition in 1981; in 2006 she was appointed Textile Conservation Director. During her 25 years at the OPD, she has carried out conservation treatments on a wide variety of projects focused on textile and textile-related artefacts and also conducted their four-year textile conservation training course, the Scuola di Alta Formazione (SAF). She collaborates with state institutions and museums on research projects, displays, publications, conferences and training courses, both in Italy and abroad. Sylvie François holds the Postgraduate Diploma in Textile Conservation from the Textile Conservation Centre (Courtauld Institute of Art, University of London). Her previous studies were in Canada where she received a Bachelor of Fine Arts (Major in Art History and Studio Arts) from Concordia University. Her early interest in textiles and costumes led her to pursue technical studies sanctioned by a Professional Collegial Diploma in Fashion from Cegep Marie-Victorin. She is currently Conservation Officer at Cirque du Soleil, an international performing arts company that has its headquarters in Montreal, Canada. In 1999, she developed their Heritage Costume Collection and, since 2000, manages this collection and the art collections. In this role, she oversees the development

and application of conservation policies and strategies. Kathryn Gill gained a BA (Hons) in Textiles and Fashion at Manchester Polytechnic, after which she trained and worked as a textile conservator, specialising in upholstery conservation at the Textile Conservation Centre, Hampton Court Palace. In 1984 Kate moved to the USA to set up the upholstery conservation section at the Metropolitan Museum of Art. After seven years as Senior Upholstery Conservator, she took a post at the Textile Conservation Centre, University of Southampton. Kate is Senior Conservator and Lecturer, combining practical conservation (textiles and upholstered furniture) with teaching and research. She is principal contributor to and co-editor (with Dinah Eastop) of Upholstery Conservation: Principles and Practice (2001). Graham Hart BSc, MSc, MIPEM, MSRP worked as a medical physicist within the National Health Service for 30 years and has for the last 15 years worked as a radiation/laser/non-ionising radiation protection adviser at the University of Bradford in the health, research, educational and veterinary sectors. Currently, Graham is a member of the Association of University Radiation Protection Officers’ Technical Coordinating Committee, the Society for Radiological Protection’s Non-Ionising Radiation Topic Group and an assessor for RPA 2000, an assessing body for Radiation and Laser Protection Advisers in the UK. He is also an independent radiation protection consultant. Robert Holmes BA (Hons) is Senior Conservator of Antiquities and Fine Metalwork for the Hampshire County Council Museums and Archive Service. He has worked for the service since 1983 and has recently taken on the Keepership of the Firearms Collection. He joined the Metals Section of the Conservation Department of the British Museum in 1978, becoming a specialist in the restoration of Roman and Anglo-Saxon metalwork. Prior to joining the British Museum he took his degree in Silversmithing at Birmingham Polytechnic. Sarah Howard BA (Hons), DipTexCons, ACR graduated from the Textile Conservation Centre in 1992 after completing the Postgraduate Diploma in Textile Conservation. After completion, she undertook a number of short-term contracts with freelance conservators and institutions such as the Victoria and Albert Museum. She joined

Contributors xix

Hampshire County Council Museums and Archives Service in 1996 and is now Principal Conservator and Senior Textile Conservator working with their historic dress and textiles collection. Marion Kite is Head of Furniture, Textiles and Frames Conservation at the Victoria and Albert Museum where she has worked since 1974. Her specialist discipline is textile conservation. She has published and lectured widely on many aspects of textile conservation and the conservation of organic materials associated with textile objects. During the past 20 years she has developed a particular interest in the conservation of animal products and unusual materials incorporated into textiles and dress accessories. She is co-editor with Roy Thomson of Conservation of Leather and Related Materials, published by Elsevier in 2005. Marion served on the Directory Board of the International Council of Museums Committee for Conservation between 1993 and 1999 and as Treasurer between 1993 and 1996. She is a Fellow of the International Institute for Conservation and currently serves on the IIC Council. She is Chairman of the Executive Council of the Leather Conservation Centre and also sits on the Council of the Museum of Leathercraft. She is a Fellow of the Royal Society of Arts and a Trustee of the Spence and Harborough Collections of Gloves administered by the Worshipful Company of Glovers of London. Jason Maher has a degree in Archaeological Sciences from the University of Bradford and an MSc in Osteology, Palaeopathology and Funerary Archaeology from the Universities of Bradford and Sheffield. He was appointed Osteoarchaeological Technician in 1996 in the Department of Archaeological Sciences, University of Bradford. In 2000 he left the university to become an IT consultant and freelance trainer. Since then he has combined his interests in anthropology, computing and radiography particularly through the application of digital imaging solutions to archaeological problems. Sonia O’Connor Dip Cons, FIIC, ACR trained as an archaeological conservator at the Institute of Archaeology, University of London, passing with distinction, and has worked in archaeological conservation at the National Maritime Museum, Greenwich, London; University College, Cardiff; and the York Archaeological Trust. In 1995, she joined the Department of Archaeological Sciences, University of Bradford. She now holds the post of Research Fellow in Conservation. Her research

forms part of the activities of the AHRC Research Centre for Textile Conservation and Textile Studies based at the University of Southampton in partnership with the University of Bradford. Her areas of expertise include the radiography of cultural material and she received the 2002 Nemet Award of the British Institute of Non-Destructive Testing for her work in this field. Elizabeth E. Peacock is Professor in Conservation and Senior Research Conservator on the staff of Vitenskapsmuseum at the Norwegian University of Science and Technology (NTNU) in Trondheim, Norway. She holds a BA in Mathematics, an MAC in Art Conservation, and a PhD in Textile Science. During her conservation training she interned with a number of institutions in the USA and Europe. She is a Fellow and Council Member of the International Institute for Conservation. Elizabeth’s main interests are: the interface between archaeological science, conservation science and conservation; the interaction of organic materials with the buried environment; and conservation education without borders. She has wide experience with organic archaeological materials, especially textiles and leather. Primarily this experience is with materials recovered from wet sites, but earlier in her conservation career she worked with finds from sites in the Eastern Mediterranean and Egypt. This extensive experience includes laboratory-based and field conservation, experimental and object-based research, and teaching and publishing. Josie Sheppard has worked with costume and textile collections in museums for 25 years, specialising in fashionable and everyday dress of the last two centuries. Since 1988 she has been Curator of Costume and Textiles at York Castle Museum, where her remit includes the extensive collection of dolls. Previous posts have included working at the Warwickshire Museum, and at Worthing Museum and Art Gallery, where she was Assistant Curator of Costume. David Starley initially trained as a metallurgist before studying Archaeological Sciences and undertaking a PhD studying steel in medieval armour (both at Bradford University). He worked in English Heritage’s Ancient Monuments Laboratory before moving to the Royal Armouries, Leeds, in 1999, as Science Officer.

This page intentionally left blank

Part 1 Textile X-radiography

This page intentionally left blank

1 Introduction Sonia O’Connor and Mary M. Brooks

Radiography is a non-invasive and non-destructive technique that can reveal information hidden from visual examination. Sometimes radiography shows information that could not have been disclosed in any other way or only obtainable through destructive investigation or sampling. On other occasions, without the confusion of colours, textures or shadow, it may provide a clearer image of features that might not otherwise have been spotted. In either event, radiography can greatly enhance our understanding of cultural material. In generating new information or supplying corroborative evidence for already known or suspected features, it enables better quality questions to be asked and so generates new lines of research. Radiography is also a direct and visually appealing way of communicating information about objects to a wider public. There is a long and well-established tradition of using radiography for the assessment, investigations and documentation of cultural material such as sculptures, decorative metalwork, ceramics, paintings, paper, archaeological metalwork, human and animal remains, and using it to answer research questions or inform conservation decision making (Lang and Middleton, 2005). However, its use in the investigation of textile objects, until recently, has only infrequently occurred and the results have rarely been published. Radiography has been used to examine upholstered furniture (see Gill, Chapter 12) although this pioneering work tended to focus on understanding substructures rather than examining textile elements in depth. It has most frequently been used to study metal threads, but not the textiles that support them. As it was not normally perceived as a useful tool for investigating textiles and dress, the principles of radiography were not usually taught to textile conservators and appropriate X-ray facilities are rarely available in textile conservation laboratories. The purpose of this book is to change this state of affairs.

The catalyst for this book In 1988, the conservation of an unusual hybrid object brought together two conservators from different disciplines in a collaboration that was to prove the catalyst for this book (Figure 1.1a). Mary Brooks was then Assistant Keeper of Textiles (Conservation) at York Castle Museum and Sonia O’Connor was Deputy Head of Conservation at York Archaeological Trust, both based in the historic city of York, England. The object in question was an 1860s women’s shoe, with a decorative tongue, which was to go on display. The addition of a group of small dolls had turned the shoe into a toy representing the English nursery rhyme ‘There was an old woman who lived in a shoe, who had so many children she didn’t know what to do’.1 The mother doll sits in the throat of the shoe, holding one of the smallest dolls in her arms. This seems to have been a home-made toy, reusing a worn shoe as a plaything, although similar representations of the rhyme, some of them based on card ‘shoes’, were made during the American War of Independence and sold to raise funds. The condition of the shoe was a cause of some concern. The silk covering was in a fairly degraded condition, the jaunty flag was weak and many of the dolls had been lost, their presence only indicated by the stains of the adhesive which had once held them in place. As a textile conservator, Mary was most concerned about the rust formed around the hands of the mother doll and staining the clothes of the baby in her arms. It was necessary to establish the cause and extent of this degradation (Figure 1.1b). Radiography seemed an ideal technique to gain information about the substructure and its condition without interventive, and possibly damaging, examination. Being unfamiliar with radiographic techniques, Mary asked Sonia, an experienced radiographer of archaeological materials, 3

4

Textile X-radiography

(a)

(c)

(b)

Figure 1.1 ‘The Old Woman in the Shoe’ after conservation, (a) photograph, (b) photograph detail; note the staining due to corrosion of the metal armature, (c) radiograph. (YORCM: BA2573; © Sonia O’Connor, University of Bradford; reproduced by permission of York Castle Museum, York Museums Trust.)

to radiograph the toy to investigate the position, extent and condition of any metal within the doll so that she could formulate an appropriate conservation strategy. The resulting radiographs were a revelation (Figure 1.1c). Not only was the original purpose of the radiography fulfilled but the images also provided detailed information about the non-metal materials of the toy. An entirely unexpected organic component was revealed – a chicken furcula (wishbone), forming the body and legs of the mother doll. The mother doll’s head and hands appear to be sculpted from very radio-opaque material, possibly pigmented lead putty. The wooden joints and paint on the baby dolls’ heads could be seen. Unexpectedly, more ephemeral materials were also visible in the radiograph, including the textiles forming the dolls’ clothes and the silk covering of the shoe. Re-covered edges and previously hidden stitched repairs to the shoe itself became evident. The source of the rust was an armature made from two iron wires which passed from each hand of the mother doll and around the top of the tightly fanfolded paper which formed a supporting ‘petticoat’

for her skirt and were joined at the front. The radiograph enabled greater understanding of the biography of the shoe and how it had been made, revealing unexpected materials as well as previous alterations. It informed the development of the treatment and storage strategy while reducing the need for interventive treatments. The results of this radiography led to further collaborations into the potential of radiography. The unusually bright image of a repair thread on one side of the shoe towards the heel triggered a new line of research into weighted silks funded by the British Academy (Brooks et al., 1996). Further radiographic investigation of historic toys, dolls and textiles was presented as a poster at the International Council of Museums (ICOM) Conservation Committee Triennial Meeting in Dresden in 1990. The results of these projects convinced us of the benefits of conducting a systematic investigation of the potential of radiography to aid conservators and curators in the understanding of textile objects and in conservation decision making (Brooks and O’Connor, 2005; O’Connor and Brooks, 2005). This book is the culmination of that work.

Introduction 5

The history of textile radiography As work on this book proceeded, we became aware of a body of knowledge and practice in the radiography of textiles that was not widely known. It became clear that there was a ‘hidden history’ of the radiography of textiles. Within months of Wilhelm Röntgen’s discovery of X-rays in November 1895, its demonstration in lecture halls and exhibitions had become a public entertainment frequently involving the imaging of hands or the contents of bags. Many of the earliest surviving radiographs are of feet in shoes or boots. Those growing up in the 1950s and 1960s might remember fluoroscopic X-ray machines in shoe shops, used to check whether new shoes fitted children’s feet. Common fixtures in shoe shops in North America since the 1930s these continued in use, in the UK and Canada at least, through the 1960s despite mounting safety concerns (Duffin and Hayter, 2000). Mould’s instructive history of radiography includes a full body radiograph of a dressed female figure, first published in volume II, number 1 of the Archives of the Roentgen Ray, July 1897 (Mould, 1993: 38, Figure 4.2). Her jewellery is evident but little detail of her clothes is captured except for her hat pin, the metal pins of the heels and the buttons of her high buttoned boots and, perhaps, the shadow of a whalebone corset. The application of radiography for security was also quickly realised. Even before the end of the nineteenth century radiography was being used by customs officials in the fight against terrorism and smuggling. Mould reproduces radiographs of an ‘explosive book’, taken in 1896, and a Polish soldier’s boots with gold coins concealed in the soles, taken in 1920 (Mould, 1993: 106, Figures 17.8 and 17.11). Perhaps one of the first, and possibly one of the most colourful, people to experiment with the radiography of textiles of cultural significance was Junius Bird, Curator of South American Archaeology at the American Museum of Natural History, New York, USA, from 1931 onwards. Bird had developed a special interest in pre-Columbian textiles and, by the 1960s, was looking into the possibility of producing ‘auto-radiographs’ of archaeological textiles following neutron activation.2 It has not been possible to establish whether this rather alarming procedure was ever put into practice. However, correspondence survives from 1964 between Bird and Charles F. Bridgman of X-ray Technical Services, Eastman Kodak Company, Rochester, New York, who had expertise in the use of radiography for cultural material. These letters make

it clear that Bird had experimented with wetting textiles using a barium solution as a radiographic contrast agent to improve the visibility of textile features ‘to determine structure in pieces too compact and delicate to stand any physical prying into their secrets’.3 Bridgman emerges as the unsung hero of textile radiography. He started using radiographs to examine textiles in the 1950s and contacted Bird to source ancient textiles to enable him to undertake this research. In his pioneering article, he notes, ‘there is almost no literature dealing with the radiography of textiles and yet it is a technique that is most valuable in gaining knowledge of weaving methods’ (1973: 7). Bridgman is notable for taking radiographs that actually focused on the structure of textiles themselves rather than on supporting substructures or purely the metal threads. In a letter of 1 July 1964, he writes to Bird that he is using exposures around 7–10 kV. This enabled him to produce radiographs in which single yarns or threads could be seen and he studied these images further under magnification. He discusses the results of radiography of a piece of modern crêpe cloth and undertook test radiographs of textiles from the Brooklyn Museum, including some Peruvian ‘brocades’. Bridgman chose these deliberately to understand the textile structures and noted the differential image density between the coloured figures and the background weaving. One notable example among the early uses of radiography of textiles was the stereoradiography carried out in the investigation of a fragmentary cushion found in the tomb of Archbishop Walter de Gray (1216–1255) in York Minster (see Chapter 3, p. 51). The images taken by Miss M. Bimson in 1971 at the British Museum revealed details of the metal thread work embroidery and proved ‘extremely useful as a guide to the layout of the design’ during the conservation treatment by Sheila Landi at the Victoria and Albert Museum (Werner, 1971: 139). France-Lanord (1975), a metals specialist involved in excavations at the Basilica of Saint Denis, France, used radiography to locate braids embroidered with gold metal threads on the grave garments of Queen Arnegonde of France who died in 570.4 Robert W. Mottern of the Sandia Laboratories, USA, led the radiographic research into the Shroud of Turin as a member of the 1978 Shroud of Turin Research Project (STURP) team. Radiography was undertaken at approximately 15 kV using a wire grid to provide orientation (Mottern et al., 1980). The resolution was sufficiently good to visualise individual linen yarns; these were about 0.15 mm in diameter. The resulting radiographs were magnified and contrast

6

Textile X-radiography

enhanced. They clearly showed the differences in the herringbone twill weave of the shroud and the backing fabric. Various stitches, patches, water stains, and tiny foreign bodies were also visible; neither the body image nor the stains thought to be blood showed up on the radiographs. Evidence from the radiographs was used in the debate over the nature of the so-called ‘seam’ between the side strip and the main section of the shroud. Information from the radiographs allowed differences in the individual yarns and weaving flaws to be tracked on either side. They also enabled the search for cut or frayed ends and the nature of the stitches in the seam to be examined. As a result, researchers argued that this feature was not a seam but a tuck (Adler et al., 1997; Schwalbe and Rogers, 1982; Whanger and Whanger, 2005). Visual examination of individual yarns supported this conclusion (Adler et al., 1997). In 1979, Hannelore Herrmann (Bayerisches Landesamt für Denkmalpflege) had some small reliquary bags and a mitre, traditionally thought to have belonged to Saint Wolfgang of St Emmeram, Regensburg, Germany, radiographed at the Germanisches Nationalmuseum, Nürnberg. The radiography was undertaken to enable greater understanding of the mitre without dismantling it; some hidden script was revealed. The contents of the textile reliquary bags were also established without having to open them (Herrmann, 1980). The Germanisches Nationalmuseum continued to explore the

radiography of cultural material, publishing a calendar featuring radiographs with Siemens Medical Radiography in 1993. These illustrated the study of museum artefacts through radiography, including a nineteenth century dressed papiermâché doll (HG 10286) and Martin Behaim’s Erdapfel or ‘earth apple’, a renowned fifteenth century globe (WI 2826; see Figure 3.19a, p. 53). Typically, these images focused on the substructures rather than the textiles themselves (Mould, 1993: 102). Similarly, Noël’s research while a student at the Institute National du Patrimoine, Department des Restaurateur du Patrimoine, Saint-Denis La Plaine, France, was centred on using radiography to explore the substructure of dolls from the Koenig Collection (Noël, 2004: 28–30). The radiographs revealed the shape and extent of the shoulder plates and metal armatures as well as recording the location of jewellery, metal lace, sequins and beads. The National Aeronautics Space Administration (NASA) used radiography to ensure that no sharp foreign bodies that might cause punctures, such as needles or pins, were present in the pressure suits worn by astronauts. NASA also made radiographs of people wearing space suits to check their fit.5 Although taken for practical safety purposes, the preflight radiograph of Neil Armstrong’s moon boots also provides an excellent record of their materials and construction (Figure 1.2). Once in collections, the left glove worn by Michael Collins on an Apollo

Figure 1.2 Pre-flight radiograph of Neil Armstrong’s boots taken on 7 July 1969 by Jack A. Weakland, NASA. (Scanned by Ulrich Lotzmann; reproduced with acknowledgements to NASA.)

Introduction 7

mission was radiographed in 1994 at the California Institute of Technology and an Apollo space suit was radiographed using a computed tomography (CT) scanner at the Smithsonian National Museum of Natural History as part of the ‘Save America’s Treasures’ (SAT) millennium project to document and preserve the Apollo space suit collection. The pressurised flight suit made from rubberised parachute fabric for the pioneering aviator Wiley Post (1899–1935) has also been radiographed by the Smithsonian Centre for Materials Research and Education (SCMRE).6

Radiography at the Textile Conservation Centre Several projects undertaken at the Textile Conservation Centre (TCC), Winchester School of Art, University of Southampton, in recent years have involved some use of radiography, often in the investigation of underlying structures. Dinah Eastop, Senior Lecturer and Associate Director of the AHRC Research Centre for Textile Conservation and Textile Studies, and Kathryn Gill,

(a)

Senior Conservator/Lecturer have both made significant and pioneering contributions. One of the earliest examples may be the radiography of flood-damaged seventeenth century hanging wall pockets. In 1983, as a student conservator, Lynn Grant took the pockets to be X-rayed at the Courtauld Institute of Art, London, before starting the conservation treatment which involved the removal of the fungi layer obscuring the metal thread embroidery (Grant, 1983). The radiographs were used to good effect in recording and deciphering the iconography of the raised metal thread work. The radiography and treatment of the pockets forms the basis of a case study in this book (see Chapter 18). Kim Leath’s final year student conservation project was the investigation and treatment of a mixed media bonnet dated 1840–1850 (Leath, 1995). This drawn bonnet had an outer layer of rouched silk fabric over a system of semi-rigid hoops. It is lined with a cotton fabric and trimmed with silk ribbon (Figure 1.3a). Radiographs taken at a dental facility7 were used to explore the construction, number and condition of these hoops (Figure 1.3b). There proved to be three types of hoop: flat iron or steel hoops, threestrand plaits of thin wire and hoops of wood ‘chip’, a

(b)

Figure 1.3 Bonnet, (a) photograph before conservation, (b) drawing showing location of dental radiographic film. (TCC 1943; © Textile Conservation Centre, University of Southampton; reproduced by permission of Barnet Museum.)

8

Textile X-radiography

(a)

(b)

Figure 1.4 Tog, a stop animation puppet, (a) photograph, (b) radiograph. (TCC 2875; © Textile Conservation Centre, University of Southampton; reproduced by permission of Peter Firmin.)

millinery term for plaits made from three strips of wood covered in sugar paper. The images also revealed breaks in the wire hoops. Rebecca Smith (2004) investigated the puppet Tog for her dissertation. Tog was created by Peter Firmin for Pogle’s Wood, a children’s TV programme written and produced by Oliver Postgate, which was originally broadcast by the BBC between 1964 and 1968 (Figure 1.4a). The materials used in Tog’s construction were many and diverse, including cotton and acrylic fabrics, rabbit fur, foam resin filling, an epoxy resin moulded body and an internal jointed skeleton of wood, nickel plated brass Meccano components and steel wire. Radiographs taken by the Conservation Centre, Salisbury, recorded most of these materials including the metal, wood, stuffings and the entire outline of the moulded body. These also revealed cracks and repairs

to the ears (Figure 1.4b). The information gained helped complete the documentation and influenced the recommendations made for Tog’s storage and display (Smith, 2004).

Imaging textiles using mammography Professor Elia Flores, Director of the Radiology Programme, Blinn College, Bryan, Texas, had been using conventional clinical radiographic techniques to look at human mummies in conjunction with Dr Sonia Guillen, Director of the Centro Malliqui, Bioanthropology Foundation of Peru. In 2002, she evaluated the potential of radiographic equipment usually used for mammography for radiographing textiles from the mummies. Dr Guillen’s question

Introduction 9

(a)

(b)

(c)

(d)

Figure 1.5 Peruvian mummy textiles: fragment with pile and fringe, (a) clinical radiograph, (b) mammograph; pattern woven fragment, (c) photograph, (d) mammograph. (Photographs © Sonia O’Connor, University of Bradford; radiographs © Elia Flores, Director of the Radiology Programme, Blinn College, Bryan, Texas; reproduced by permission of Sonia Guillen, Bioanthropology Foundation of Peru.)

10

Textile X-radiography

was whether such radiographic imaging of textiles could provide enough information to determine not only wear, damage and repair for informing future conservation strategies but also give detailed information about obscured weaving techniques. Flores had previously used a single phase X-ray generator, double emulsion medical film and par speed screens (fluorescent intensifying screens) but found the results were not satisfactory, having low resolution and contrast (Figure 1.5a). However, when she turned to mammography and employed single emulsion mammographic film, Flores had greater success (Figure 1.5b). The X-ray tube in mammography is engineered to provide higher resolution and contrast, using lower beam energies than are customarily used in general clinical radiography. Flores imaged three pieces of textile selected to show wear, decoration or weaving technique (Figure 1.5c and d). It was concluded that the use of mammography was a great improvement on the medical equipment used earlier. It was capable of bringing out details of the yarns and threads and providing information about wear, damage and repair.8

Developing practice The significant shift in the radiography of textile artefacts has been from using radiography to look at substructures to using it to look at the textile itself. As increased understanding of appropriate radiography techniques for textiles is becoming known, textile conservators are embracing the methods wholeheartedly. As with every conservation analysis and treatment, it is important to balance time, energy, costs and potential risks to the object against the possible gains. When estimating costs, access to facilities, development time and image analysis should all be included in the budget. In comparison to the costs of other analytical techniques – and even that of commissioning professional photographs – the costs are relatively minor in comparison to the information that is obtainable together with the creation of a one-to-one record of the artefact. Radiography is also excellent for communicating ideas, information and excitement about objects and conservation to the public. An early example of this is the 1967 exhibition of thirty-five sixteenth and seventeenth century paintings and their radiographs from the Museum of Utrecht’s collection (Houtzager et al., 1967). Helmus (forthcoming) notes the museum was a pioneer in introducing the public to the use of radiographs in art historical research. Bowyer’s

case study in this volume (Chapter 15) discusses the dramatic use of radiographs in the Hidden Structures section of The Fashion Gallery at Snibston Discovery Park, Leicestershire. Radiographic images of quilts were included in a quilt exhibition at York Art Gallery (Brooks, O’Connor and Sheppard, Chapter 22). The radiographs provided the public with further insights into the construction of the quilts as well as being beautiful images in their own right. As understanding of techniques and the potential information gain grows, radiography should, hopefully, become established as a standard procedure for documenting textiles and contribute to the development of conservation strategies by providing information which enables clearer treatment decision making.

Notes 1.

2.

3. 4.

5. 6.

7.

8.

This nursery rhyme could refer to several mothers famous for their large families, such as Queen Caroline, George II’s wife, who had eight children (Opie and Opie, 1951: 434–435). GG’s Garland (1784) contains the first published version. Unpublished letter to Junius Bird, 16 November 1962, from E. L. Kropa, Chief Chemical Consultant, Battelle Memorial Institute, Columbus, Ohio. We would like to thank Vuka Roussakis, American Museum of Natural History, for providing copies of these letters. Unpublished letters, 1 and 7 July 1964. We would like to thank Vuka Roussakis, American Museum of Natural History, for providing copies of these letters. The excavations at Saint Denis, directed first by Edouard Salin, were carried out by the staff of the Nancy archaeological metal laboratory from 1953 to 1954 and 1957, and continued until 1961 under the direction of Michel Fleury. Personal communication, Amy Ross, Space Suit Project Engineer, NASA, Johnson Space Centre, 12 April 2005. Personal communication, Mary Ballard, Senior Textile Conservator, Museum Conservation Institute, Smithsonian Institution, Maryland, via Amanda Young, Curator, Division of Space History, National Air and Space Museum (NASM), Smithsonian Institution, Washington, 2005. Mr John S. White, BDS (Syd.), Dental Surgeon, Hampton Wick. Personal communication, Dinah Eastop, Senior Lecturer, Textile Conservation Centre, 16 August 2006. Personal communication, Elia Flores, Director, Radiology Programme, Blinn College, Bryan, Texas, 26 July 2006.

Introduction 11

References Adler, A. D., Whanger, A. and Whanger, M. (1997). Concerning the side strip on the Shroud of Turin. In Actes du IIIème Symposium Scientifique International du CIELT, Nice. Imprimerie de la Neuvelle Clemacy. Bridgman, C. F. (1973). The radiography of museum objects. Expedition, 15(3), 2–14. Brooks, M. M. and O’Connor, S. A. (2005). New insights into textiles. The potential of X-radiography as an investigative technique. In Scientific Analysis of Ancient & Historic Textiles, Informing Preservation, Display and Interpretation. Post-prints of the AHRB Research Centre for Textile Conservation & Textile Studies, 13–15 July 2004 (R. Janaway and P. Wyeth, eds), pp. 168–176, Archetype Press. Brooks, M. M., O’Connor, S. and McDonnell, J. G. (1996). The application of low energy X-radiography in the investigation of degraded historic silk textiles. In Preprints of the 11th Triennial ICOM-CC Triennial Meeting Edinburgh, Scotland, 1–6 September 1996 (J. Bridgland, ed.), pp. 670–79, James & James. Duffin, J. and Hayter, C. R. R. (2000). Baring the sole: the rise and fall of the shoe-fitting fluoroscope. Isis, 91(2), 260–282. France-Lanord, A. (1975). Les textiles brodé d’or dans les tombes princieres d’époque mero-vingienne à SaintDenis. In Conservation in Archaeology and the Applied Arts. Preprints of the Contributions to the Stockholm Congress, 2–6 June 1975, pp. 15–18, International Institute for Conservation. Grant, L. (1983). The Conservation of a Set of Seventeenth Century Hanging Pockets TCC 0499. (Unpublished Diploma Report, Textile Conservation Centre.) Helmus, L. M. (forthcoming). The Museum Curator: A Vital Link between the Conservator and the Public. (Paper presented at the ICOM Conservation Committee 14th Triennial Meeting, The Hague, 12–16 September 2005.) Herrmann, H. (1980). Röntgenaufnahmen als hilfsmittel bei technischen untersuchungen in der textilrestaurierung. Arbeitsblatter Gruppe 10 Textilien, 1(10), 54–65. Houtzager, M. E. et al. (1967). Röntgenonderzoek van de Oude Schilderijen in het Centrall Museum in het Utrecht. Centraal Museum, Utrecht.

Lang, J. and Middleton, A., eds (2005). Radiography of Cultural Material (2nd ed.), Elsevier. Leath, K. (1995). The Investigation and Treatment of a MultiMedia Bonnet, c.1840–1850. TCC 1943. (Unpublished Diploma Report, Textile Conservation Centre.) Mottern, R. W., London, J. R. and Morris, R. A. (1980). Radiographic examination of the Shroud of Turin – a preliminary report. Materials Evaluation, 38(12), 39–44. Mould, R. F. (1993). A Century of X-rays and Radioactivity in Medicine. Institute of Physics Publishing. Noël, G. (2004). La Collection de Poupées Koenig: son Histoire la Restauration de Quatre Poupées Costumées. Etudes et Réalisation de Supports de Conservation Presentation. (Unpublished thesis, Institute National du Patrimoine, Department des Restaurateur du Patrimoine, Textiles, Saint-Denis La Plaine, France.) O’Connor, S. A. and Brooks, M. M. (2005). Making the invisible visible: the potential of X-radiography as an investigative technique for textile conservation decisionmaking. In Pre-prints of the 14th Triennial Meeting of the ICOM Conservation Committee, The Hague, 12–16 September 2005 (I. Verger, ed.), pp. 954–962, James & James. Opie, I. and Opie, P. (1985). The Oxford Dictionary of Nursery Rhymes. Oxford University Press, pp. 434–435. Schwalbe, L. A. and Rogers, R. N. (1982). Physics and chemistry of the Shroud of Turin: a summary of the 1978 investigation. Analytica Chimica Acta, 135(1), 3–49. Smith, R. (2004). TV Puppets from the 1960s and 1970s: Creation, Materials and Conservation. TCC 2875. (Unpublished MA dissertation, Textile Conservation Centre.) Whanger, A. D. and Whanger, M. (2005). Radiological aspects of the Shroud of Turin. (Excerpt from conference paper given 10 September 2005 at the 3rd International Dallas Conference On the Shroud of Turin, Dallas, Texas, USA; see www.shroud.com/dallas3.htm#Conference.) Werner, A. E. A. (1971). The scientific examination and treatment of objects from the tombs. In The Tombs of Archbishop Walter de Gray (1216–55) and Godfrey de Ludham (1258–65) in York Minster and their Contents (H. G. Ramm, ed.), p. 139, Society of Antiquaries of London.

2 Principles of X-radiography Sonia O’Connor

Introduction Conservators in many disciplines, and other heritage professionals, already receive some instruction in radiography as part of their training and it is to be hoped that this will become an integral part of the education and practice of textile conservators. However, radiography is rarely taught at anything other than a basic level, and any improvement of knowledge or skills depends on opportunities for continuing professional development. Taking high quality radiographic images of textiles can be difficult and a radiographer with no experience of imaging organic materials may have little idea where to begin. It is not uncommon to find that even professional medical and industrial radiographers’ first reactions are to suggest that no information would be gained from trying to X-ray materials as ephemeral as silk, cotton or wool because these materials are so far outside their own experience. This book is not intended to be a course book for delivering training in the radiography of cultural materials. For accounts of radiographic imaging theory and practice, see Halmshaw (1995) on industrial radiography and Lang and Middleton (2005) for the application of radiography to cultural material. Graham and Cloke (2003), although dealing with the medical applications of X-rays for diagnostic imaging and radiotherapy, provide a clear and comprehensive account of the principles of radiological physics. The purpose of this brief discussion of conventional radiography is to provide sufficient understanding of radiographic principles and practice to allow those with some knowledge and experience of the radiography of cultural materials to obtain optimal results when radiographing textiles. Additionally, the aim is to give those who commission radiographs the understanding and vocabulary required to select the facilities appropriate to their needs and to communicate those needs precisely to radiography professionals. 12

This chapter therefore deals with scientific principles and the details of practical procedures, making it quite dense and information-rich. Examples and illustrations of the points made here are given in Chapter 3 ‘High definition radiography of textiles: methods and approaches’ but readers are advised to persist with the current chapter before turning to the more textilebased examples in order to gain maximum benefit from the chosen examples.

X-rays and the electromagnetic spectrum X-rays are streams of photons of electromagnetic energy. Photons are mass-less particles that travel in straight lines at the speed of light, producing waves of vibrations of electric and magnetic fields at right angles to the direction of travel. The energy of these photons determines the properties of the different forms of radiation in the electromagnetic spectrum (Figure 2.1). The wavelength of the vibrations is dependent on the energy of the photons: the higher their energy, the shorter the distance between the peaks of the waves and, therefore, the greater the number of peaks in a given distance. The photons in gamma (γ) radiation are high energy so this is short wavelength, high frequency radiation while radio waves are low energy, long wavelength, low frequency radiation. The most energetic radiations are γ-rays, then X-rays, ultraviolet (UV) light, visible light, infrared (IR) light, microwaves and finally radio waves. Each group has particular properties that define it but the variation of the photon energy throughout the spectrum is continuous so there is a little overlap of particular properties across the boundaries of adjacent radiation groups. For instance, high energy UV radiation has some characteristics in common with very low energy X-rays, also called ‘soft’ X-rays. For this reason, the energy limits quoted for each radiation group are

Principles of X-radiography 13

Figure 2.1 The electromagnetic spectrum. (Drawing by Jason Maher.)

usually given as an approximation. However, the energy ranges of γ-rays and high energy or ‘hard’ X-rays actually overlap. The difference between them is really the means of their production. Gamma rays are generated from within the nucleus of an atom, for instance, due to radioactive decay or high energy particle collisions. X-rays are generated either by electrons changing energy levels within the shells of an atom or by the interaction of free electrons with, for instance, a magnetic field in a synchrotron or charged particles in the target anode of an X-ray machine. In the area of overlap, photons of a particular energy will be identical in their properties whether it is γ- or X-radiation.

Properties of X-rays X-rays can travel through a vacuum and are able to penetrate deeply into, or through, materials that are opaque to visible light. Like most electromagnetic radiation, X-rays are invisible, as the retina of the eye is only sensitive to the narrow band of the electromagnetic spectrum known as visible light. However, X-rays cause the same chemical changes as light in photographic film emulsions. They are termed ionising radiation because they have enough energy to move electrons out of orbit around the atoms of the material with which they interact. Most importantly they can damage or destroy living cells so the safe use of X-rays is paramount (see Chapter 7).

As a beam of X-rays passes through matter some of the photons will be absorbed or scattered by interaction with the atoms of the material. The resultant reduction in beam intensity is referred to as attenuation. To put it very simply, the bigger the atoms, the more closely they are packed, and the more atoms there are in the path of the beam, the greater the probability of these interactions occurring. Therefore, the thicker and denser a material is, and the higher its atomic number, the more it will attenuate an X-ray beam. If a beam of X-rays passes through a heterogeneous object, variations in its shape, structure and chemistry will produce differential attenuation of the transmitted beam. These variations in the X-ray flux (beam intensity and energy range) can be used to form a radiographic image of the object either using specially formulated radiographic film, in the same way that a photographic negative is formed by light, or by using a receptor that detects variations in the ionisation effect of the beam (see Chapter 4).

Production of X-rays Although X-rays and γ-rays have overlapping energy ranges, differences in the ways that they are produced mean that they behave rather differently when they are used for imaging objects. Gamma

14

Textile X-radiography

rays emitted by a radioactive substance have only a few discrete energies. Plotting intensity against energy produces a series of peaks, called a line spectrum. The intensities and energies of the lines in the spectrum depend on the radioisotope involved and cannot be altered. Gamma rays can be very useful where very high penetrating power is required, for instance, when imaging features in thick metal specimens but in medical and most industrial imaging applications. In the radiography of textiles, electrically generated X-rays are used. These X-ray machines produce emissions of many energies which form a continuous spectrum and the range and intensity of the beam can be controlled. Safe working practice ensures that no personnel are exposed to the X-rays during imaging and once the electricity is turned off no X-rays can be produced (see Chapter 7). In very simple terms, X-rays are electrically generated in an X-ray tube. This is a sealed metal–ceramic or heatproof glass vessel in which a very high vacuum is maintained (Figure 2.2). The tube encapsulates two electric poles, separated by a small gap. At the cathode (negative pole) is a wire filament that glows and emits electrons when an electric current is passed through it. The anode (positive pole) has a target that is made from a heavy metal such as tungsten or molybdenum recessed into its surface. When a high electric potential (voltage) is set up across these poles, electrons emitted by the filament are attracted to the anode and stream across the gap. When these accelerated electrons collide with the target they are suddenly Electrons

Sheilding

Focal spot

Cathode

decelerated and release energy in the form of photons of electromagnetic radiation mostly as heat, a little light and about 1% X-rays. The small area of the target bombarded by the electron beam is called the focus or focal spot but, depending on the angle of the target relative to the electron beam, the effective focal spot size of the X-ray beam produced can be much smaller. The X-rays leave the tube through a window, producing a beam in the form of an expanding cone. Figure 2.3 shows a typical X-ray spectrum where the number of the X-rays emitted (intensity) rises and then falls away again with increasing energy. The very lowest energy X-rays of the spectrum penetrate the vacuum of the tube but are absorbed by the material of the tube window. The smooth line of the spectrum is disrupted by a number of peaks (spectral lines) indicating high intensities of X-rays of discrete energies that look quite out of place. This occurs because the X-rays are formed in two different ways in the X-ray tube. Electrons passing near the nucleus of atoms of the target are bent from their path by the attractive forces of the nucleus. The closer they pass to the nucleus, the more they are deflected from their original path and they may suddenly emit energy in the form of electromagnetic radiation and slow down. These emissions form a continuous spectrum of energies as some electrons may slow down more than others depending on their interaction with the nucleus. Those that collide directly with a nucleus will lose all their kinetic energy. Some of the emissions

Vacuum

Target

Anode

Glass tube

Filament

Window

X-rays

Figure 2.2 Diagram of an X-ray tube. (Drawing by Jason Maher.)

Principles of X-radiography 15

will have energies within the X-ray range and this is termed Brehmsstrahlung from the German Brehms for braking and Strahlung for radiation. On top of this X-ray spectrum are superimposed spectral lines characteristic of the metal of the target. These are produced when the accelerated electrons collide with inner shell electrons in the atoms of the target. The electron that is hit is ejected from the atom and an electron from a higher orbit drops down to take its place with a consequent loss of energy which is emitted in the form of an X-ray photon of a specific energy. Beam energy and intensity The electrons emitted by the filament form the tube current, which is measured in milliamps (mA). The voltage across the tube is measured in kilovolts (kV). This voltage may vary cyclically due to the rectification of the unit’s power supply and so the tube voltage is expressed in terms of the peak voltage of this cycle (kVp). Increasing the mA increases the intensity of the X-ray beam. The kVp determines the maximum energy of the electrons (measured in kilovolt electron potential or keV) hitting the target, which in turn determines the energy of the X-ray photons emitted by the target. An X-ray tube operated at 25 kVp will produce X-rays with a maximum energy of 25 keV. However, because the X-ray beam produced is not of a single energy but a range of energies, it would be misleading to describe it as a 25 keV beam. Instead, it is conventional to describe the beam produced at 25 kVp as 25 kV

X-rays. Increasing the kVp also increases the intensity of the X-ray beam to a small extent.

Conventional radiography In conventional, or transmission, radiography the object is placed between the source of radiation and the image receptor, most commonly radiographic film (Figure 2.4). The object is exposed to the X-rays and the image, or radiograph, of the object is formed on the film due to the differential attenuation of the beam by the object. A radiographic image is essentially a twodimensional representation of variations in radioopacity produced by a three-dimensional object. It is similar to a black and white photographic negative; the image will be darkest where the greatest number of X-rays has reached the film and lightest where the greatest proportion of the X-rays is absorbed by the object. The higher the atomic number of the materials of an object, the thicker it is, and the denser its structure, the more the X-rays will be absorbed. In its simplest terms a radiographic exposure is a combination of two parameters, the kV of the beam and the overall X-ray dose.

X-ray Source

X-rays

Unfiltered in vacuum Spectral lines

Brehmsstrahlung 50

100

Specimen 150 Photon Energy (keV)

Figure 2.3 Graph of X-ray energy plotted against intensity, showing the shape of the spectrum typically produced by an X-ray tube at 150 kVp. (Drawing by Jason Maher.)

Film

Cavity

Shelf

Figure 2.4 Typical layout for conventional or transmission radiography. (Drawing by Jason Maher.)

16

Textile X-radiography

Beam kV Most importantly, the kV of the X-ray beam affects the contrast of the image and has to be matched to the object being X-rayed. X-rays of different energies are absorbed to different extents as they pass through material. Low energy X-rays are more readily absorbed than the more penetrating, higher energy X-rays, so a small variation in the radio-opacity of an object will have a greater effect on the attenuation of a low kV beam than a high kV beam. As the amount of radiation reaching the film affects the density of the image, the greater the variations in the X-ray flux, the bigger will be the differences in the shades of grey formed. Lower energy X-rays will therefore produce a higher contrast image of an object than will higher energy X-rays. Low energy X-rays are useful for radiographing thin, low atomic number materials like organic textiles. With increase of thickness, material density and atomic number, higher and higher energy X-rays are required to penetrate the material. If the kV is too low, the object will absorb all the X-rays and the image will only be a white silhouette with no internal detail, no matter how long the exposure is run. If the kV is too high for the object, the image will have very low contrast and features of similar radio-opacity will be lost in the too-close shades of grey. Low kV, high contrast images, however, have a narrower exposure latitude than higher kV images. This is because the bigger the difference between the shades of grey, the fewer shades there are between black and white. High contrast is ideal for imaging fairly homogeneous objects but if the radio-opacity of an object varies due to diverse materials or thicknesses, some areas may be black on the film due to overexposure while other areas, which have not been penetrated, may be white. If the variation in radio-opacity is not too great, raising the kV a little may overcome these problems by increasing the exposure latitude but the image contrast will be reduced and the fine detail lost.

expressed as a combination of current and time – mAs (milliamps per second). However, if the X-ray focus to film distance (FFD) is varied, the intensity of the beam will also change because the X-rays are diverging as they move away from the focus, producing a widening beam in which the X-rays are spread increasingly thinly. This change in intensity conforms to the inverse square law1 so doubling the FFD reduces the intensity of the beam by a factor of four, while halving the beam distance increases its intensity by a factor of four. If an object is overexposed, too many X-rays will have penetrated it during the exposure and its image may be so blackened that it is not distinguishable from the background of the radiograph. If an object is underexposed, insufficient X-rays will reach the film during the exposure and the image may only appear as a white or relatively featureless, pale grey silhouette against the darkened background.

Optimising image quality A good quality image is one that is adequate for the task; that is, the sensitivity of the image is sufficient to show the features that are being sought. The process of obtaining high definition radiographs starts with the selection of appropriate X-ray facilities but the quality of the image can be detrimentally affected by many factors, such as how the X-ray unit is used and how the image is recorded. Gauging the quality of the image can be particularly problematic when X-raying objects like cultural material where there are greatly varied and unknown features to be revealed. It is difficult to be confident whether an absence of features on the image is a true reflection of the nature of the object or due to poor, or inappropriate, radiographic technique. To ensure good quality, usable results that capture the maximum amount of information, all the following issues need to be addressed in relation to the object being X-rayed. Selection of X-ray facilities

X-ray dose X-ray dose affects the general brightness or darkness of the radiographic image. The dose is a combination of the beam intensity and the time for which the exposure lasts. Beam intensity is directly proportional to the tube current. A high X-ray tube mA and a short exposure can be used to deliver the same X-ray dose as a long exposure at a low mA. In clinical radiography, where calculating the total X-ray dose given to the patient is very important, the exposure is

Not all radiographic facilities are designed to produce high definition images as this is not necessarily the purpose of the process. Medical radiography facilities are, in general, not well suited to producing high definition radiographs of cultural material because the priority is to balance image quality against patient safety. Exposures have to be very short to prevent blurring of the image due to patient movement and the X-ray dose has to be kept as low as practically possible. To protect

Principles of X-radiography 17

patients further, the X-ray beam is heavily filtered to remove all X-rays with energies up to about 40 kV as these lower energy X-rays are particularly damaging to living cells. Unfortunately, this is just the energy range that is particularly useful for producing high contrast images of organic objects. With industrial radiography, there is no necessity to compromise image quality for patient safety and industrial equipment and techniques are better suited to the imaging of cultural material. However, not all industrial X-ray units are suitable for this work. For instance, real-time radiography equipment, used to record the behaviour of molten metals flowing into moulds or exploding munitions, may be designed to capture hundreds of frames per second but to the detriment of image detail. The dual energy X-ray inspection equipment used by airport security to screen passenger baggage is also not designed to produce high definition images. Here the important factors are to use the lowest possible X-ray dose, to prevent damage to photographic films and sensitive electronic equipment, while producing virtually instantaneous ‘on-screen’ images from digitally captured and processed data which are colour coded to pick out metals from organic materials, batteries and detonators. Even equipment designed for high definition radiography may present problems if the kV range does not match the material to be radiographed. The form of the anode and the material of the target

are important in determining the maximum kV of the beam and the proportion of lower kV X-rays produced. Both the material and the thickness of the X-ray tube window can also have a great effect on the spectrum of the beam particularly by absorbing lower energy X-rays. An X-ray unit with a low maximum kV may not generate a beam with a high enough energy to penetrate thick metal objects, but some machines may not operate at all below a fairly high minimum kV, making them unsuitable for low kV imaging of organic materials such as paper or textiles. Beam positioning The relative positions of the X-ray focus, the object and the film can all have a great effect on the quality of the image. As a general rule the object, or area of interest, should be directly under the centre of the beam where the X-rays are most nearly parallel and the features will be recorded in their correct relative positions and at life size (Figure 2.4). To optimise the capture of detail, more than one viewpoint may be required. Moving out towards the edges of the beam the X-rays diverge more and more from the perpendicular and the image becomes spread out. Features here will be slightly magnified and distorted, and fine features, such as vertical cracks, may become obscured (Figure 2.5a). At the edges, the

Effective Focal Spot

X-rays

Focus to Film Distance (FFD)

Specimen Penumbra

(a)

Film (b)

(c)

Figure 2.5 Diagram showing the effect of beam positioning on image quality, (a) object far from the beam centre, (b) object at beam centre with a short FFD, (c) object at beam centre with a long FFD. (Drawings by Jason Maher.)

18

Textile X-radiography

images of features above the film plane may fall beyond the edge of the film and be lost. Another effect of the diverging beam is that it can cause the image to be unsharp. X-rays do not come from a point source but from anywhere in the effective focal spot of the beam. Those X-rays originating from either side of the effective focal spot and passing through a single point in an object will cross each other and continue to diverge until they reach the film (Figure 2.5b). This means that, in the image, every point in the object will be blurred by a penumbra, limiting the resolution of fine detail. All X-ray units produce geometric unsharpness but this is more of a problem in those with a large effective focal spot. Both these problems are least noticeable when the object is flat and in close contact with the film because the X-rays cannot diverge significantly between passing through the object and reacting with the film. To reduce problems with curved objects the film can be supported to maintain contact with the surface of the object, but with more three-dimensional objects it is not possible for all features to be close to the film. Features furthest from the plane of the film will be most fuzzy, magnified and displaced, and will overlay the sharper detail of the features against the film. It may be necessary to turn the object over and take a second exposure to get good quality images of all the features. Increasing the FFD of a system will reduce these problems because the X-rays in the beam will be more nearly parallel (Figure 2.5c). Unfortunately, increasing the FFD greatly decreases the beam intensity so the length of the exposure may quickly exceed the capability of the equipment. Increasing the FFD will also change the character of the beam spectrum as more of the lower energy X-rays are absorbed by the air before reaching the film. The maximum FFD will also be limited by the X-ray equipment or the size of the X-ray room. Cabinet X-ray units provide little room for manoeuvre and a small effective focal spot is the only way of ensuring good quality images. Exposure parameters Wildly differing exposure parameters can produce a radiographic image of an object but obtaining an image with the greatest amount of discernible information generally requires using the lowest practical kV to gain optimum image contrast and exposure latitude. It may not be possible to capture all the information in a single image if the variation in

the radio-opacity of the object is very wide. Raising the kV to improve the exposure latitude may reduce the image contrast to the point that fine detail is no longer discernible. This problem might be overcome by using image receptors with a greater dynamic range than radiographic film (see Chapter 4) or techniques designed to maintain image contrast while extending the exposure latitude, such as lead screen intensifiers (see Chapter 3, pp. 41–3). Alternatively, several exposures can be taken varying the kV and X-ray dose to suit each area of the object, or bi- or tri-pack film (different speed films packaged together) can be used to gain good coverage from a single exposure. Scatter Scattered radiation is another cause of unsharpness in radiographic images. The primary beam X-rays, those passing directly from the focus through the object to the film, will form the image of the object but some of the X-rays will inevitably interact with the materials through which they pass (including the air, surfaces around and below the film and the object itself ) producing lower energy, scattered radiation (see Chapter 7, p. 100). These X-rays radiate in all directions, fogging the film so the edges of features appear blurred. Very low energy X-rays that cannot penetrate the object do not contribute to the formation of the image but can form a significant component of the scattered radiation, increasing the fogging. The amount of scattered radiation produced is reduced by restricting the width of the beam, so that it only covers the area of the image, and by the use of filters. Filtration Filters are metal sheets which are used to modify the beam spectrum and to sharpen images by reducing scattered X-rays. A filter will reduce the intensity of the whole beam but, because lower energy X-rays are less penetrating than higher energy X-rays, it will selectively remove the scatter. The material and the thickness of the filter used depend on the purpose of the filter, the kV of the beam and the object being X-rayed. A filter placed between the X-ray focus and the object can be used to remove the low energy, non-image forming, X-rays from the beam, including scatter. If it is placed between the object and the film, it will also remove the scatter formed in the object. This is particularly useful when radiographing irregular objects that do not lie close to the film. Filters used in this position must be uniform in thickness and free of damage as

Principles of X-radiography 19

the image of any irregularities will appear on the film and be confused with the features of the object. Lead sheets placed beneath the film are used to filter out backscattered radiation from surfaces below the film, such as the floor or the base of an X-ray cabinet, which can be a significant cause of fogging. In removing the ‘soft’ (lower energy) X-rays, filtration is said to ‘harden’ the X-ray beam. Removing these energies alters the image contrast, giving the appearance of a higher energy image without increasing the kV of the beam. This can occasionally be useful with heterogeneous objects as it will reduce the overexposure of thinner areas of the object. Although the image contrast may be reduced, the visibility of features may actually be improved as their edges will be sharpened by the consequent reduction in scatter. In medical X-ray units the filtration cannot be removed and this limits their use for the radiography of cultural material. Industrial units are more versatile but inherent filtration, a product of the materials of the X-ray tube through which the beam is emitted, may mean that the lower energy X-rays are still lost from the spectrum.

Figure 2.6 A stepwedge made from copper alloy sheet stuck together with epoxy resin for use as an IQI when radiographing copper alloy artefacts. (© Sonia O’Connor, University of Bradford.)

when trying to identify equipment faults or film processing problems. X-ray film selection

Image quality indicators To be able to gauge properly the quality of an exposure it is necessary to include something in the radiograph that has known features. This is called an image quality indicator (IQI). The quality of the image of the IQI on the radiograph of an object whose characteristics are not well understood should give the confidence to determine whether certain features are not there to be found or that they cannot be detected by the radiographic method used. In industrial and medical radiography the form and use of IQIs is highly regulated as it is very important to have confidence in the sensitivity of the images. Those used in industry are the most relevant to the radiography of cultural material and typically consist of a set of wires or metal plates of differing thickness drilled with small holes of various diameters. These are made to conform to recognised standards, are expensive and only made in a limited range of materials. For cultural material, however, it is a relatively simple matter to make an effective IQI in the form of a stepwedge from strips of material of similar radio-opacity to the objects to be radiographed (Figure 2.6). The image of the stepwedge provides a lot of information including the sharpness of the image, the limits of the exposure latitude and the contrast of the exposure. It also facilitates comparisons between exposures and can prove invaluable

The information potential of high definition radiography will not be realised if the film or image receptor used to capture the image is unsuitable. Medical radiographic films, for instance, are, in general, designed for speed rather than high definition. Their fast reacting film emulsions have a coarse grain size, which limits the resolution of detail, and the film contrast is low as the maximum optical density (Dmax) formed in the background of the images is only very dark grey, rather than black. Used with fluorescent screens, to cut the X-ray dose required still further, these films produce relatively low resolution, low contrast images. Only mammography X-ray units and films are designed to produce higher definition images. Slow speed, high definition, ‘direct type’ (designed to be used without fluorescent screens) industrial films are most commonly used for the radiography of cultural material. These are double-sided films, that is they have an emulsion on both sides of the film base so that there is no ‘correct side’ when making an exposure, unlike photographic film. Industrial films are high contrast films because they can develop a very much higher Dmax than photographic or medical radiographic films, allowing them to capture many more shades of grey between black and white. This gives better image contrast than medical film and better discernibility of detail. Industrial films come in a range of speeds, all with slightly different characteristics, the

20

Textile X-radiography

slowest ones having the finest grain and the highest Dmax but also being the most expensive. Choosing the best film for high definition radiography depends on the equipment available and the work to be undertaken. There is little point in using an ultra fine-grained film with equipment that cannot produce very high resolution images or for images that will not be inspected under magnification. On the other hand, if the film grain is too coarse the image quality will suffer. Ideally the resolving power of the film needs to be slightly finer than the resolution of the image, unless the higher film contrast of the slower films is desirable. Although using a faster film will compromise the resolution of the image, it may be needed if the duration of the exposure is an issue. Other image receptors are being increasingly used to capture X-ray images. Those most useful to high definition radiography are discussed in Chapter 4 on digital radiography. Film enclosures and cassettes Radiographic films can be purchased in a range of sizes and enclosures to suit most purposes. However, for infrequent users it is sensible to standardise on only one or two varieties as X-ray film has a relatively short shelf life. Films may be pre-packed in individual lightproof enclosures or be unencapsulated. The latter are particularly versatile as they can be used as they are or loaded into cassettes, either of flexible plastic or rigid aluminium, to suit the situation. Any enclosure will attenuate the beam and so the material and thickness of the enclosure must be matched to the kV of the exposure. For instance, a rigid aluminium cassette should not be used below about 60 kV. Above this kV its filtering effect reduces scatter but at lower kVs it attenuates too much of the beam, extending the exposure time needed and producing low contrast images degraded by the texture of the cassette surface. Film handling and processing Radiographic film stock must be stored in cool, dry conditions. It should be handled only at the edges as the emulsions scratch easily, especially when they are wet during processing. Clear patches can be caused by mechanical damage prior to exposure, including pressure during storage. Creasing film before exposure can lead to the formation of clear crescent shapes on the image, while dark crescent shapes indicate creasing after exposure (Halmshaw, 1995: 210). Film can easily become charged with static electricity, especially in conditions of low relative

humidity, and the discharge of static can also produce dark patches. Contamination with processing chemicals, or even water, before development will lead to dark or light patches. Overdate or badly stored films will develop a grey fogging. All these handling and processing faults and image artefacts will degrade the quality of the radiographic image in some way (see Chapter 5, pp. 88–90). These and other image artefacts and their causes are described in detail by Halmshaw (1986: 119–124) and to a lesser extent in his more recent work (1995: 209–211). Quality control of the film processing is essential to ensure that the images are fully and properly developed. Processing involves developing the latent image, fixing the film so that it will no longer react to light, washing out the chemicals and unconverted silver salts from the emulsion, and, finally, drying the film. The correct chemicals must be used or it is possible that the film will not develop its full density range and the image will lack contrast. Industrial film should not, for instance, be processed using a medical film processor. It is equally important to use the recommended chemicals at the correct dilutions, temperatures and timings. Extending development time to compensate for underexposure will only result in excessive contrast, losing the information in the intermediate tones. Shortening the development time to compensate for overexposure will result in low contrast, while general lack of image density is a characteristic of exhausted developer. If the processing is not consistent it will not be possible to assess if the exposure parameters were correctly selected for the object. Perhaps the most common cause of quality loss at this stage is poor manual processing. Manual processing is still widely practised by infrequent users as the set-up and running costs are considerably lower than automatic processing. Lack of appropriate agitation of the film during development, the oxidation of the developer between periods of use and the temptation to terminate the development when the image ‘looks right’ are just some of the causes of degradation of image quality. Properly regulated manual development using film hangers and tanks with temperature control can produce high quality results, while dish development is more likely to result in physical damage to the emulsion, uneven development and cross contamination of the processing chemicals. This reduces their performance and can form dichroic fogging of the film so parts of the film look greenishyellow in reflected light but pink in transmitted light. Whether automatic or manual processing is used there is no quality control unless the activity of the

Principles of X-radiography 21

chemicals is quantitatively assessed on a regular basis. The activity of the developer can be assessed by developing pre-exposed test films, such as Agfa Structurix PMC (Processing Monitoring Control) strips, and measuring the optical density of the different parts of the image using a densitometer. The activity of the fix is assessed by timing how long it takes to ‘clear’ a strip of unexposed film. Both chemicals should be replaced when these measurements fall outside stated limits. Insufficient fixing and insufficient washing will both lead to image degradation in the shorter or longer term. Unlike the times for development recommended by the manufacturers, which should be adhered to rigorously, the times for fixation and washing (at the stated temperatures) should be seen as a minimum, which guarantees the permanency of the image for perhaps five to ten years under optimal storage conditions. For cultural material radiography this would not be considered a ‘permanent’ archive and considerably longer fixing times are used, followed by slightly extended washing, with the intention of extending the image permanency beyond forty years.

Viewing film radiographs A radiographic film image viewed on a light box will have a tonal range dependent on the local densities of the developed emulsion. The photographic or optical density (D) of a given point depends on the ratio of the light that is incident on the film to that which is transmitted through it, expressed as the logarithm (log10) of that ratio. The difference between the minimum densities (Dmin) the film can record (in the unexposed areas) and the maximum densities (Dmax) formed (in the fully exposed areas of the black background) defines the dynamic range of the radiograph. Medical radiographs can capture optical densities over 3, but high contrast industrial films can develop densities of nearly 5. This is a logarithmic scale so the difference is two orders of magnitude, despite the apparently small numerical difference. Depending on the quality and brightness of the illumination, it is possible to use a light box to examine densities up to 3.5 (Halmshaw, 1986: 103). Information captured in areas denser than this can only be detected and examined by digitising the image (see Chapter 4, p. 58). The fine detail and subtle changes of image density produced by textiles will not be seen if the radiograph is viewed in poor conditions. Optimum

conditions are provided by viewing the film against the diffuse, white light of a light box situated in an area with a low ambient light level. A black card mask should be used to prevent any light escaping around the edges of the film. This would otherwise flood the eye, making it impossible to distinguish detail in the darker areas of the image. Closing window blinds and turning off overhead lights will also prevent reflections from obscuring the image. Alternatively, a viewing booth can be constructed around the light box from something as simple as a large cardboard box. With these precautions, the eye will adjust to the level of light being transmitted through the radiograph so that it will be at its most sensitive in distinguishing the variations of shades of grey within the image. The film should be viewed using a low magnification (c 5 to 10) lens or loupe, or a low power binocular microscope to examine the finer detail of the image.

Working with film radiographs One of the biggest problems encountered with film radiographs is the problem of making faithful copies. Whenever X-ray films are handled or viewed they are placed at risk of degradation by exposure to light, atmospheric pollution, temperature and humidity changes, and contamination by finger grease. The film emulsion is easily scratched and even sliding the film out of its protective sleeve or moving it across the surface of a light box can produce scuff marks on the surface. Ideally a working copy of the radiograph is needed so that the original can be kept in a safe archive. One way round this would be to take multiple radiographs of the same exposure but this is expensive and in practice is rarely done. Producing contact copies with special duplicating film is time consuming and usually results in some loss of contrast. Because the Dmax of photographic film is typically much lower than that of industrial X-ray films, information recorded in the darker areas of the radiograph may be lost from the copy altogether. Photographing the X-ray films on a light box and printing copies from the black and white negative is another solution but one which is quite difficult to do well. In addition to loss of contrast, the resolution can be reduced noticeably because the grain of the photographic film emulsion becomes superimposed on that of the original radiograph. Photography can also introduce lens distortions and variations in image density due to uneven lighting. Consequently, it is commonly the case that each

22

Textile X-radiography

unique radiograph is handled, viewed and physically sent, by more or less secure means, to different specialists working with the object, gradually accruing damage until the project finishes and the image is placed in an archive. Even then the lack of a copy creates a problem. Should the radiograph be kept with the institution’s records or passed to the client? In either case, one partner will receive an incomplete record. Converting radiographs to digital images means that it is now possible to extract more information from them than is possible by viewing them on a light box and to create faithful copies for archive, dissemination and publication (see Chapter 4).

Storage and archive of radiographic films Radiographs must be kept clean, free from fingerprints, dirt and dust, and away from any materials which might contain water, damaging chemicals or fungal spores. The recommendations for the archival storage of radiographs, based on those for photographic films (British Standards Institution, 2000), include dark, stable conditions at temperatures below 21°C and relative humidity between 20% and 50%. The radiographic film manufacturer Kodak recommends a lower relative humidity limit of 30%. Storage below this level, combined with higher temperatures, can occasionally lead to emulsion cracking. Damp conditions can lead to fungal growth. High storage temperatures will increase the rate of chemical deterioration leading to yellowing of the film and greatly shortening its useful life. The films should be separated from each other by acid-free paper envelopes or inert, transparent polyester sleeves. The light-tight paper envelopes from the

pre-packed films and the interleaving papers of the unencapsulated films provide useful temporary protection. If films are stored flat they must be protected in boxes from pressure damage due to stacking. Ideally they should be stored vertically, supported to prevent bending.

Summary Radiography is an extremely versatile imaging tool though its true potential for the investigation of cultural material is often masked by poor practice. Through better understanding of the principles of radiography, the rigorous application of procedures and quality control, it is possible to produce images with the sensitivity to enhance greatly our understanding of objects as varied and ephemeral as textiles.

Note 1.

The intensity of the beam is inversely proportional to the square of the distance from the focal spot.

References British Standards Institution (2000). Imaging Materials – Processed Safety Photographic Films – Storage Practices, BS ISO 18911:2000. British Standards Institute. Graham, T. and Cloke, P. (2003). Principles of Radiological Physics. Churchill Livingstone. Halmshaw, R. (1986). Industrial Radiography. Agfa-Gevaert. Halmshaw, R. (1995). Industrial Radiology. Theory and Practice (2nd ed.). Chapman and Hall. Lang, J. and Middleton, A. (2005). Radiography of Cultural Material (2nd ed.). Elsevier.

3 High definition X-radiography of textiles: methods and approaches Sonia O’Connor

Introduction It is perfectly possible to obtain good quality images of textiles with radiographic equipment already widely used in the radiography of cultural material. The skill is in selecting the correct kV for the materials involved and in taking high definition radiographs. The aim of this chapter is to explain best practice in textile radiography, both for those commissioning and those taking radiographs, so that the maximum benefit is gained from its application.

Why textiles seem difficult to X-ray There are several reasons why textiles seem hard to radiograph successfully, including the nature of materials from which they are made, their structure and the shape of the individual components. The organic molecules from which natural and most man-made fibres are formed, typically carbon, hydrogen, oxygen with nitrogen and some sulphur in proteinaceous fibres like wool and silk, are all of low atomic number. In relation to their diameter, textile fibres may have quite low densities, due to thin walls and cell cavities in cotton and linen or medullary cavities largely filled with air in wool. They may also contain some water; the state of hydration depends on factors such as the fibre type and the ambient relative humidity. However, a single yarn or thread may be very radio-lucent, doing little to attenuate an X-ray beam on its path through the textile to the film. The structure of a textile fabric may also be very thin. The difference in beam attenuation between a single yarn and two yarns crossing will have very little effect on the overall beam intensity. As a result,

even when using the least penetrating, low energy X-rays, radiographic images of textiles are often of very low contrast. If textiles are layered with other materials of rather greater radio-opacity, the image of these materials may make the textile difficult to see or mask its presence altogether (see Chapter 19). Textiles are mostly very finely detailed structures and, in order to capture that detail, the image quality or ‘sensitivity’ of a radiograph, that is its ability to show details of a given size, has to be very good to allow examination under magnification. Under perfect conditions an abrupt change in radio-opacity should produce a sharp change in image density but textile yarns or threads generally have ‘soft’ edges. They are thickest in their centre but gradually thin towards the sides, which may fray to individual fibres with very rounded cross-sections. This produces a gradual transition in density in the radiograph between the yarns and the background. If the textile is tightly woven the definition between one yarn and the next may be lost altogether. The image may appear to be a featureless grey shadow to the naked eye but information may still be retrievable through digitisation and digital image manipulation (see Chapter 4). Such apparent lack of information may be made worse by problems of geometric unsharpness in the X-ray equipment, X-ray beam scatter, incorrect film selection, poor quality film processing and viewing the radiographs in unsuitable conditions, all topics discussed in general terms in Chapter 2. In addition, many textile objects are far from flat so some parts will not be in close contact with the film. Images of these areas will therefore be magnified and blurred to some extent. In the case studies presented in Part 3 of this book, there are several higher beam energy images 23

24

Textile X-radiography

(a)

(b)

Figure 3.1 Painted face of the Madonna from a chalice veil; see Chapter 17, (a) photograph, (b) radiograph. (Private collection; © Sonia O’Connor, University of Bradford; reproduced by permission of James Spriggs.)

taken to examine the more substantial elements of composite textile objects, such as metal thread work, criers and joints in dolls or frames in upholstered furniture. The textiles, if visible at all in these images, may appear as nothing more than faint, thin lines where they are caught on edge. This can help to reinforce the misconception that nothing can be gained from the radiographic imaging of the textiles themselves. Where textiles are detected at these higher beam energies, it is often due to the presence of more radio-opaque materials. For example, the paint on the face of the Madonna from the centre of the chalice veil discussed in Chapter 17 has made the detail of the textile visible in the radiograph (Figure 3.1a and 3.1b). Sometimes the textile is not visible but its presence is inferred from the pattern of absence or thinning of the more radio-opaque material. When radiographing paintings on canvas it is the presence of the ground in the interstices of the weave that attenuates the X-ray beam and reveals the presence and detail of the textile (see Chapter 29). A similar effect to this is seen in the card and textile structure of the doll in Figure 20.6e (see p. 256).

Successfully acquiring high quality radiographic images of the organic textiles themselves requires an understanding of low energy, high definition radiography.

Low energy high definition radiography Low energy or ‘soft’ X-rays, below 40 kV, are not particularly widely used in radiographic imaging. Industrial radiography is commonly associated with the non-destructive testing and evaluation of metal components and pipelines but there are increasing numbers of applications for low energy X-ray imaging such as the quality control of food and plastic products and in the development of modern textiles and textile-reinforced composite materials. In clinical radiography only mammography, an imaging technique used in the diagnosis of the early stages of breast cancer, utilises low energy X-rays. The radiography of most cultural material generally requires relatively high beam energies (Lang and Middleton, 2005) so the majority of people working

High definition X-radiography of textiles: methods and approaches 25

with cultural material are not familiar with the particular benefits and problems of very low energy imaging. Objects in stone and metal are usually imaged above 60 kV and large objects with constituents of high atomic number, such as cast bronze cannons, may require several hundred kV, or even γ -rays, to provide sufficient penetration. A few materials, however, may require low beam energies (below 40 kV), for imaging, for instance the radiography of ceramics and paintings on wood or canvas. O’Connor and O’Connor (2005) discuss the radiography of pathology in bird bones and many of the observations that are made relating to equipment choice, film selection and radiographic technique are equally valid for the radiography of textiles. The problems encountered when attempting to record variations in ceramic composition and structures at low kV are not dissimilar and Carr and Riddick (1990) provide a detailed account of essential concepts and procedures to optimise image quality. They assess both medical and industrial X-ray facilities and, although there have been many technical developments during the past decade particularly in the fields of mammography and digital capture, many of their conclusions are still valid. However, none of these materials usually requires beam energies below about 25 to 30 kV. The very lowest energy X-rays, between 5 and 30 kV are called Grenz rays (see Chapter 6, p. 94, Note 1). They are particularly useful for radiographing low atomic weight organic materials, producing very high contrast images as these X-rays are very readily absorbed. Graham and Thomson (1980) examine in detail the history, production and practical applications of Grenz rays. In forensic science they have been used in the examination of documents, counterfeit bank notes, the authentication of stamps, the distribution of gunshot residues on textiles and the detection of additional adhesives on envelopes and packages that have been opened unlawfully and resealed. Other uses include the examination of biopsy specimens, post mortem samples and the morphological studies of small animals and botanical specimens. In cultural material radiography, only the imaging of paper and textiles utilises these very low energy X-rays. Daniels and Lang (2005) discuss several radiographic techniques, including Grenz rays, used in the study of historic documents and works of art on paper. Not all of these techniques are applicable to the study of textiles. Papers are generally flat and limited in size so textiles present imaging challenges not encountered in paper.

Choosing X-ray equipment and facilities X-ray unit specification Industrial X-ray units designed for high definition imaging are best suited for working with textiles. The unit must be capable of operating down to 5 kV or 10 kV and ideally have an upper operating limit of at least 120 kV to allow techniques such as lead screen intensification to be used (see pp. 41–43). Modern mammography units can produce very good results with small objects requiring exposures from around 16 kV to about 35 kV (see Chapter 1 p. 9, Figure 1.5) but general medical X-ray equipment is not suitable for low kV imaging. Where beam energies over 40 kV are required, and appropriate industrial equipment is not available, medical equipment might prove useful. However, the image contrast and resolution obtained will not be as good as that which could be gained with a suitable industrial system. An X-ray tube with a high output in the low kV range will help to keep the exposure times relatively short. A high atomic number target, such as tungsten or molybdenum, will produce a high percentage of low energy X-rays but these will be filtered out of the beam if the material of the tube window is inappropriate. A thin beryllium window (atomic number 4) is ideal as this will transmit almost the entire X-ray spectrum above 2 kV. A 1 mm thick beryllium window will transmit over 90% of 5 keV radiation and practically 100% of radiation above 8 keV (van Aken, 2002: 105, Figure 2). Windows of half this thickness are not uncommon in equipment designed for low energy imaging. A small effective focal spot is also important because of the fineness of detail in textile structures. At higher energies the effective focal spot is not so critical because geometric unsharpness can be reduced by increasing the focus to film distance (FFD) (see Chapter 2, p. 18). This reduces the beam intensity, as the X-rays diverge further with increasing FFD. Increasing the exposure duration by the square of the increase in FFD compensates for this. However, in lower energy imaging this no longer works because the column of air in the beam’s path also acts as a filter, attenuating the beam below about 20 keV. The effect on the reduction in the intensity of the beam increases with decreasing X-ray energy and is particularly marked below 17.5 kV (Graham and Thompson, 1980: 22–25). Increasing the length of the air column increases its filtering power. With a high kV beam, increasing the FFD hardens the radiation to some extent which

26

Textile X-radiography

reduces the image contrast, but this has little effect on the overall beam intensity so the exposure correction for beam divergence alone is sufficient. The effect on a low energy beam is far more dramatic. For example, when compared with the spectrum of the beam as it emerges from the X-ray tube, a modest FFD of 430 mm reduces the intensity of 10 keV X-rays by 25%, 5 keV X-rays by 90% and those with energies below 4 keV are fully absorbed (van Aken, 2002: 105, Figure 2). As a result, increasing FFD at very low kV both reduces the image contrast significantly and also has a marked effect on the beam intensity, far beyond that caused by beam divergence alone. This places a practical limit on the FFD and, therefore, the size of the effective focal spot. Finally, the unit should be capable of continuous running for long periods. When imaging with a 10 or 15 kV beam, low tube output, the effects of air filtration and the use of slow films combine to produce exposures of several minutes duration. Producing single images of large objects by using a long FFD can easily extend the exposure to over an hour (see Conti and Aldrovandi, Chapter 13, p. 187). Selecting an X-ray facility In choosing an X-ray facility for textiles there are additional factors that must be considered. It may require a site visit, prior to transporting the textiles, in order to answer the following questions: ● ● ●

●

●

Is the work environment suitable for the textiles? Is it possible to access the building with the object? Are there physical restrictions to the X-ray enclosure? Can all parts of the object be brought under the X-ray beam centre? Will it be possible to support the object in appropriate positions for radiography?

Some industrial radiography applications require working environments as clean, and possibly as sterile, as medical radiography. At the other extreme, they could be heavily contaminated with dirt, metal swarf and oil. Even in the relatively clean conditions of a museum X-ray facility, it would be advisable to assume that surfaces could be covered with dust from previous work and to protect the textiles by isolating them from the supporting surfaces with, for instance, tissue paper. In certain instances, such as when radiographing archaeological textiles or concealed garments, the objects may be quite soiled and it may not be appropriate to introduce these into some clean

working environments. If unencapsulated film is to be used, the X-ray unit needs to be housed in an area which can be blacked out and illuminated to an appropriate level with safelights. Unless it is safe for personnel to remain in this area during exposures, the door to the enclosure will need to incorporate a light trap so that personnel can come and go without compromising the film. With textile objects wider than a standard doorway or which are awkwardly shaped and cannot be tilted, folded or rolled, access is an important issue. Accessing the building where the X-ray facility is housed may not be problematic but internal doorways, lifts, stairs and limited turning spaces on landings may all need to be negotiated between the point of ingress and the X-ray unit enclosure. Unless the facility is specifically designed for radiographing large objects, the door of the enclosure, and the enclosure itself, also may be of restricted size. The form of the X-ray unit and its enclosure may also provide constraints that influence the usefulness of a facility for X-raying textiles. Small cabinet-type X-ray units are commonly used in museum and laboratory settings because they are fully shielded to contain the radiation they produce and safety interlocks ensure that they can only be used when the cabinet door is shut (Figure 3.2a). The X-ray tube is fixed in position in the roof of the cabinet facing downwards. The object is placed on the film on the base of the cabinet or on a shelf which can be inserted at different levels to vary the FFD. The main drawback with this type of unit, however, is that the cabinet limits the size of object that can be X-rayed. Even if the object will fit into the cabinet it may not be possible to bring all areas under the beam centre, unless the object can be folded over rollers or similar supports. At maximum FFD the beam coverage may still have quite a small diameter which limits the size of film that can be used. However, a large percentage of the objects X-rayed during this research project, from shoes and hats to embroideries and archaeological fragments, easily fitted into the chamber of a Faxitron X-ray unit, model 43855 (internal dimensions approximately 370 mm high by 390 mm deep by 460 mm wide). Free-standing X-ray units do not have built-in shielding and are usually housed in a shielded enclosure such as a substantial brick and concrete structure, a lead-lined room or cabin from which all personnel are excluded during X-ray exposures. The thickness and materials used to form the enclosure will be dictated by factors such as the maximum X-ray energies generated and beam direction. Shielding is heavy

High definition X-radiography of textiles: methods and approaches 27

(a)

(b)

(c)

(d)

Figure 3.2 Different types of industrial X-ray units, (a) cabinet, (b) free-standing stand mounted, (c) ceiling mounted, (d) mobile head. (© Sonia O’Connor, University of Bradford.)

28

Textile X-radiography

and expensive so the enclosure and its door may be as small as the intended purpose of the X-ray facility will allow. Figures 3.2b and c show two free-standing industrial X-ray units, both from facilities designed for X-raying cultural material. These are more versatile than the cabinet units. In both the head containing the X-ray tube can be raised or lowered so that the FFD is continuously variable and has a greater range. A bigger maximum FFD means that bigger films can be used and the image distortions associated with very three-dimensional objects can be largely ameliorated. Usually the head can also be tilted to some extent which makes it possible to Xray objects at angles other than directly from above. This can be very useful with complex pieces, such as upholstered furniture, particularly where rotating the object might cause damage. A stand-mounted unit of the type in Figure 3.2b can be raised to about 2 metres above the ground, giving quite a wide beam. However, the arm supporting the head limits its reach and it may not always be possible to bring the centre of a large object directly under the beam. This may be overcome if the object and film can be supported vertically and, if local safety rules allow, the X-ray head is rotated so that the beam is horizontal. Not only will the stand not be in the way when placing the object but, depending on the dimensions of the enclosure, a longer FFD might be obtainable. A ceiling-mounted unit has the advantage that its maximum FFD is only restricted by the height of the room and there is no stand to restrict the movement of the object being radiographed. The object can be supported on a variable height table or the X-ray head can be raised or lowered from the ceiling, as in Figure 3.2c. X-ray units with fully mobile heads are produced for a number of applications. Figure 3.2d shows an Art Gil X-ray unit made by Gilardoni for X-raying paintings. The X-ray head can be supported at any angle and is connected to the power and control unit in the adjoining room by a flexible cable. This degree of mobility means it is not necessary to keep moving an object to bring each area under the beam centre. Instead the X-ray head can be moved around the object. This can also facilitate X-raying objects from below, as in Figure 3.2d. Rather than the film having to be inserted below the object, the film is placed on top and, being in full view, is easier to place accurately when several exposures have to be taken to provide full coverage. Inevitably there are some situations where it is impossible to move an object from its collection or the conservation workspace to an X-ray facility. In

such instances it may be possible to bring a mobile radiographic unit to the object. Gill’s case study on the radiography of upholstered furniture (see Chapter 14) shares her experience of working with a specialist medical X-ray team whose portable equipment is designed to be used in the patient’s own home. Portable industrial equipment suitable for working at similar, lower or higher kVs is also available. This is usually hired with a specialist team who will undertake the health and safety assessments and submit the work plan to the appropriate authority (in the UK this will be the Health and Safety Executive) for approval before the radiography is undertaken. However, because of the risks inherent in using ionising radiation in an environment not designed for its safe use, the assumption is that objects will normally be taken to the X-ray facility unless a strong case can be made to the contrary. Testing radiographic facilities To assess the sensitivity of an X-ray system, it is worth testing it by taking a radiograph of an object of which the properties are known and which closely match the character and range of materials to be studied. From the outset of this research project, a 1950s plastic doll has been used to evaluate the sensitivity of the equipment and techniques being investigated (Figure 3.3a). This doll was chosen on account of the variety of materials involved and because, as a non-museum object, it could safely be explored to confirm the image interpretation. The doll is dressed in synthetic fabrics with an underskirt fashioned into a bag to contain dried lavender. The radiograph in Figure 3.3b was taken using a Faxitron model 43855 at 20 KV and approximately 2 mA on Agfa Structurix D4 film with an aluminium foil filter and an FFD of 550 mm. It spectacularly demonstrates the ability of radiography to provide a wide range of detailed information such as the: ● ●

● ● ●

● ● ●

●

construction of the clothing weave structure of fine fabrics and net (Figure 3.3c) decorative plastic sequins and glass beads concealed metal pin moulded plastic of the doll with mineral filler and details of joints and hollow interior (Figure 3.3d) shut-eye mechanism in the head detached legs secured in the ‘bag’ underskirt remains of the elastic band securing the arms to the torso dried lavender flower heads (Figure 3.3e)

High definition X-radiography of textiles: methods and approaches 29

(c)

(a)

(d)

(b)

(e)

Figure 3.3 The test doll, (a) photograph, (b) good quality, high definition radiograph and radiographic details, (c) fabrics, beads and sequins, (d) hollow moulded arm, (e) dried lavender flowers. (Private collection; © Sonia O’Connor, University of Bradford.)

30

Textile X-radiography

This image was the benchmark against which other radiographs of the doll, acquired at other facilities and with other capture systems, were measured during the project. Other useful test objects might include an upholstery sample for higher kV work or a fine weave, undyed cotton or unweighted silk fabric for very low kV work. For evaluating computed tomography, a three-dimensional imaging system, a suitcase containing a range of textile objects was also used (see pp. 54–5).

Practical approaches to textile radiography Depending on the materials involved, the radiography of textile objects may involve very different X-ray beam energies and radiographic techniques. The selection and application of these are discussed here and this chapter builds on topics introduced in Chapter 2 (‘Principles of radiography’). Perhaps the first question to consider is why the radiographic study is being undertaken. This will determine which equipment, techniques and viewpoints are most useful and whether radiography is even appropriate. When to take radiographs of textiles Radiography should be as much a part of the study and primary recording of textiles as visual examination, photography and microscopy. Each provides information but it is only through combining the results of all these investigative techniques that understanding is advanced. The examples explored in Part 2 illustrate the ability of radiography to show features more clearly than can be seen by the eye, to reveal information that cannot otherwise be gained without at least partial deconstruction of an object and to initiate completely new lines of enquiry. There may be several stages in a conservation project when it is appropriate to use radiography. In many of the case studies in Part 3 radiography was used to record construction and condition and to provide information vital to the formation of the conservation plan. It can also be used to review the efficacy of past treatments, to record and quantify changes during conservation and to detect changes over time. For example, the radiograph in Figure 9.24c (see p. 144) is of a seam in which the stitches have been pulled and extended. Comparison with radiographs taken before display would have shown if this was due to stresses caused by the display method or if it was already present.

It is worth noting that certain views may no longer be obtainable if an object is not radiographed until after conservation. Image quality can be compromised significantly because the image of mounts or supporting textiles will become superimposed on that of the object. Even a single layer of paper, card or closely woven fabric will produce a reduction in image contrast. Open weave or net textiles will not reduce the general contrast of the image but they will become confused with the image of the object. The image of the silk net support in the double layered mount in Figure 3.4a is quite faint in the radiograph because, compared to the archaeological wool fragment, it is very radio-lucent (Figure 3.4b). Although the eye does not discern the image of the net against the wool, it will have reduced the clarity of the detail of the fragment. The card, padding and textile layers of conservation mounts can produce such a dramatic loss of image contrast that it may not be possible to see details such as stitches without digital image manipulation. Where the material in conservation mounts or traditional frames has a distinctive structure, such as wood and corrugated plastic board, the image of this will obscure that of the object. Both these problems are illustrated by the radiography of the raised work embroideries from the collections of the Ashmolean Museum, Oxford, although better contrast would have been gained for some of the pieces if a lower energy beam could have been used (see Brooks and O’Connor, Chapter 19). Protecting and supporting the textile As when laying out a textile on any surface, it is advisable to place clean tissue paper below the object to be X-rayed. This isolating layer can also reduce direct handling of the object as its position is adjusted under the beam. However, wherever possible the film should be placed between the paper and the textile as otherwise folds, creases, edges and tears in the paper may be visible in the image. Even a single layer of tissue paper will significantly attenuate the beam, particularly when imaging with unencapsulated film at very low energies, such as 15 kV. Similarly, when an object cannot be laid flat, the film should be placed between an object and its support. Selecting most appropriate views With single layer or thin, flat multiple layer textiles (such as bed coverlets) it is usually only necessary to take a single image of an area of interest as the object is, to all intents and purposes, two dimensional.

High definition X-radiography of textiles: methods and approaches 31

(a)

(b)

Figure 3.4 Archaeological woollen fabric fragment supported between two layers of net in a card mount, (a) photograph, (b) radiograph. (YAT 1972 .12 Cat. No. 669; © Sonia O’Connor, University of Bradford; reproduced by permission of York Archaeological Trust.)

Information from all the layers will be captured. Whether the object is radiographed from one side or the other should make no discernible difference to the clarity of detail or measurements taken from the image. It is only when the depth of the object exceeds about 5 mm that the consequences of the conical projection of the beam make it worth considering which surface should be closest to the film. With increasing depth, features furthest from the film will be magnified and less sharp than those at the film surface and it may be useful to take images from both sides to aid interpretation. If the object is not flat, the side placed against the film should normally be that which affords best contact. The radiography of three-dimensional objects is discussed in more detail on pages 45–6. The fewer layers there are in an object, the better will be the rendering of detail and the easier the image interpretation. When trying to characterise stuffings and fillings using radiography, the best definition of such features will be found in the thinnest areas (see Chapter 8, Figures 8.8a and b, p. 114). The sensitivity of the image, that is its ability to show the detail of a textile, decreases with an increase in thickness and numbers of layers. This is because the difference in contrast between an individual yarn and the

density of the general background against which it is viewed gets less and less (Figures 3.6c and d). To reduce the complexity of images, objects should be laid out for radiography so that the beam passes through as few layers as possible. The radiograph of a jacket left buttoned up will have superimposed on it the features of the overlapped right and left front sides and the back, including outer fabrics, interlinings, facings, pockets and linings. However, unfolding an object may not always be appropriate, as in the case of concealed garments (Eastop and Brooks, 1996), where the folds are significant or the purpose of the radiography at that stage of the study is to record the folding and soiling, and any objects concealed among the folds. If an object cannot be opened out flat, inserting film between layers may enable one side to be radiographed at a time. This technique was used to record the detail of an eighteenth century, metal thread embroidered stocking (see O’Connor and Brooks, Chapter 16). Film was cut to size and the corners were rounded to avoid them catching on the interior. The film was then inserted between the upper surface of the stocking and the supporting roll of tissue paper. Between exposures the stocking was turned so that different areas were placed between

32

Textile X-radiography

the beam and the new strip of film. A similar approach was used to investigate the two sides of a sixteenth century mitre (see Conti and Aldrovandi, Chapter 13). Consistency of image layout With no colour, texture and right or wrong side, a radiographic image of a textile may look surprisingly different from the object itself. It can become a time-consuming process to determine right from left, top from bottom, the back of the film from the front or even which image belongs to which object. The use of small lead numbers and letters to provide a unique identifying code on the image and some consistency in layout will overcome these problems. It is a quick and easy matter to check that these are the right way up and the right way round. The lead characters used on the radiographs taken for this project were usually placed at the top right-hand corner of the film or as near as possible to this position if the object covered that area (Figure 3.5). Determining image quality Textile objects can contain such a variety of materials requiring quite different exposures so finding an IQI (see Chapter 2, p. 19) that will suit all situations is not possible. Simple iron or copper alloy stepwedges such as those in Figure 2.6 are appropriate when working at higher kVs, as when imaging metal threads, mechanisms, or wire armatures (Figure 3.5) but are of little use below 60 kV. Graham and Thomson (1980: 22) developed a stepwedge for their experimental work with Grenz rays made from eight 1 mm strips of Perspex held together with Tensol, a suspension of Perspex in acetone. This stepwedge would make a very useful IQI for the more substantial textile objects when working with energies between 15 and 30 kV, and possibly higher. For this textile research project IQIs were made by attaching layers of white cotton tape or layers of unbleached, undyed, Habouti test silk to photographic slide mounts, so that there were areas of one to four layers of fabric (Figures 3.6a to d). These proved very useful when taking exposures around 15 kV. Placed next to the object they gave a record of the contrast, exposure latitude and sharpness (Figure 3.6b). However, when placed on the object they gave a guide to the sensitivity of the image. Figures 3.6c and d show the same cotton tape IQI on two identical exposures of different areas of the same object: a feather and down-filled bedcover.

Figure 3.5 Radiograph of jockey doll; see Chapter 22. (YORCM: 31.53; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

The filling has settled unevenly. In Figure 3.6c the exposure was taken through the cover and lining fabrics, a total of four layers. The image of the tapes in the IQI is detailed and has good contrast. In Figure 3.6d there is also at least 30 to 50 mm depth of filling in the cover. As a result, it is no longer possible to distinguish the fine detail of the textiles of the IQI as the sensitivity of the image is reduced. The data to record Whether taking or commissioning textile radiographs, keeping good records is essential. Details of each exposure and its outcomes will prove invaluable in refining technique, predicting exposure parameters in future work and for briefing radiographers who are not experienced with this sort of material. Table 3.1 lists the categories of information that are most usefully recorded.

High definition X-radiography of textiles: methods and approaches 33

(a)

(b)

(c)

(d)

Figure 3.6 IQIs made from textile and plastic photographic slide mounts, (a) photograph; radiographs on, (b) film background, (c) layers of cotton textile, (d) four layers of cotton textile and a filling of down and feathers. (© Sonia O’Connor, University of Bradford.)

34

Textile X-radiography

Table 3.1 Radiography record Subject

Contents

Reason

Radiograph number

Numeric or alpha numeric code

Date

Date taken

Object identifier

Object name and client’s object number or conservation record number Brief account of object type and materials, including any supports

Running list of unique numbers is the easiest way to tie up image with original handwritten record made as the images are taken and the more formal searchable database Can be useful when troubleshooting, i.e. if processing chemical activity changes Objects can look extraordinarily different on a radiograph Always useful when looking through records for inspiration as to what exposures might work well with a new object. A note of the position and material (tissue paper, foam plastic, textile, etc.) of any supports used during radiography will greatly aid image interpretation All units will perform differently so same exposure parameters may give different results Needed to relate exposure to quality of result, to allow a good exposure to be repeated as closely as possible on a similar object or to translate the exposure to a different facility These will also greatly influence the quality of the resulting image and may influence image interpretation These will also greatly influence the quality of the resulting image and may influence image interpretation. If a filmless capture technique is used that could also be recorded here A guide for future exposures and useful when troubleshooting and in image interpretation Tracking film location (film sent to client/no film – digital copy only, etc.). A place to state the unusual

Object description

Name of facility

Place and X-ray unit details

Exposure parameters

mA, kV, time, FFD, effective focal spot size (if variable)

Filters and intensifiers Filter characteristics or use of lead screens

Film and enclosure

Film type/make. Unenclosed or enclosed film – details of enclosure (paper wrap, plastic or aluminium cassette, etc.)

Result

Image quality, problems, visibility of any supporting materials Anything thought relevant

Notes

Films and film enclosures Ultra fine-grain industrial X-ray film is the best film for capturing the detail of textiles. This film will produce high definition images with good exposure latitude and contrast and will allow examination of the textile’s image under magnification. Film of similar specification to Agfa Structurix D4 and Kodak MX125 (Halmshaw, 1995: 82–85), or better, should be used. In the event that only a mammography unit is available, it may not be possible to use industrial films both because of the extended exposure times they require and incompatibility with high speed film processors. In addition, the radiologist will know how to get the best results with this equipment

and the mammography film and may not be familiar with the characteristics of industrial film. The results will not be of the quality obtainable with industrial film and their usefulness will depend to a great extent on the nature of the object and the purpose of the study. When imaging textiles at very low energies, such as below 20 kV, the film should be used unenclosed with the bare film in direct contact with the object as any cassette or film envelope will attenuate the beam so reducing the image contrast. This can only be done if the area around the X-ray unit can be blacked out to the same standard as a processing darkroom and only lit with a safelight. Where this is

High definition X-radiography of textiles: methods and approaches 35

not possible, the film enclosure used must absorb as little as possible of the X-ray energies transmitted by the object. The paper or plastic sleeves of the prepacked films do reduce image contrast but usable results may be gained down to about 15 kV. However, when the resulting image is magnified, the paper’s structure will be seen superimposed on the detail of the textile so reducing image definition. Graham and Thomson (1980: 15–20) describe how to use a laminator to produce enclosures for film that are ideal when taking radiographs as low as 5 kV. In a darkroom, place the film in a black plastic bag before passing this through a laminator and sealing the package between two layers of polyester film coated with heat reactive adhesive. When done, the package is flat, light-tight but very radio-lucent. Graham and Thomson were working with flat paper specimens, such as bank notes, which could be sealed into the bag with the film. As the air was expelled from the package during lamination, this gave the closest possible contact between the specimen and film and produced very sharp images. This might prove a useful approach to imaging fabric samples such as those used for the study of historic silks by Brooks et al. (1996) but most textiles must be placed on top of the package because of their size or shape and because of the risk of damage. If black bags are used, on their own or with a laminator, the material and quality of the bag are important. Black polythene is suitable but polymers containing chlorine (atomic number 17), for instance polyvinyl chloride (PVC), will absorb the low energy X-rays. If the thickness of the sheet forming the bag varies this can produce bands of different density across the radiograph. Reusable flexible plastic film cassettes are not recommended for use below 30 kV as they are very much thicker than bags and significantly reduce image contrast (Figure 3.7a). Some may be fabricated from PVC. In addition, these plastic envelopes are often textured and this can produce an image on the radiograph at low energies, visible even without magnification. In Figure 3.7b this is very prominent and degrades the image of the textile. Daniels and Lang (2005: 101) suggest using transparent red gelatine or cellulose acetate sheet (the same colour as used in safelights) to provide a window in the film cassette which will transmit very low energy X-rays without degrading the image. The radiograph in Figure 3.7c shows dark and light banding across the fringe of a pattern woven woollen fabric. This banding does not relate to variations in the textile but to the distortion in the plastic

cassette in Figure 3.7a. This cassette had been stored folded over in a drawer because of its size and consequently would not lie entirely flat, even when loaded with the film and weighted down by the textile. As a result the X-rays have passed through a greater thickness of the plastic on either side of the fold where the surface of the cassette was not perpendicular to the beam direction. This has produced the bands of increased beam attenuation. Film processing Low energy radiographs seem to be very sensitive to poor processing. Changes in image quality caused by deterioration of the processing chemicals, variations in temperature or development time which might have remained undetected by casual inspection in higher kV images seem to be greatly exaggerated at these low energies. This is particularly noticeable when imaging very thin specimens as the background image densities can be quite low. A gradual drop in the image density from the edge to the centre of the film and diffuse, waving bands of variation in density have both been observed. Producing even development of such low kV exposures when dish developing seems to be particularly difficult (see Chapter 5, pp. 88–9).

Determining correct exposure parameters This piece of reverse appliqué was radiographed at two different beam energies (Figure 3.8). The radiographs in Figures 3.8b and c were taken at 25 kV and 15 kV respectively. The difference in contrast and the effect this has on the visibility of the detail is selfevident. The low contrast of Figure 3.8b cannot be improved by increasing the X-ray dose. More X-rays will reach the film during the exposure but the proportion of the beam transmitted through each area will not change so the image will just be darker. The key to producing good quality radiographs is first to select the beam energy most appropriate for the object and then to determine the correct exposure. Test exposures Determining the correct beam energy and exposure need not take significant amounts of time or film as lead sheeting can be used to produce several images of an object, or part of an object on one film. Lead sheets are used to shield the film and are moved around so that each area is only exposed once to

36

Textile X-radiography Figure 3.7 Flexible plastic film cassette distorted by folding, (a) photograph and radiographs showing (b) image of cassette surface texture, (c) image of fold in cassette. (© Sonia O’Connor, University of Bradford.)

(a)

(b)

the X-ray beam. When doing test exposures it is best to alter only one parameter of the exposure at a time. First, select a tube voltage and take several exposures, increasing the X-ray dose each time, until all areas of the film have been exposed (Figure 3.9). Experience will suggest a reasonable starting point for a particular type of object – a ‘guesstimation’ based on the variations in thickness and the materials

(c)

involved. As the X-ray dose is a combination of the tube current and the duration of the exposure, if the current is left the same, the X-ray dose is simply controlled by altering the exposure time. If the ‘guesstimated’ exposure time is several minutes long there is little point in increasing or decreasing the time by a few seconds, as the difference in each image will be minimal. More usefully, the time should be halved or doubled.

High definition X-radiography of textiles: methods and approaches 37

(b)

(a)

(c)

Figure 3.8 Detail of an ethnographic band decorated with reverse appliqué, (a) photograph and details of radiographs taken at (b) 25 kV, (c) 15 kV. (Private collection; © Sonia O’Connor, University of Bradford.)

Once the film is processed select the ‘best’ image, examine it critically and make a note of the exposure duration. In Figure 3.9 the 1 minute and 2 minute exposures are underexposed and the film has a general lack of density – the image is very light and its background light grey. The 8 minute exposure is overexposed and the film is very dense – the background is very black and even the most radioopaque areas of the object have produced quite a grey image. The 4 minute exposure has produced the best result – the background is black and there is a wider range of densities in the image. The characteristics of X-ray film emulsions mean that both over- and underexposure will cause a loss of image contrast. If all the images are over- or underexposed, a second test film is made with either shorter or longer exposure times until the correct exposure duration is identified. Once a little experience is gained, this is unlikely to be necessary unless faced with an unfamiliar X-ray unit. Remember also that any change of FFD, focal spot size, filtration, film type or film enclosure will need to be compensated for by changing exposure duration. If none of these factors seems to be the cause of such a gross miscalculation, it is worth considering if there is a problem with the film processing before assuming that this is

purely due to unrecognised factors relating to the object itself. The second stage of the process is to determine how good the image contrast is of the ‘best’ image on the test exposure. A good quality image will show detail in both the thickest and thinnest areas of an object. If the ‘best’ image is very close to this, then no further test is required and the correct exposure parameters can be selected for the textile with little or no alteration. But if the image contrast is too high or too low, another test exposure may be required at a different kV. The test film at the new kV should use the exposure duration of the previous ‘best’ exposure as the central point for the exposure time range. Changing the kV of the beam will alter the intensity of the beam so the time or tube current needed to maintain the same general radiographic density will be less if the kV is raised or more if the kV is reduced. Figures 3.10a to c are three exposures of the test doll. If the image has very low contrast all the detail will be lost in similar shades of grey, as in Figure 3.10a, and the exposure needs to be repeated at a lower kV. If the image has very high contrast, as in Figure 2.10b, the features will be either black or white with few shades of grey and the exposure test

38

Textile X-radiography

needs to be repeated at a higher kV. The more heterogeneous an object, the more difficult it is to record the detail in all areas in a single exposure but the ‘best’ exposure is the one that provides the most information – particularly in the areas of interest. Figure 3.10c has recorded the most detail of the hidden contents and layers of the doll and is the ‘best’ image of the doll as a whole, even though Figure 3.10a provides a better record of the netting and ribbon of the bonnet. At very low voltages a change of one or two kV can make all the difference. Above 15 kV subsequent tests might be done at 5 kV increments. In general, unless the original ‘guesstimation’ was off the mark, two test images should provide enough information between them to determine the best exposure parameters for the textile object. Exposure meters

Figure 3.9 Radiographic exposure test of the test doll. Anticlockwise from the top right: 1, 2, 4, and 8 minute exposures at 15 kV. (Private collection; © Sonia O’Connor, University of Bradford.)

(a)

(b)

Devices are available which help reduce the guesswork of exposures, such as the M-Gil contrast and exposure meter by Gilardoni. This is an electronic dosimeter which is commonly used in the radiography of paintings and which proved useful in the radiography of bed quilts and coverlets (Figure 3.11). This dosimeter has three exposure probes that detect the intensity of the radiation passing through an object. With paintings all three probes are generally placed on areas of possible differing X-ray attenuation, for instance areas of white paint, dark paint or over the stretcher bar. During a test exposure the

(c)

Figure 3.10 Details of radiographic test exposures of the test doll, (a) contrast too low, (b) contrast too high, (c) contrast appropriate to subject. (Private collection; © Sonia O’Connor, University of Bradford.)

High definition X-radiography of textiles: methods and approaches 39

meter produces an average intensity reading from the three areas. As the test exposure proceeds, the kV is raised until the optimum intensity is reached on the meter. This optimum reading and the approximate exposure durations required at the kV indicated are initially arrived at empirically by making tests similar to those just described. If the FFD or film type is altered a new optimum reading must be ascertained.1

Thin homogeneous textiles Unless the FFD is very short, the filtering effect of air makes it difficult to produce radiographs of practical exposure duration on slow speed, industrial films below 15 kV with most readily available X-ray equipment. It is possible, for instance, to use the Faxitron model 43855 at 10 kV or slightly lower but the FFD has to be so short that the beam diameter will only cover a small area of the film. Working at 15 kV was not generally found to be a problem during the research project as most objects X-rayed were made of more than one layer of relatively thick fabrics or their transmission of X-rays varied from feature to feature so the penetration power and

exposure latitude of the beam was quite appropriate. Synthetic textiles showed a wider range of radioopacities than natural organic fibres. Colourants, mordants and other surface treatments could all affect the image contrast and exposure duration. However, some of the very fine fabrics, such as undyed, unweighted silks, produced very low contrast images in which the weave structure could not really be discerned without contrast adjustment of the digitised image (see Chapter 4, pp. 69–70). Graham and Thomson (1980: 25) recommend that a vacuum chamber is introduced between the X-ray source and the object to allow radiography down to 5 kV. McClung (1964) experimented with replacing air with helium (atomic number 2) in a chamber with a relatively X-ray transparent polyethylene window and unencapsulated film at 5 to 10 kV. A heliumfilled chamber at approximately atmospheric pressure has the advantage over a vacuum in that it is easier to maintain with X-ray transparent windows for transmission of the beam. The gas used is pure helium, and not the air/helium mixture supplied for party balloon inflation or the oxygen/helium mixture for divers. This has been used in forensic radio-graphy (Rendle, 1993) and van Aken (2002) discusses the practicalities of using helium for the radiography of historic papers. Objects which are not flat could only benefit from this technique if they could be introduced into the helium chamber with the film. Although it is very difficult to translate exposures from one piece of equipment to another, as a general indicator, appropriately exposed radiographs of the seam samples in Figure 9.19 were produced on Agfa Structurix D4 film, in air, using a Faxitron model 43855 with an FFD of c. 550 mm (lowest shelf position) at 15 kV, c. 2 mA and an exposure duration of 3 minutes.

Layered and more complex textile objects

Figure 3.11 Two exposure probes of the M-Gil contrast and exposure meter by Gilardoni. (© Sonia O’Connor, University of Bradford; reproduced by permission of the Quilters’ Guild of the British Isles.)

The more layers of textiles, the longer the exposure will need to be to ensure that the image is not underexposed. Exposure durations from 2 up to 8 minutes served to produce excellent images of a whole range of textile objects from bed coverlets and bodices to embroideries and purses at 15 kV. When objects with a greater variety of radio-opacities were encountered, due either to variations in thickness or materials, the exposure latitude at 15 kV was not adequate to allow detail to be captured in all areas, but this was

40

Textile X-radiography

often remedied by even a small increase in beam energy.

Thicker textiles With increasing thickness and the inclusion of denser layers, such as leather or thick card, the ability of a 15 kV beam to penetrate textile objects is quickly diminished and exposure times become excessively long. Graham and Thomson (1980: 22–25) showed through experiment that the exposure duration needed to produce a specific image density at a fixed FFD is 15 times longer at 15 kV than at 20 kV, the increase in exposure being particularly marked below 17.5 kV. The longer the exposure, the more scatter will be generated within the object and this will increasingly degrade the image. There comes a point, then, where it may become more difficult to discern the detail in a long 15 kV exposure, for instance, than in a shorter 18 kV exposure of the same object, despite the lower image contrast. The object may be so thick or dense that almost all X-rays of 15 kV or below will be absorbed before reaching the film. Even thin layers may attenuate a low kV beam completely if they have greater amounts of higher molecular weight components than are generally found in organic fibres, such as chlorine or calcium. Raising the kV to 20 or even 25 kV was sometimes necessary to penetrate such objects.

(a)

(b)

Heterogeneous textiles When imaging textile objects that vary in thickness or material, very low energy radiography may produce images that have too high a contrast for the subject so that information cannot be recorded in all areas of the image in a single exposure. This is illustrated by the radiographs of the test doll in Figure 3.12a which was a 6 minute exposure at 15 kV. The image contrast is good in most areas where there are three, four, five or more layers of fabrics and in most of the components of the plastic doll. However, the exposure latitude is too narrow, so while the detail of the thinnest areas, such as the net bonnet and its ribbon, is very dark, the X-ray beam has barely penetrated the region where the net of the skirt and fabric of the underskirt are gathered over the lavender flower-filled plastic torso. The thickest areas of the doll’s moulded body, for instance under the chin, have attenuated the beam completely. Raising the kV a little can overcome these problems as it will increase the penetrating power of the X-ray beam and widen the exposure latitude, but at the cost of reducing the image contrast. The radiograph in Figure 3.12b was a 1 minute exposure taken at 20 kV and is a much more suitable beam energy for this subject. The exposure latitude is greater than in Figure 3.12a, the main improvement being in the greater visibility of detail in the thickest areas.2

(c)

Figure 3.12 Details of radiographic test exposures of the test doll, (a) 15 kV for 6 minutes, (b) 20 kV for 1 minute, (c) 20 kV for 1 minute with aluminium filter. (Private collection; © Sonia O’Connor, University of Bradford.)

High definition X-radiography of textiles: methods and approaches 41

A reduction in image contrast ought to mean that it is more difficult to distinguish features in areas of very similar radio-opacity, for instance the detail of the fabric of the underskirt below the torso of the doll. Comparing Figures 3.12a and b, however, this does not seem to be the case. The ability to discern fine detail is influenced by a combination of image contrast and image sharpness. The great reduction in exposure duration from 6 minutes at 15 kV to 1 minute at 20 kV means that there is a consequent reduction in scattered radiation so the detail in Figure 3.12b is sharper than that in Figure 3.12a. This has compensated for the loss in image contrast. However, raising the kV has not really improved the rendering of detail in the most ephemeral areas in Figure 3.12b. The bonnet net and ribbon have hardly attenuated the higher energy beam at all and in places are entirely lost against the black background of the film. Raising the beam energy again to widen the exposure latitude further would only exacerbate this problem. In such circumstances careful filtration can improve the image by changing the shape of the beam spectrum. This is called beam hardening. In preferentially removing lower energy X-rays from the beam spectrum, a filter not only helps to sharpen an image but will also lower the image contrast and widen the exposure latitude. This is demonstrated by the radiograph in Figure 3.12c which was taken at 20 kV for 1 minute using an aluminium foil filter between the X-ray source and the object. The image is less dense overall than Figure 3.12b as no increase in exposure duration has been made to compensate for the general attenuation of the beam by the filter. Nonetheless, the image contrast is noticeably lower in Figure 3.12c than Figure 3.12b but the visibility of detail in all areas is very good. The main difference between the two radiographs is that in Figure 3.12c the weave of the bonnet ribbon is also seen. The radiograph produced with the filtration of the foil has the desirable characteristics of a slightly higher kV image without the problems of increased penetration. At the low kVs needed to image textiles, filtration is a very delicate matter as it is easy to end up reducing the contrast too far. A single sheet of aluminium cooking foil is most useful. This was used taped into a frame of corrugated plastic board which could be introduced into the filter holder of the Faxitron. At other facilities visited during the research project, if filtration proved necessary, it was usually possible to fix a sheet of foil across the path of the beam using adhesive tape or to interleave a layer of foil between the textile and the film. Aluminium foil is very

delicate but cheap and easily replaceable. Flat, not embossed, foil is required; for instance, some are made with a reinforcing diamond pattern. If the foil becomes scored, creased or punctured, it must be discarded or these anomalies will produce image artefacts on the radiograph of the object.

Mixed-media objects The greater the difference in radio-opacity between areas or components of an object, the more difficult it is to capture the detail of the object in a single image. Figure 3.13a is an ethnographic mask made from several layers of textiles and decorated with appliqué, embroidery, mirrored panels, cords and cowry shells. Figure 3.13b is a low kV image which has captured the details of the textile components but only outline shapes of the mirrored panels and shells. At a high kV, Figure 3.13c shows the internal form of the shells and the distinctly different characteristics of the round and rectangular panels but all trace of the textile backing is lost. The round panels may be thin metal disks coated with silver or tin and lacquered to keep them bright while the rectangular panels are cut from mirrored glass. Visual examination indicated that the bright frames to the glass panels were iron but it was not clear if this was a protective edging or the wrapped-over edges of a backing sheet. The radiograph gives no clues on this point because there is not enough penetration of these radio-opaque features. Even at these higher beam energies the exposure latitude is not sufficient to deal with the range of materials present and the variations in thickness in one image. However, Figure 3.13d shows a dramatic improvement in exposure latitude. Not only is it possible from this radiograph to see the edges of the glass mirror and to confirm that the iron is a continuous sheet but the shells are better rendered and even the denser areas of the textile components are visible. This image was captured simply by using lead screen intensifiers. Lead screen intensifiers Lead screens are routinely used in industrial radiography for high definition imaging of defects in metal components and in archaeological conservation for the investigation of metal artefact assemblages. Low energy X-rays readily interact with the film emulsion to produce the required photographic effect but only a few per cent of higher energy X-rays are absorbed; the rest pass straight through and have no

42

Textile X-radiography

(a)

(b)

(c)

(d)

Figure 3.13 Ethnographic shell mask, (a) photograph; details of radiographs taken at (b) low kV, (c) higher kV, (d) 120 kV with lead screen intensifiers. (Private collection; © Sonia O’Connor, University of Bradford.)

effect. By increasing this rate of interaction, the image produced will be intensified so reducing exposure times. In medical radiography fluorescent screens are used to convert some of the X-rays into light before they reach the film. Unfortunately light is

readily reflected which causes optical blurring so making these screens unsuitable for high definition radiography. Lead screens intensify the image by emitting electrons which only travel short distances and are not reflected by the film so do not degrade

High definition X-radiography of textiles: methods and approaches 43

the image to any significant extent. In addition, lead screens produce other beneficial effects that can greatly increase the amount of information visible in a radiograph. Understanding their benefits is easier once the make-up and use of lead screen intensifiers is understood. The screens are polished lead foil filters supported on card or plastic. They are used as a pair inside rigid aluminium film cassettes so that the surface of the lead is held in close contact with the emulsion on either side of the film. Their thickness is matched to the beam energy, for instance at energies between 120 and 250 kV Halmshaw (1995: 132) recommends a front screen of 0.025 mm thickness. A thicker back screen can be used to absorb backscattered radiation but this is not necessary if a lead sheet is placed below the cassette. The internal padding of the cassettes ensures proper contact between the screens and the film. This is essential as the electrons are emitted in all directions and any gap between the film and the screen will allow the electrons to spread out before they interact with the emulsion. Screen damage such as small dents can produce dark spots of fogging on the image. For situations where the film needs to be held against a curved surface, film sandwiched between lead screens and vacuum packed into flexible, opaque plastic sleeves, can be purchased from the major industrial X-ray film suppliers. X-rays that have passed through an object are filtered by the front screen before they reach the film. This removes the lower energy and scattered radiation from the beam so sharpening the image and hardening the beam, improving the exposure latitude of the image. At beam energies of 120 kV and above, both screens also emit electrons in direct proportion to the X-ray flux. These electrons are readily absorbed by the film and supplement the effect of the X-rays, improving the contrast of the image. The result is a high quality radiograph with the contrast of a low kV image and the penetrating power and exposure latitude of a high kV image. Lead screen intensifiers should be used with textile objects whenever metal components are being examined, including metal thread work, but can be equally useful for imaging objects of very diverse materials or those that vary greatly in thickness. This is illustrated by Figure 3.14. This leather-soled, early twentieth century shoe has a leather-lined, metal thread fabric upper but this is difficult to image against the heel. The radiographs of the side and top view of the heel of the shoe in Figure 3.14b have captured some detail of the fabric and revealed internal details of the shoe’s construction. A more

penetrating beam would have revealed more of the heel structure but would have lost the detail of the less radio-opaque materials. Certainly several exposures at different beam energies would have been required to capture the same quantity and quality of information in the single image as was obtained using lead screens (Figure 3.14c). Seams, stitches, fabric, leather, metal, wood and plastic components can all be identified and studied in detail in this one radiograph which also allows the relationship between these components to be recorded.

X-raying ‘special needs’ textiles Large textiles It is tempting with large objects to hope that taking perhaps one or two radiographs in particular areas will be sufficient to characterise the piece, revealing all there is to know about its structure and condition and answer any questions posed. However, wherever an exploratory radiograph was followed by a full survey during this project this rarely proved to be the case. Radiographing whole large textiles can be approached in one of two ways, either by taking several exposures – moving each area beneath the centre of the beam in turn – or by capturing the entire object in a single exposure. The size of the area that can be radiographed in a single exposure may be limited by the size of film available or the area that can be covered by the X-ray beam. A complete radiographic survey of quilts, banners or flags might require several or many tens of radiographs. X-ray facilities designed for the radiography of large paintings, where the object is supported horizontally above the X-ray unit (as in Figure 3.2d), adapt easily to this use. Textiles, unlike paintings on wood or stretched canvases, will not support themselves so additional boards may have to be used. Over the area of the X-ray beam, polythene or polyester film sheet will support the textile with minimal attenuation of the beam. The order in which the radiographs are taken will reflect the shape of the piece and factors such as ease of handling. The process requires careful planning to avoid leaving gaps in the radiographic record and it is helpful to start with a photograph of the object that can be annotated to show which area is covered by each radiograph. Each radiograph must overlap its neighbours by at least 30 mm. Lead numbers placed consistently in the same corner of each film will help with the orientation of each image. As long as the numbers are kept within the area of overlap the

44

Textile X-radiography Figure 3.14 Leather and metal fabric shoe, (a) photograph; radiographs taken at (b) 20 kV, (c) 120 kV with lead screen intensifiers. (TCC 92.10 Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

(a)

(b)

features they obscure will appear on adjacent radiographs. Using this approach to radiographing large objects has advantages. Standard-sized films and film to focus distances are used which makes the estimation of appropriate exposure durations a simple matter to those familiar with a facility. The usual film processing procedures can be followed. However, even with relatively flat objects that are in close contact with the film it is possible to observe the distortions formed by the conical radiation of the X-ray beam which cause features that appear on adjacent radiographs to be displaced or differ in size when the images are compared

(c)

(see Chapter 2, pp. 17–18). In the radiographs of objects with appreciable depth, these distortions can cause considerable problems as the image of features towards the edges and above the film plane may fall beyond the edge of the film. Additionally, textile objects are often very flexible and dimensionally unstable so care must be taken to keep the tension of the piece the same in order to avoid distortion in successive exposures making it difficult to track a feature across adjacent radiographs. If it is necessary to produce a combined image from the individual radiographs, this can be done in the same way as it is in the radiography of paintings.

High definition X-radiography of textiles: methods and approaches 45

First, copies are made of the radiographs on duplicating film. These are then used as internegatives from which contact prints are made. Finally a mosaic is made from the individual prints by pasting them onto a plywood board to produce a single image, which is then photographed using large format film (Padfield et al., 2002). Each stage of this process requires time and great skill, not least the production of the print mosaic as a certain amount of easing and stretching of the prints is required as they are stuck to the support to overcome at least the smaller distortions caused by the conical projection. Even then the result can suffer from tonal differences between adjacent images and a general loss of contrast, not to mention a loss of information at each stage. Storing the full-sized mosaic is yet another problem. Fortunately, converting radiographs to digital images means it is now possible to overcome many of the problems associated with radiographic mosaics (see Chapter 4, pp. 69–71). Large objects which have any amount of depth should be radiographed as a whole in one exposure. The main advantage of this is that, having a single viewpoint, the distortions and tonal variations between areas within the object are reduced. However, to do this the FFD must be increased to provide a beam wide enough to cover the object and consequently these images require very long exposures and equipment that can be run for perhaps several hours. A long FFD also means that very low energy imaging cannot be done, as the filtering effect of the air column will be considerable (see pp. 25–6). The image can be recorded on standard size films overlapped to provide continuous cover of the area required and the resulting radiographs are then combined by mosaicing to produce a single image. In Chapter 15, Conti and Aldrovandi discuss practical aspects of taking radiographs of large objects in this way including using a single, very large sheet of film which is processed while still sealed in its light-tight container. The latter system is described in detail in Aldrovandi and Ciappi (1995). Three-dimensional textile objects The more three dimensional the object, the more the image will be distorted by the diverging X-ray beam and the more features are likely to be superimposed on each other in the resultant twodimensional image (see Brooks et al., Chapter 20, p. 260). This makes it impossible to take accurate measurements of details, other than those in direct

contact with the film, and complicates image interpretation. Increasing the FFD will reduce the problems of magnification and displacement of features but the lower the beam kV the less leeway there will be to do this. Image interpretation can be aided by careful selection of viewpoints and film position but this assumes some pre-knowledge of the nature of the object and its internal structure. Gill considers this in relation to radiographing upholstered furniture (Chapter 12). Objects such as dolls which incorporate mechanisms or internal supports may need to be radiographed from several directions to provide sufficient information to understand their internal structure (Figure 20.6 see p. 256). Many of the lessons learned working with a collection of dolls is relevant to other complex textile objects (Brooks et al., Chapter 20). The three-dimensional nature of dolls and toys makes it essential to radiograph them at least twice, once from the front and once from the side so that the relative positions of features can be determined. Careful placing of the limbs and head is also important in the recording of joints and shut-eye mechanisms and to ensure that as few components as possible overlie each other. For instance, when radiographing a side view, raising the arms will allow a clearer view of the details of both these and the torso. Radiographing shut-eye mechanisms with eyes open and closed might also prove useful in characterising the components. The eyes usually close when the doll is laid on its back but laying it face down should mean that they remain open. Carefully moving the doll from a sitting or lying position onto its side can ensure that the eyes remain open or closed for the side view. When radiographing larger dolls that require several exposures to give full coverage, the relative position of the limbs and clothing needs to be carefully maintained to provide continuity across the images. Because of the problems inherent in producing a two-dimensional representation of a threedimensional object, conventional radiography may not always be the best approach. Stereoradiography (see pp. 51–2) should also be considered but if computed tomography (CT) is available, cost allowing, this is preferable as it can provide fully three-dimensional information about the internal and external features of an object (see pp. 52–6). CT should not, however, be seen as superior to conventional radiography in all circumstances or even the technique of choice. Rather, these are complementary techniques. Taken first, radiographs will provide the fine detail

46

Textile X-radiography

of the textiles and will indicate whether CT is necessary, or even appropriate, to complete the understanding of the object. Metal thread work In the past, with a few notable exceptions, radiography was most often used on textiles for recording metal thread work. In some of these radiographs, the metal threads were revealed as an undifferentiated bright image against a black background which rather limited their information value. Ideally two images should be made of metal thread work; a low kV image to record the organic fabrics of the textile and the location of the metal threads within them, and an exposure at 120 kV using lead screen intensifiers to provide information about the metal threads themselves. The value of this approach is illustrated by the radiography of this highly decorative embroidery on velvet (Figures 3.15a and b). Shaped like a chasuble (a vestment worn by priests when celebrating mass) but only 550 mm long, this single-sided miniaturised version may have been a commercial sample but its origins are unclear. Radiography was employed to help investigate both its construction and materials (Brooks and O’Connor 2005: 170–171). The low energy images, combined in Figure 4.8 (see p. 71) to form a radiograph of the whole object, reveal the internal structure. The detail in Figure 3.15c shows a canvas interlining and the thread attaching the metal embroidery thread passing continuously between similar parts of the designs. The same area, in Figure 3.15d, imaged at high kV using the lead screen intensifiers, shows instead the differences between each metal thread type. Differences in absorbency relate both to the thickness of the material and the atomic weight of the constituent metals (see Figure 10.3, p. 153). Some modern metal threads hardly attenuate the beam at all and may be penetrated even at 15 kV (see Figure 10.8). Metal threads wrapped in gold do not corrode but if the gold is alloyed with silver or copper, or is only a coating on a silver or copper foil, the detail of the wrappings can become obscured by corrosion. Bridgman (1973: 10–11) used radiography to identify wrapped metal threads hidden beneath an area of green corrosion on a cloth bundle from a Native American burial near Rochester, New York. The radiograph revealed the characteristic interlocking triangles formed by the helically wound foil (see Figure 10.4, p. 154), in this case possibly copper with a silver or gold wash.

X-raying archaeological textiles Textiles can only survive in archaeological contexts if conditions inhibit biodeterioration. Desiccation, exclusion of oxygen, very low temperatures or the presence of biocides can all lead to their preservation but may influence the quality of the information that can be gained from their investigation by radiography. Textiles lifted and X-rayed in a block of sediment may only appear as a less dense feature within the matrix. Nevertheless, images may reveal and record the location of non-textile objects attached to the textiles, such as brooches or buckles, or concealed within them. In the case study discussed by Barham, radiography was used to map metal thread work lifted from a burial where the textile had decayed completely, leaving only a stain in the damp sandy ground (see Chapter 25). The amount and nature of the archaeological matrix lifted with the textile remains will make a great difference to the quality of the information that can be gained through radiography. Silica sand is less radio-opaque than, for instance, lime plaster rubble or sediment derived from ferruginous sandstone. The more matrix through which the beam has to pass, the more scatter will be generated and the more the image of objects within it will be degraded. Sediment still trapped in the fibres of the textiles in Figures 9.13c (see p. 136) and 9.14b (see p. 137) shows as bright speckles. Soluble salts crystallised on the surface of textiles would also produce an image. Before cleaning, these deposits would have obscured the weave structure completely. Many textiles recovered from dry environments such as those buried in Egyptian tombs or concealed in buildings (see Barbieri, Chapter 14) may not have been buried in sediment or soil but may still have become heavily soiled. Alternatively, the radio-opacity of textiles might be enhanced if the yarns are impregnated with resins or become suffused with minerals. Permafrost can create very dry environments and textiles preserved in these conditions will produce radiographs every bit as detailed as those preserved in warmer desiccated environments (see Peacock, Chapter 24). However, textiles recovered in ice present different problems as the ice will attenuate the beam quite significantly at low kV. The shape and location of denser objects may be obtained but the image of the textiles may be very low contrast and radiographic exploration of these is best done after the ice has been eliminated. Water is even more dense than ice so although moisture in the yarns of a

High definition X-radiography of textiles: methods and approaches 47

(a)

(b)

(c)

(d)

Figure 3.15 Miniature velvet chasuble, (a) photograph, (b) detail of metal thread work; radiograph of same detail taken at (c) 15 kV, (d) 120 kV with lead screen intensifiers. (TCC IM 17.33 Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

48

Textile X-radiography

fabric will increase beam attenuation, detail in waterlogged textiles may be lost. This is illustrated by the radiography of a thirteenth century drawstring purse found in the waterlogged excavations of Swinegate in York, England (Figure 3.16). Originally a leather purse decorated with silk piping and tassels, the leather has rotted away revealing the silk lining fabric. After the initial cleaning to remove the obscuring deposits, the purse was X-rayed to

investigate two areas of metal concretion and to see if there was anything inside. The detail in Figure 3.16b is from a radiograph taken at 50 kV while the purse was still waterlogged. It shows only the bright granular clusters of trapped sediment and the thicker components of the purse, which appears to be empty. That the textile is visible at all at this kV is largely due to the attenuation of the beam by the water. Although some bright yarns are apparent in

Figure 3.16 A thirteenth century leather and silk drawstring purse (York Archaeological Trust 989.28 ctx. 3204 sf. 267), (a) photograph still waterlogged but cleaned, (b) radiograph detail, waterlogged (© YAT), (c) photograph detail, after drying, (d) radiograph detail, after drying. (© Sonia O’Connor, University of Bradford; reproduced by permission York Archaeological Trust.)

(a)

(b)

(c)

(d)

High definition X-radiography of textiles: methods and approaches 49

the tassels, the surface tension of the water is holding everything close together and no other detail is visible. After controlled drying, the detail of the fabric and construction of the purse could be appreciated (Figure 3.16c). Re-radiographing it at 15 kV not only demonstrated the true potential of the technique for recording and exploring archaeological textiles (see detail in Figure 9.5 see p. 130), it showed that the purse was not entirely empty but contained some seeds (Figure 3.16d). Textiles recovered in situ on human remains are often difficult to radiograph because underlying bone will completely attenuate a low kV beam so that the textiles are only visible between the bones. This was a particular problem for Javér (1997) in her study of a sprang cap which had survived still on the head of a Roman period Egyptian woman. The presence of the skull meant it was only possible with radiography to see anything of the cap at the back of the head where there was a tangled bundle incorporating a knotted cord. Elsewhere the cap was only a halo around the edge of the head which gave only an oblique view of the structure. Today, computed tomography (CT) can be used to produce radiographic images from which the underlying bone can be digitally removed, allowing the textile layers to be seen and explored in three dimensions (see p. 54). Textiles may also be preserved in archaeological contexts where they become infused with inorganic compounds derived either from mineral deposits, such as potassium alum, or from corroding metals

(a)

(Janaway, 1987). These are known as mineral preserved organic (MPO) remains. Radio-opacity of such textiles will be enhanced to a greater or lesser extent depending on the contaminant and its concentration. Solutions from corroding copper or lead alloys are biocidal and organic materials which become contaminated with them may be preserved in what would otherwise be considered adverse environments. The pallium of Godfrey de Ludham, Archbishop of York (d. 1265), is one such example. This rare survival was found when his tomb in York Minster was opened in 1969 (Ramm, 1971: 131–132). Tablet woven in wool and decorated with three silk Maltese crosses, the pallium was very degraded (King, 1971: 136). When radiographed by English Heritage in 2002, the textile was unexpectedly visible on images taken at 90 kV. This was because the textile was impregnated with lead salts from the coffin and from the weights sewn into the ends of the pallium. Even if the organic material continues to decay, its structure may be preserved if the corrosion precipitates, forming a mould or cast or a combination of the two (Janaway, 1989). This is most common where the organic material is in close contact with corroding iron (Figure 3.17). However, such remains, by the nature of this preservation mechanism, are often found attached to the surface of metal objects and obscured by layers of corrosion so that the details are often obscured to visual examination. Experimental radiographic work using micro-computed tomography (see pp. 54–5) has

(b)

Figure 3.17 Sixth to seventh century Saxon copper alloy buckle from St Peter’s, Kent. Photograph, (a) front, (b) detail of back of buckle loop with MPO textile on corroded iron buckle pin. (© Sonia O’Connor, University of Bradford.)

50

Textile X-radiography

shown that it is possible to explore MPO textile remains in detail while still in situ. Blumer et al. (2005) and Nowak-Böck et al. (2005) have used this technique to produce virtual sections and threedimensional images of multi-layered MPO textiles, from which they were able to determine weave structure, direction of spin and to make thread diameter measurements.

Special radiographic techniques The information obtainable by conventional radiography can be supplemented by a number of special techniques. Electron transmission radiography and β-radiography Both these techniques are used in the study of historic paper and their application and relative merits are discussed by Daniels and Lang (2005). In electron transmission radiography, lead foil is used to generate electrons in the same manner as lead screen intensifiers. However, the object is placed between the foil and the film in a high energy X-ray beam, heavily filtered to remove all X-rays below at least 150 kV. The electrons from the lead produce the high contrast image of the object while the remaining X-rays are too high energy to expose the film significantly and produce only a weak image. β-radiography also utilises electrons to form an image on radiographic film but these are in the form of β-particles produced by the decay of a radio-isotope. In paper radiography, the most commonly used β source is carbon 14 in the form of a plastic sheet. Exposures may take a day to complete because of the low output of β-particles and, although this is a conveniently portable source, for health and safety reasons its use is very highly regulated through the Ionising Radiations Regulations (The Stationery Office, 1999). Both electron transmission and β-radiography produce very high contrast images allowing watermarks and paper structure to be recorded but their usefulness in the imaging of textiles is limited to thin, relatively small and flat items. To produce a sharp image and to ensure that the air does not unduly absorb the electrons, it is necessary to have close contact between the electron emitter, the object and the film. With paper this is not difficult to achieve, but textile objects are often far from flat and considerably larger than stamps, bank notes, book pages or drawings. However, in an experimental

project such as that carried out on degraded silks by Brooks et al. (1996) where the fabric samples were both small and flat, the higher image contrast resulting from either or both of these techniques would have been advantageous. Xeroradiography Xeroradiography, originally developed for medical imaging, is a form of radiography which was quite popular for the study of cultural material for a few years. Instead of radiographic film, an electrostatically charged plate is used to capture the image. On exposure to X-rays the charge at any one point is dissipated in proportion to the amount of radiation striking the plate. The image is made visible by exposing the plate to a cloud of negatively charged powder which clings to the positively charged parts of the plate. At the boundary between higher and lower charged areas, the powder creeps towards the higher charge so producing a strong outline to features. Using heat and pressure, the image is transferred to a plastic-coated paper to form a permanent record (Rockley, 1964: 93–98). A blue powder is commonly used, producing blue and white images as seen in Figure 26.1 (see p. 309), 26.3 (see p. 312) and 27.5 (see p. 317). Xeroradiography has a wider exposure latitude than film, is less affected by scatter and the edge enhancement effect makes it very useful for detecting features in a matrix of very similar radio-opacity. However, the edge enhancement can be disadvantageous as the bright outline of one feature may hide the detail of another, particularly at boundaries where the change in X-ray flux is quite significant. Having fallen out of use in medical imaging, xeroradiography is really no longer available. O’Connor et al. (2002) showed that high definition radiography coupled with careful digital image processing can produce images which reproduce the virtues of xeroradiographs without the same loss of fine detail. Micro-focus radiography It is not possible to produce a radiographic image of a feature that is smaller than the effective focal spot of the X-ray tube, no matter how fine the grain of the film used. For instance, in images taken with a Faxitron model 43855 it is possible to distinguish individual wool fibres (Figure 9.1c, p. 127), but not those of silk. Micro-focus radiography employs electrostatic focusing of the electron beam and specially designed anodes to produce an X-ray source with an effective focal spot down to 10 μm or less in diameter (Halmshaw, 1995: 173). Much sharper higher

High definition X-radiography of textiles: methods and approaches 51

Real-time radioscopy

Figure 3.18 Detail of fine silk fabric taken using a low energy digital capture micro-focus radiographic unit at X-Tek Systems Ltd, Tring, Herts. (Sample No. 26; © Sonia O’Connor, University of Bradford.)

resolution images are produced with such equipment; see Figure 10.4, p. 154. These images can be viewed under far greater magnification than those produced with a conventional tube but the graininess of the film can become a limiting factor. However, this problem can be overcome if the image is magnified by projective techniques. By spacing the film, or other image receptor, at some distance from the object, the beam continues to diverge after passing through the object, forming a larger than life-sized image without the blurring which would be created by a larger effective focal spot. Figure 3.18 is a micro-focus image of a group of yarns from the silk featured in Figure 9.7 (see p. 131). In the contact radiograph of this sample taken using the Faxitron model 43855, individual warps and wefts can be distinguished under low magnification. With a micro-focus beam and projective magnification, it is possible to image individual silk fibres (Figure 3.18). Obviously there are practical limits to the magnification achievable because of the geometric unsharpness of the system and loss in beam intensity with increasing FFD. In the case of low energy imaging, filtration of the beam by the air column in its path further restricts the FFD, restricting the area of the object that can be covered by the beam in a single image.

Real-time systems are designed to capture X-ray images of moving subjects. As the term ‘real-time’ suggests, these systems are designed to show what is happening as it happens, such as a beating heart, in-line inspection of components moving through the X-ray beam on a conveyor belt or molten metal flowing into complex moulds. Whether the radiographic images are produced using fluorescent screens, image intensifiers or scintillators, and recorded in an analogue or digital form, real-time systems are unlikely to capture the subtlety of tone or fine detail required in textile radiography because image quality is compromised for speed of acquisition. However, rotating a complex object can help in the interpretation of the radiographic information by allowing the relationship of different features of the object to be explored as they are seen to move relative to one another. This technique has been used for the safety testing of soft toys. Metal detectors could prove that metal foreign bodies were present inside imported teddy bears but only through real-time micro-focus radiography was it possible to identify and locate the potentially dangerous metal inside each bear so it could be removed in a cost-effective procedure (Towell, 2002). Real-time images can also be useful in selecting the most informative views from which to take the higher quality static images. Stereoradiography Animals with stereoscopic vision, such as humans, judge distance by detecting differences in the relative position of features in the images received by each eye. Replicating this with radiographic images creates an element of depth, making it possible to judge the positions of features through the thickness of an object. This technique was used in one of the earliest radiographic investigations of an archaeological textile (see Chapter 1, p. 5). The stereoradiographs were taken in order to gain an understanding of the embroidered design of the metal thread work on the remains of a cushion found in Archbishop Walter de Gray’s tomb (Werner, 1971: 138–139). Available records suggest that the exposure was probably about 50 kV as the metal embroidery threads are highlighted but the organic textile components are not visible. In simple terms stereoradiography is achieved by taking pairs of radiographs of the same object from two viewpoints separated by the distance between the pupils of the eye, approximately 65 mm. These two images are then presented, one to each eye,

52

Textile X-radiography

and the brain combines the information to produce a perception of depth within the object. This is not truly a three-dimensional image and small movements of the head quickly demonstrate to the viewer the narrow limits over which the effect can be seen. Nevertheless, stereoradiography is often a very effective way of improving image interpretation as it facilitates the separation of image artefacts produced by a combination of overlapping features and so determining which features are towards the front or back of an object. The radiographic stereo pairs are created by moving either the object or the X-ray tube between exposures so that the beam centre will have moved by the requisite distance relative to the features of the object. For instance, the first image is taken with the beam centre slightly to one side of the centre of the object, or the feature of interest, and the second with it equidistant to the other side. Combining the images can be done in a number of ways using, for instance, separate light boxes to illuminate each image and a pair of mirrors between them angled to reflect the images to the eyes of the viewer. The images can also be viewed through a device called a binocular stereoscope, which brings the images together using lenses (Rockley, 1964: 169–171). Some people can bring the images together without the aid of the stereoscope by crossing their eyes but this takes some concentration and the effect may be fleeting. Alternatively the radiographs can be photographically reproduced and either projected towards the viewer using light polarised horizontally for one image and vertically for the other or printed as a single (anaglyph) image using a different colour for each image, such as red and cyan (Lang and Middleton, 2005: Plate 2.1). In both cases ‘3-D spectacles’ are worn to view the images. No lenses are involved so these viewing devices are very cheap and usually made from card or plastic frames with plastic sheet panes. Each pane is designed to let through the light from one image of the pair only, either by matching the plane of polarization or the colour. The image produced by the stereoscope can be emulated on computer from digital radiographic stereo pairs. Image software originally designed to produce partially rotatable ‘three-dimensional’ images by combining digital photographs of objects taken from different viewpoints can be used with radio-graphs. Digital radiographic images are also easily converted to different colour layers which can be combined to produce an anaglyph image and displayed on screen or printed for viewing with the appropriate 3-D spectacles. Rockley (1964) describes the practical aspects of taking stereoradiographs and Spicer (1985) explores

its application to the study of archaeological material as part of a discussion of stereoscopic imaging in general. Although stereoradiography has largely fallen out of use since the advent of computed tomography, which can produce real three-dimensional images, its main advantage remains that it can be performed with conventional radiographic equipment. Conventional tomography This X-ray imaging technique was commonly used in clinical radiography to produce an image of an area or plane within a patient without the superimposition of the shadows of other structures from above or below. It is achieved by coordinating the movement of both the X-ray tube and the film during the exposure about a pivot point. Everything away from the pivot point will be magnified, distorted and blurred because it will have moved relative to the source and film. Only the features in the plane of the pivot point will be sharp and easily distinguishable against the blurred structures. It has occasionally been used in the radiography of cultural material; for example, by Lindegarrd-Andersen et al. (1988) to explore a pattern-welded sword and in the investigation of the contents of the Anglian period Coppergate helmet (Tweddle, 1992: 898–900), but no published records of its use with textiles have so far been found. Computed tomography (CT) Conventional tomography and stereoradiography have largely been superseded in medical and industrial radiography by computed tomography (CT). Developed in the 1970s for medical imaging, CT combines elements of both these techniques with digital capture, producing an altogether more powerful and versatile imaging modality. CT scanners measure the differences in attenuation of an X-ray beam as it passes through an object in different directions and computes this information to construct a set of digital images of cross-sections, or slices, through its internal features. The information from a series of adjacent slices can be used to create sections through the object in other planes or to build up a three-dimensional image of the object that can be turned and viewed from almost any angle. Figure 3.19a is Martin Behaim’s fifteenth century globe from the collections of the Germanisches Nationalmuseum (WI2826). The CT images reveal the laminated structure of the two hollow hemispheres, which are joined at the globe’s equator using wooden pegs to secure them to a wooden hoop (Figures 3.19b and c). The weight of the globe has caused it to sag around the South Pole

(b)

(c)

(a)

(d)

Figure 3.19 Martin Behaim’s fifteenth century globe, (a) photograph; CT images originally published in Siemens Medical Radiography calendar 1993, (b) section through the diameter, (c) section through laminated structure of the skin, (d) peg securing internal wooden hoop at the equator. (WI2826 © Germanisches Nationalmuseum; reproduced with permission.)

54

Textile X-radiography

where it revolves on an iron axis. Beneath the vellum map the laminated shell structure includes many layers of canvas (Figure 3.19d). CT also allowed accurate measurements to be made of the internal features directly from the digital images (Mould, 1993: 102). The CT scanner used in clinical applications consists of a gantry that conceals the rotating X-ray source and detectors, with a hole in the middle through which a patient passes (Figure 3.20a). As the source and detectors move around, the patient is moved through the gantry to acquire information for successive slices. Despite issues of high cost and limited availability, CT has increasingly been used in the investigation of cultural material, including mummified human remains, stone, wood, ceramic and geological material (Lang and Middleton, 2005). In the case of Figure 3.20a the ‘patient’ is an early twentieth century suitcase that was scanned to explore the versatility of CT for the investigation of mixed-media textile objects. This figure illustrates that although quite long objects can be placed in the scanner, the width and depth are quite restricted. The maximum width for medical scanners is currently about 0.5 m. The corduroy-covered leather case is also lined with corduroy and fitted out with a range of toiletry equipment, cosmetic bottles and brushes. It housed a family collection of baby dresses (Figure 3.20b). Together these represent a very diverse range of materials, from textile, card and leather to wood, glass and various metals. Figure 3.20c is a CT section from about two thirds of the way along the case (approximately where the red light guides cross the case in Figure 3.20a) through one of the cosmetic bottles and the handle of the mirror in the lid. All the metal components, including the enamelled mirror handle, the bottle top, the wire of the folding coat hanger lying on the bottom of the suitcase and the nails attaching one end of the carrying handle, are radiating dark and light lines. Image artefacts like these are due to the very high density of the materials being scanned. These are far beyond the normal range for medical imaging but this problem can be overcome to a great extent by adjusting both the scanner’s calibration and the digital image. Despite this interference, the image reveals the construction of the suitcase and the layers of textile both outside and within the case. The study of the suitcase was carried out in 2002. The spatial resolution is quite coarse in these images as the thickness of the slices from which the sections were constructed is about 1 mm at best. Even so, Figure 3.21, a CT section of a rolled silk ribbon, demonstrated its potential to distinguish very thin

layers of low density material. Since then CT technology has continued to develop at a tremendous pace. Improvements in image quality, speed of acquisition, computing power, image manipulation software and a reduction in X-ray dose have made three-dimensional imaging of CT data a quicker and more commonplace matter. Blumer et al. (2005) have used high resolution industrial CT equipment to produce three-dimensional digital images of textiles preserved in metal corrosion products (pp. 49–50). This form of investigation allows viewers to ‘fly through’ objects. A virtual reality film of Nesperennub’s mummy using animated CT data featured in the British Museum’s exhibition Mummy: The Inside Story. This took the viewer on a three-dimensional journey through the embalmed body without having to unwrap the mummy. The smallest element of a CT image is a voxel – the three-dimensional equivalent of the pixel (see Chapter 4, p. 58). The value of each voxel relates to the beam attenuation at that point in the object and is called its CT number, which is measured in Hounsfield units (HU). Materials of different density attenuate the beam differently so they have different CT numbers. Water has a CT number of 0 HU, air is 1000 HU and bone is 1000 HU. This allows different materials to be distinguished from each other and tentatively identified. To aid image interpretation further, materials of different density can be displayed with different colours or selectively removed from the image to reveal underlying structures. Lang and Middleton illustrate these techniques with a three-dimensional CT image of the head of a Twenty First Dynasty mummy (2005: 136–137 and Plate 7.2). The skin and soft tissue of the head have been selectively removed from the image to show the bone and teeth which are coloured blue. The top of the cranium has also been removed to reveal the resin-soaked linen packing of the cranial cavity which has been rendered in purple. Techniques like these could revolutionise understanding of multi-layered structures, such as complex costume elements, whose conventional X-ray images can be so hard to interpret because so much information is superimposed. It should be possible to examine individually the components of a multi-layered structure assigning a different colour to each and peeling them away one by one – outer fabrics, linings, interlinings and facings. For small objects, up to a few centimetres in diameter, very much higher spatial resolutions are now obtainable. Cone-beam micro-computed tomographic scanners (micro-CT) combine the

High definition X-radiography of textiles: methods and approaches 55

(a)

(b)

(c)

Figure 3.20 Early twentieth century suitcase, (a) entering a medical CT scanner at Bradford Royal Infirmary, (b) contents, (c) CT section. (Private collection; © Sonia O’Connor, University of Bradford.)

three-dimensional capabilities of a CT scanner with the resolution of a microscope. The object is rotated in the path of a micro-focus X-ray beam and spatial resolutions in the order of 5 μm and features as small as 1 μm can be visualised. These machines are used for examining small animal specimens (Ford et al., 2003) or electronic components but can also be used to produce detailed structural images of a wide variety of objects as diverse as bone, seeds, diamonds, wood and fibre reinforced composites. Micro-CT is used in textile technology studies in many ways, such as in the evaluation of textile

scaffolds for tissue engineering, the definition of three-dimensional non-woven structures and to map wear. Weder et al. (2006) used micro-CT to investigate moisture distribution in multi-layered textiles. For samples of a few millimetres diameter, nano-CT can achieve resolutions below 0.5 μm. It is now even possible to ‘fly through’ textile samples, passing between the very yarns of the weave by producing animations of the three-dimensional data. Blumer et al. (2005) have also used CT data to generate threedimensional resin copies of the structure of archaeological objects using computer-aided design (CAD)

56

Textile X-radiography

Figure 3.21 CT section through two rolled silk ribbons taken at Bradford Royal Infirmary. (Private collection; © Sonia O’Connor, University of Bradford.)

software and stereolithography, also known as 3-D printing or 3-D layering. Stereolithography uses a liquid photopolymer that hardens when exposed to an ultraviolet laser light, to build an object layer by layer (typically five to ten layers per mm) from digital data. With this technique, it would be possible to replicate the detail of textile objects, features or surfaces revealed by CT that could not themselves be exposed or which were too fragile to handle.

Notes 1. 2.

Personal communication, David Crombie, Senior Paintings Conservator, National Museums, Liverpool. The upper part of the torso in Figure 3.12a is filled with lavender flowers, obscuring the elastic band securing the arms. This filling has settled lower down in Figures 3.12b and c.

References Aldrovandi, A. and Ciappi, O. (1995). La radiografia di grande formato: problemi e soluzioni tecniche. In OPD Restauro. Rivista dell’Opificio delle Pietre Dure e Laboratori di Restauro di Firenze, 7, 163–168, Edizione Centro Di. Blumer, R-D., Butenuth, J., Nowak-Böck, B. and Peek, C. (2005). Inventarisation und dokumentation.

Neue einsatzmöglichkeiten digitaler medien in der archäologischen Denkmalpflege. Denkmalpflege in Baden-Württemberg, 1, 29–36. Bridgman, C. F. (1973). The radiography of museum objects. Expedition, 15(3), 2–14. Brooks, M. M. and O’Connor, S. A. (2005). New insights into textiles. The potential of X-radiography as an investigative technique. In Scientific Analysis of Ancient & Historic Textiles, Informing Preservation, Display and Interpretation. Post-prints of the AHRB Research Centre for Textile Conservation & Textile Studies, 13–15 July 2004 (R. Janaway and P.Wyeth, eds), pp. 168–76, Archetype Press. Brooks, M. M., O’Connor, S. and McDonnell, J. G. (1996). The application of low energy X-radiography in the investigation of degraded historic silk textiles. In Preprints of the 11th Triennial ICOM-CC Triennial Meeting Edinburgh, Scotland, 1–6 September 1996 (J. Bridgland, ed.), pp. 670–679, James & James. Carr, C. and Riddick, E. B. (1990). Advances in ceramic radiography and analysis: laboratory methods. Journal of Archaeological Science, 17, 35–66. Daniels, V. and Lang, J. (2005). X-rays and paper. In Radiography of Cultural Material (2nd ed.) (J. Lang and A. Middleton, eds), pp. 96–111, Elsevier. Eastop, D. and Brooks, M. M. (1996). To clean or not to clean: the value of soils and creases. In Preprints. ICOM Conservation Committee. 11th Triennial Meeting, Edinburgh, pp. 687–691, James & James. Ford, N. L., Thorton, M. M. and Holdsworth, D. W. (2003). Fundamental image quality limits for microcomputed tomography in small animals. Medical Physics, 30, 2869–2877. Graham, D. and Thomson, J. (1980). Grenz Rays. Pergamon Press. Halmshaw, R. (1995). Industrial Radiology. Theory and practice (2nd ed.), Chapman and Hall. Janaway, R.C. (1987). The preservation of organic materials in association with metal artefacts deposited in inhumation graves. In Death, Decay and Reconstruction. Approaches to Archaeology and Forensic Science (A. Boddington, A. N. Garland and R. C. Janaway, eds), pp. 127–148, Manchester University Press. Janaway, R. C. (1989). Corrosion preserved textile evidence: mechanism, bias and interpretation. In Evidence Preserved in Corrosion Products: New Fields in Artifact Studies (R. Janaway and B. Scott, eds), pp. 21–29, United Kingdom Institute for Conservation of Historic and Artistic Works. Javér, A. (1997). Determining a Conservation Strategy for a Sprang Cap Preserved on a Naturally Dried Head. TCC 2271. (Unpublished report.) Textile Conservation Centre. King, D. (1971). The textiles. In The Tombs of Archbishop Walter de Gray (1216–55) and Godfrey de Ludham (1258–65) in York Minster and their Contents (H. G. Ramm, ed.), p. 136, Society of Antiquaries of London.

High definition X-radiography of textiles: methods and approaches 57 Lang, J. and Middleton, A., eds (2005). Radiography of Cultural Material (2nd ed.). Elsevier. Lindegarrd-Andersen, A., Vedel, T., Jeppesen, L. and Gottlieb, B. (1988). Film-based X-ray tomography combined with digital image processing: investigation of an ancient pattern-welded sword. NDT International, 21(6), 407–409. Mould, R. F. (1993). A Century of X-rays and Radioactivity in Medicine. Institute of Physics Publishing. McClung, R. W. (1964). Studies in contact microradiography. Materials Research and Standards, 4(2), 66–69. Nowak-Böck, B., Peek, C. and Pfeifer-Schäller, I. (2005). Ertse ergebnisse zum einsatz der 3D-computertomographie bei der untersuchung archäologischer textilien. Die 3D-röntgen-computertomographie, Archaeological Textiles Newsletter, 40 (Spring), 11–18. O’Connor, T. and O’Connor, S. (2005). Digitising and image-processing radiographs to enhance interpretation in avian palaeopathology. In Documenta Archaeobiologae 3 (G. Grupe and J. Peters, eds), pp. 69–82, Verlag Marie Leidorf. O’Connor, S., Maher, J. and Janaway, R. (2002). Towards a replacement for xeroradiography. The Conservator, 26, 100–114. Padfield, J., Saunders, D., Cupitt, J. and Atkinson, R. (2002). Improvements in the acquisition and processing of X-ray images of paintings. National Gallery Technical Bulletin, 23, 62–75. Ramm, H. G. (1971). The tomb of Archbishop Godfrey de Ludham. The opening of the tomb and its identification. In The Tombs of Archbishop Walter de Gray (1216–55) and Godfrey de Ludham (1258–65) in York

Minster and their Contents (H. G. Ramm, ed.), pp. 131–134, Society of Antiquaries of London. Rendle, D. F. (1993). The use of soft X-rays in forensic science. British Journal of Non-Destructive Testing, 35(7), 381–383. Rockley, J. C. (1964). An Introduction to Industrial Radiology. Butterworth. Spicer, D. (1985). Stereoscopic representation of archaeological data – a case for drawing conclusions in depth. Science and Archaeology, 27, 13–24. The Stationery Office (1999). Ionising Radiations Regulations (Statutory Instrument No. 3232). Her Majesty’s Stationery Office. Towell, T. (2002). Non-destructive testing – micro-focus X-radiography. Materials World (Special Testing Supplement), April, 12–14. Tweddle, D. (1992). The Anglian Helmet from Coppergate. Council for British Archaeology, Archaeology of York 17/8. van Aken, J. (2002). An improvement in Grenz radiography of paper to record watermarks, chain and laid lines. Studies in Conservation, 48(2), 103–110. Weder, M., Brühwiler, P. A. and Laib, A. (2006). X-ray tomography measurements of the moisture distribution in multilayered clothing systems. Textile Research Journal, 76(1), 18–26. Werner, A. E. A. (1971). Scientific reports. The scientific examination and treatment of objects from the tombs. Archbishop de Gray. In The Tombs of Archbishop Walter de Gray (1216–55) and Godfrey de Ludham (1258–65) in York Minster and their Contents (H. G. Ramm, ed.), pp. 136–140, Society of Antiquaries of London.

4 Textile X-radiography and digital imaging Sonia O’Connor and Jason Maher

Introduction Combining radiography with digital imaging greatly increases the value of radiography as an investigative approach. This topic is covered in relationship to the radiography of cultural material by O’Connor and Maher (2001) and Lang and Middleton (2005). Digital radiographic images can be acquired in two ways, either through filmless digital radiographic techniques or by digitising a film radiograph. Once a digital radiograph has been obtained the flexibility of this medium means that images can be altered with ease thereby making digital radiographs more useful for the investigation of textiles than conventional film. When assessing the suitability of a digital acquisition system for the radiography of textiles, it is necessary to understand the basics of digital imaging. Digital imaging consists of four stages, capture, storage, manipulation and output. Rather than reference specific technologies or computer programs which will quickly become out of date, this chapter details the considerations that should be foremost when evaluating digital images and the equipment and programs used to capture, manipulate and store them.

Digital versus analogue Radiographic film images are analogue images. As in paintings and photographic film images, the tones of a radiograph can take a continuous range of values along any line across the image. Digital images, such as those produced by digital cameras or image scanners, may look like analogue images but they are composed of discreet units, called pixels or pels (both terms are contractions of the phrase ‘picture element’), each having a single colour value. By converting from film radiographs to digital radiographs, it becomes possible 58

to produce endless numbers of copies, each identical in every detail. They can be viewed on a computer, printed out on transparent film or paper, sent electronically to researchers elsewhere, included in conservation reports, publications and presentations, linked to searchable databases or displayed on a website. Unlike vulnerable film images, viewing does not degrade the digital image. In an archive, the digital image takes up far less space than film and copies can be stored in different locations for added security. In addition, digital radiographs can be explored in ways which are not possible with an analogue image thus more fully revealing the potential of radiography as an investigative tool for textiles. A good quality digital copy of a film radiograph will appear faithful to the film original in its tonal quality. Hundreds of shades of grey can be captured in the digitised image and each one displayed on a computer monitor. The eye can distinguish less than a hundred shades of grey between black and white and it is not equally sensitive in all regions (Okkalides, 1996). The darker tones are more difficult to separate by eye, rendering it impossible to see some details of a radiograph. However, whereas film might only be viewed with a dimmer or brighter light, the grey tones of the digital image can be altered more subtly. At first, it might not be possible to see any more detail in the ‘as captured’, or raw, image than was visible on the film but changing the colour values of the pixels through digital image processing (DIP) enables previously invisible features to stand out boldly from the background (Figure 4.1). The image can be manipulated in other ways to improve the visibility of the captured information but it is important to remember that this manipulation does not create information – it merely makes existing features more discernible. Digitising film radiographs allows the original film to be protected in an archival state so it can

Textile X-radiography and digital imaging 59

(a)

(b)

(c)

Figure 4.1 Turkish waist cloth detail, (a) photograph, (b) radiograph raw ‘as captured’ image, (c) radiograph after image processing. (Private collection; © Sonia O’Connor, University of Bradford; reproduced by permission of Rosemary Payne.)

always be redigitised in the future if the need arises, for instance when better quality digitisers become available. Filmless digital radiography cannot offer this but does away with wet processing and issues related to film storage. Some filmless systems offer an even greater dynamic range than industrial radiographic film, improving information capture and image contrast while maintaining a wide exposure latitude and reducing the need for repeat exposures. The digitalisation of radiography has also revolutionised the process of making mosaics from multiple images of single objects. Using a computer, it is possible to crop, rotate and resize images and assemble a mosaic from them to produce a single digital radiographic image in a fraction of the time required to do this photographically.

Components of a digital image When evaluating digital images and the equipment used to generate them there are three important characteristics used in determining quality: resolution, bit depth and dynamic range. The same characteristics are used to describe the equipment used to capture the digital image (scanner, camera or imaging plate, etc.), the digital radiographs themselves and the hardware used to output the images (printer, screen, etc.). Unlike an analogue image where the physical properties are fixed (width, length, tonal range, etc.) a digital image can be manipulated and its resolution, bit depth and contrast altered. Simply put, the greater the values associated with each of these characteristics,

the greater the ability of the image to resolve fine detail and subtle changes in colour, contrast or optical density. This is true for all digital images whether they consist of multiple colours or, as in the case of digital radiographs, shades of grey from black through to white. Resolution The term ‘resolution’ is employed in slightly different ways depending on whether it is being used to describe the hardware used to capture the images, the properties of the digital image itself or the equipment used to display or print the image. When used in relation to a digital image, resolution is an indication of the viewable size of the image. Unlike an analogue image a digital image has no physical width or height. Its dimensions are expressed in horizontal and vertical pixels; for example, an image described as having a resolution of 1024 768 consists of 1024 pixels horizontally and 768 pixels vertically, giving a total of 786 432 pixels in the whole image. The input ‘resolution’ of the digitisation hardware used to sample or capture an analogue image indicates how many times per inch a reading is taken of the colour value in the analogue image and therefore is measured in points per inch (p.p.i.). Each sampling point equates to a single pixel in the digital image so increasing the p.p.i. increases the number of pixels created per inch of the analogue image. This effectively reduces the relative size of each individual pixel and increases the ability of the digital image to resolve fine detail (Figure 4.2).

60

Textile X-radiography

(a)

(b)

Figure 4.2 Matchbox doll, (a) photograph of head, (b) matrix of radiographic images of head showing effect of increasing the number of pixels and bit depth on image quality. (Private collection; © Sonia O’Connor, University of Bradford.)

Textile X-radiography and digital imaging 61

In theory, the size of the grains in the emulsion of a radiographic film will limit the size of the smallest details that can be resolved in the image and, ideally, the sampling resolution should reflect this or detail will be lost. For fine-grained film, with a grain size of approximately 50 μm, this equates to a sampling resolution of approximately 512 p.p.i. For ultra finegrained film with a grain size of 10 μm, this would equate to approximately 2540 p.p.i. However, the actual resolution of the film image is never this fine due to unsharpness caused by the geometry of the X-ray equipment (see Chapter 2, pp. 17–8) and factors inherent to the film itself. A sampling resolution of 512 p.p.i. was found to be perfectly adequate to capture the detail visible in even the ultra finegrained film images digitised for this book in all but a few cases. The resolutions associated with various forms of output are differentiated by reference to the output resolution (general), print resolution (printer) or display resolution (screen). Output resolution is described in a similar manner to input resolution. Each pixel of the image is represented by a single dot in the output and the resolution of this is expressed in the number of dots per inch (d.p.i.). As d.p.i. is increased, a digital image with a fixed number of pixels will be printed or displayed in a smaller area. Unless the sampling resolution and output resolution are equal, the resultant image will not be the same size as the original analogue image. This is most evident when viewing digital images on a computer screen. If a monitor has a display width of 15 inches (380 mm), a height of 12 inches (305 mm) and a resolution of 1280 1024 pixels, this will be equivalent to 85 d.p.i. An image sampled at 512 p.p.i. will look six times larger than the original when displayed on the screen. To see the digitised object at roughly the same size as in life, it is necessary to view the digital image at 17% of its current size. Bit depth The term ‘bit depth’ or bits per pixel (b.p.p.) refers to the way in which colour information is stored in the digital image. The greater the bit depth, the wider the range of colours that can be represented. Colour images are comprised of three channels (shades of the same colour); red, green and blue. Radiographic images should not be captured as colour images but as greyscale images. These utilise a single channel to record the shades of grey between black and white. Within a greyscale image, each pixel has a specific colour value, from black with a value of zero through

to white, the value of which is determined by the bit depth of the image. The relationship between bit depth and the number of colours that can be stored is exponential (a bit depth of 2 can record 4 different shades, a bit depth of 3, 8 shades, etc.), so a small increase in bit depth can mean a big difference in the ability of the image to capture subtle changes in colour (Figure 4.2). If only a few colours can be recorded, similar colours in the analogue image have to be grouped together when it is digitised and represented by a single shade so detail is lost. In practice, radiographic images need to be captured at a minimum of 8 bits, although images with a wide dynamic range such as those produced with the very high contrast industrial films will need a greater bit depth. For a computer to display the full bit depth of a radiographic image, the graphics card must be set to a bit depth three times that of the greyscale image. If the computer’s bit depth is too low, very similar shades of grey will be displayed as single values with the result that not all of the detail that was captured will be visible. Dynamic range As with film, the dynamic range of a digital image is a measure of the range of densities in the image (see Chapter 2, p. 21). The term is also used to describe the sensitivity of the digital capture hardware. The maximum density a piece of hardware can capture or display is referred to as Dmax and is based on a logarithmic scale. It is important to match the Dmax of the digitising equipment with that of the film. Digitising equipment with a lower Dmax than the radiograph will have a narrower dynamic range than the film and this will lead to the digitised image lacking information in the lightest and darkest regions. Most document scanners and digital cameras do not have a high enough Dmax to capture all of the information in a radiograph. Conversely, some filmless digital capture receptors have a wider dynamic range than radiographic film and can record information in regions that would have been either underor overexposed in a film image taken at the same beam kV. The wider the dynamic range of a given image, the greater the bit depth needed to express all of the density information. In the medium to long term, the biggest change to digital images will occur as a result of the move to High Dynamic Range (HDR) formats. These will encode 32 (and greater) bits per pixel and thus bring with them a significant increase in dynamic range. This reference to HDR formats is

62

Textile X-radiography

distinct from the creation of HDR images themselves, which are usually produced from the merging of multiple exposures of the same object (Reinhard et al., 2005). Currently, even the highest quality computer monitors do not have the capability to display HDR images to their full potential but HDR displays are being developed.

Storage of digital images Digital images are stored as electronic files on physical media, either magnetic or optical. No matter at what resolution or bit depth a radiographic image is digitised, it is the file format in which the digital image is saved, and the medium on which it is stored, that governs the long-term usefulness of the image. For this reason it is important to pay particular attention to these two areas when considering what method of storage to use for archiving digital radiographs. File formats Almost all radiographic digitisation hardware will initially store digital images in a ‘proprietary’ format, one owned and used only by the manufacturer of the digitisation equipment. The sharing of these images will be limited to other users who have the same hardware and software so, in most cases, digital images will also need to be saved in a file format that is more widely accessible. However, a copy in the initial digital format, even if proprietary, should always be retained. Of the many file formats available, only a few are suitable for the storage of digital images. For primary storage, a format must be chosen which records the maximum amount of detail while remaining relatively compact and is sufficiently widely supported as to be accessible to as many different programs as possible, now and in the future. This does not preclude the same digital image being saved in a variety of secondary formats, each one being chosen for a specific purpose. For instance, image formats for the web tend to be optimised for size ( JPEG, GIF, PNG) while formats for digital image processing are optimised for content (TIFF, DICOM, PNG), and formats for print focus on accuracy of reproduction for output to a printer or display on a screen (PDF, EPS). Table 4.1 compares the uses and features of some of the most popular image file formats. The selection of a primary format should also take into consideration the type of compression used, the addition of image information in the form of metadata and the permanency of the file format.

Compression Different file formats incorporate different methods of compressing the image in order to make storage more efficient. Compression can be either lossy or lossless and influences the choice of file format for particular applications (see Table 4.1). Lossy compression reduces the size of the file by reducing the complexity of the visual information in the image and thereby the amount of space needed to store the file (Gonzalez and Woods, 2001). This reduction of complexity can often introduce errors or artefacts into the image. If possible, file formats that utilise lossy compression should be avoided for primary storage formats. Repeatedly saving a file in a lossy format will result in further degradation of the image. Lossless compression compacts a file by finding more efficient ways of encoding the same level of detail and is artefact free. Even then not all lossless formats will efficiently compress the continuous tonal range of a radiographic image; the LZW compression used in GIF images, in particular, is unsuitable. It is possible to forgo compression all together although the resulting files will be larger. Metadata Metadata stores information that describes the contents of an image. Rather than having to rely on an association between an image reference number and information in a database, metadata encodes details (such as the date taken, keywords for cataloguing the contents of the image, the capture hardware used, etc.) within the image file itself. Two elements of metadata are of specific importance; the way the metadata is encoded into the image and the vocabulary used to describe the various elements of the object in the image. Digital images captured by modern digital cameras will automatically include metadata encoded in the Exchangeable Image File (EXIF) format. EXIF records the settings (the date, time, aperture, ISO, focal length, etc.) of the hardware used to capture the image. Additional metadata can also be added by the user to a pre-existing image but this standard is not supported by radiographic capture equipment and so is of little use. The most widely supported metadata standard for digital images at the moment is named after the International Press Telecommunications Council (IPTC). While there are other metadata initiatives, such as DIG35 and MPEG-7, it is XMP (eXtensible Metadata Platform) that is the most promising.

Textile X-radiography and digital imaging 63 Table 4.1 Uses and features of some of the most popular image file formats Metadata4

Use5

Notes6

XMP None

Web Archive

16

EXIF/IPTC/XMP

Web/Archive

High/Low

16

EXIF/IPTC/XMP

Archive

Lossy7

High

16

EXIF/IPTC/XMP

Web/Archive

Lossless JPEG JPEG2000

Lossless Both

Low High/Medium

16 16

EXIF/IPTC/XMP EXIF/IPTC/XMP

Archive Web/Archive

JPEG-LS

Lossless

Medium

16

EXIF/IPTC/XMP

Archive

DICOM8

Both

High/Low

16

DICOM

Telemedicine9

PDF

Both

High/Low

8

XMP

Print

EPS

Both

High/Low

8

XMP

Print

Not suitable Low bit depth, little compression Best compressed format Use version 6, proprietary Versatile, but lossy. Poor support Poor support for lossless option Poor support Need specialist programs Print only, use LL compression Print only, use LL compression

Format

Compression type1

Compression ratio (Ly/LL)2

GIF BMP

Lossless Lossless

Low Low

PNG

Lossless

Medium

TIFF

Both

JPEG

Greyscale bit depth3 8 8

1

The type of compression available in the file format; some file formats offer a choice of lossy and lossless compression. A relative comparison of the compression offered by the 11 formats considered here. For formats with different algorithms the most compacting was used, ratings are given separately for lossy (Ly) and Lossless (LL). Test image was a radiographic film digitised at 8 b.p.p. with dimensions of 7080 8670 pixels. 3 This is the maximum bit depth supported for greyscale images, which is accessible by most imaging packages, e.g. theoretically, BMP can encode 16 bits per pixel but few programs can read these files, therefore the maximum, readable BMP image records 8 bits per pixel. 4 The most common metadata standards supported by the file format. 5 Most common use for each file format. 6 Recommendations for each file format. 7 Even when compression is set to zero or ‘Highest Quality’ a JPEG file is still lossy compressed. Using no compression in a JPEG file is not the same as lossless JPEG. 8 Digital Imaging and Communications in Medicine. DICONDE (Digital Imaging and Communication in Nondestructive Evaluation) is a standard based on DICOM used in industrial radiography. The metadata categories in DICONDE are more suitable for conservation work than those used in DICOM, but standards compliance is still a complicating factor. 9 Telemedicine refers to the delivery of medicine at a distance using communication and information technologies. 2

Established by Adobe and based on the Resource Description Framework (RDF) it uses the Dublin Core Metadata Initiative (DCMI) description building blocks (International Organisation for Standardisation, 2003). Currently IPTC is more widely used for digital images than XMP. However, XMP has been adopted as a standard by industry. This support, together with the creation of mapping that allows IPTC data to be embedded using XMP, means that XMP is gradually becoming the metadata standard of choice for long-term storage and compatibility.

Stability of file formats Finally, the longevity of a file format is of concern when considering long-term storage. The file format governs which programs can be used to read and manipulate a file. Unpopular file formats quickly die out. When a file format is no longer supported by programs it becomes unreadable, resulting in all the information being lost. Currently the file format most frequently recommended for archival purposes is TIFF (Technical Advisory Service for Images, 2002a; Technical Advisory Service for Images, 2004)

64

Textile X-radiography

although PNG has garnered support as it is an open standard (the rights to it are not owned by a company but are made available free to everyone), supports metadata and does not use proprietary compression algorithms. Lossless JPEG 2000 should also be considered if its popularity increases (Technical Advisory Service for Images, 2002a; Technical Advisory Service for Images, 2003). Storage media The choice of storage medium used for saving files governs the physical durability of the data. While it can be difficult to obtain accurate information on the stability of different media for use in an objective comparison, this is, nevertheless, a factor that should impact on the selection of a suitable storage medium (O’Connor and Maher, 2001; Technical Advisory Service for Images, 2002b; Byers, 2003). Of equal importance is the availability of equipment to read the chosen storage media. The drives to read the 51⁄4 and 3 disks used in the 1970s and 1980s have, in less than thirty years, become obsolete and hard to find. Images contained on these media would be difficult to recover even if the file formats and physical medium have persisted. With increasing commercial competition in the removable storage market, the uncertainty over which media standards will be preserved is greater than ever. Therefore, no matter what the merits of an individual storage medium may be, the most sensible archival strategy is that of data migration. Archives should use the most suitable medium for storage at the time but this should be checked for consistency once a year and migration to new storage media undertaken as these become available. As with all digital data storage, digital images should be seen as part of a dynamic archive that must be monitored and maintained.

Digital image capture When considering the methods available for digital image capture, it is worth noting that, no matter how cutting edge the technique or equipment is now, it will be superseded by something even better in the coming years. There is no substitute for having the original object available for re-imaging in the future but few people are in that position. It is important, therefore, that the best technology available at the time is employed to acquire the image required. For infrequent users, the best system may currently seem to be film radiography coupled with

digitisation because the conversion to filmless capture involves a much larger capital outlay. Film digitisation equipment also has the advantage that it can be used with archived film holdings. However, as with photography, the radiographic imaging world is changing quickly. Once frequent users have converted to digital capture, film and equipment related to its processing may become increasingly difficult to source. Many museums in North America are already having to meet the challenges of converting to filmless radiography because of the environmental issues surrounding the disposal of processing chemicals and waste water from the film washing. Ultimately, the only way to judge these systems is to use them. When evaluating a system, use test objects which reflect the range of materials and complexities of the textiles expected to be radiographed most frequently. A mid-twentieth century plastic doll was one of the objects originally selected as a test subject to refine radiographic techniques for textiles (see Chapter 3, pp. 28–30 and Figure 3.3 p. 29). It has proved equally useful for gauging the quality of various digital capture systems. Whatever test objects are selected, it will be necessary to have high quality film radiographs of them to act as a benchmark for comparison with the digital images. The comparisons will only be valid if the film and digital images are acquired using the same X-ray unit. If the digital capture system is to be added to an existing X-ray facility, this is where the tests need to be conducted. Digitising film radiographs The quality of the original analogue image and the selection of the equipment used to acquire the digital image are both important in determining the quality and usefulness of the digital copy. The condition of the radiograph is also important. The less the film has been handled, the fewer finger marks and scratches will be on the surface to interfere with the transmission of light through the image during capture. Dust will also be recorded as dark spots on the digitised image so ideally film radiographs should be scanned immediately after processing and drying – even before they are studied on a light box or labelled. The large size, high resolution and density range of radiographic films can all present problems for digitisation. Digital scanners, video and stills cameras have all been used to digitise radiographic films with varying degrees of success. Comparison of the results produced by a range of equipment with the same film image show that the best option for digitising whole radiographs is a dedicated industrial

Textile X-radiography and digital imaging 65

radiographic scanner (O’Connor and Maher, 2001). Such facilities are available in the heritage sector but usually it is necessary for the work to be outsourced. Digitising a radiograph in-house with a camera or photographic transparency scanner usually means that image quality will be compromised. The importance of the consequences of this will depend on the intended use of the image. Digital photography, for instance, may be of no use for copying a complete film for archive purposes or for exploration through DIP but perfectly adequate for capturing a detail of an image for a report. The capture software and file format used with many of these devices may automatically apply pre-processing to the digital image, such as sharpening algorithms, before it is saved. It may not always be possible to disable these features or reverse their effects. The radiographic images in Figure 4.3 are details from a film radiograph digitised using a range of

equipment.1 The detail is from a radiograph of the test doll (Figure 3.3, p. 29) and shows the top of the detached left leg with the remains of the elastic band, beads, sequins and various layers of textile. With the exception of Figure 4.3b, the whole radiograph was captured as a single image. By reproducing the same detail from each, at the same scale, it is possible to evaluate the quality of the resulting digital image. Figures 4.3a and b were taken using a relatively cheap, 4 megapixel camera, a Nikon Coolpix 4300, on a copy stand with the radiograph illuminated by a light box. Figure 4.3a was cropped from the complete image but Figure 4.3b was taken in close-up for maximum resolution. When using a camera and light box, it is important to surround the film with a nonreflective matt black surface so that only light coming through the film reaches the camera and to photograph it in low ambient light. This avoids interference from reflections on the surface of the film. Figure 4.3c

(a)

(b)

(c)

(d)

(e)

(f)

Figure 4.3 Film radiograph of the plastic and nylon net test doll. Detail digitised using a range of different capture systems, (a) and (b) relatively cheap camera, (c) high-end studio camera (photographed by Andy Chopping, Museum of London Archaeology Service and reproduced by permission), (d) high-end document scanner (scanned by Andy Chopping, Museum of London Archaeology Service and reproduced by permission), (e) medical X-radiographic scanner, (f ) industrial X-radiographic scanner. (Private collection; © Sonia O’Connor, University of Bradford.)

66

Textile X-radiography

was taken using a high-end studio camera; a Sinar P2 monorail camera with a digital Sinaron 105 mm lens and a Leaf Volare digital camera back. The Leaf Volare is a 6 megapixel back designed to produce images with a high degree of colour fidelity. Figures 4.3d, e and f were captured using three different types of digital scanner. Figure 4.3d was captured on a high-end document scanner, a UMAX Powerlook 2100 XL Pro with a Dmax of 3.4. This has a built-in light source for scanning photo transparencies. Figure 4.3e was produced using a Howtek medical radiographic film scanner and Figure 4.3f with an industrial radiographic scanner, an Agfa FS50B. Unsurprisingly, the relatively cheap digital camera has produced a low resolution image and its narrow dynamic range has meant that no information has been captured in the higher and lower density areas of the image (Figure 4.3a). The close-up taken with the same camera has a greatly improved resolution but the dynamic range is just as narrow (Figure 4.3b). The exposure is slightly different, giving a little more detail in the darker areas but with a consequent loss of detail in the brighter areas. Neither of these images can be usefully explored further through DIP. The high-end camera gives greater dynamic range and a higher resolution while the camera’s preprocessing (the ‘unsharp’ mask) has produced a fine edge to the detail making the image appear sharper (Figure 4.3c). To reduce exposure inconsistency for this image, the radiographic film was supported on a glass plate suspended about half a metre above the light box. This provided a more accurate exposure range and retained a greater degree of fine detail, particularly in the lighter areas of the film but information in the darker areas was lost. Exploring the dark areas using DIP reveals some indication of the fine net mesh but the detail is poor and overlaid with electronic ‘noise’ in the form of bright speckling. The document scanner produced an image that is biased towards the very lowest density areas (Figure 4.3d). The exposure was adjusted as far as possible to cope with the contrast range of the original film but this has compressed much of the mid-range detail and has coped least well with the very dark regions. Although the resolution is good, the results do not seem focused when compared with Figure 4.3c because the image has not been pre-processed to improve the sharpness of features. Again, when the darker regions are explored with DIP something of the fine net becomes visible but so does electronic ‘noise’. The resolution of the images produced by the medical and the industrial radiographic scanners is the same but the quality of the images is very different.

Both scanners have a much wider dynamic range than the document scanner or the cameras, capturing information from the lightest to the darkest areas of the object. The medical scanner, however, suffers greatly from electronic ‘noise’ where the light is not bright enough to penetrate the higher optical densities developed in the industrial film image. The presence of this ‘noise’ means the black background of the film appears as dark grey in the digital image and the detail of the object is degraded (Figure 4.3e). The Dmax and dynamic range of the industrial radiography scanner is exactly matched to that of the film. This has produced an image whose tonal range matches the original film and which has captured information from even the lightest and darkest regions (Figure 4.3f ). The resolution and bit depth are both relatively high – sufficient to enable enhancement of detail through DIP. The majority of the film images in this book were digitised in preparation for publication with the industrial scanner used in this test. Filmless radiography As noted, medical and industrial radiography increasingly uses filmless radiography techniques which capture the image in a digital form. Digital radiography has many advantages over film radiography. There are no film costs, no film stock to go out of date and no film storage problems. The health and safety procedures relating to wet processing of films and environmental issues related to disposal of waste films and processing chemicals and the quantity of water used for film washing are no longer relevant. Filmless radiography can also lead to a considerable saving in time. Manual wet processing to produce archival standard results can take the best part of an hour before the film is dry and fully viewable. Automatic processors are faster although the resulting radiograph may have a shorter life expectancy. However, the delay between taking and viewing the digital image can be a matter of seconds rather than minutes. Digital systems also tend to have a wider exposure latitude than film, reducing the need for retakes and making it is easier to get a reasonably good exposure of features with quite diverse radio-opacity on a single image. If the exposure is poor or the view not useful, the digital image is simply erased and the exposure repeated with no waste and little time delay. The main disadvantage of digital radiography is that the capital outlay is very high compared with film and film processing costs. Unless facilities can be shared, this makes the real cost, per digital image, much higher than that of film for the low volume user.

Textile X-radiography and digital imaging 67

Filmless digital capture can be direct or indirect. Direct digital capture is where the variations in X-ray flux are detected (e.g. by an amorphous silicon detector) and converted into a digital signal. The detector sits in place of the film and is connected to a computer where the data is processed to form the digital image. Direct capture can produce high quality images but it is very expensive and the panels used for image capture are inflexible and easily damaged. Indirect capture involves an intermediate stage where the X-ray flux forms a light image, using a material which scintillates or fluoresces. It is this that is detected and used to form the digital image. Not all digital capture techniques produce images of a quality suitable for the radiography of textiles. For instance, any technique designed for virtually instantaneous acquisition of images, such as is required for real-time radiography, may not produce images with sufficient image contrast or resolution of detail (see Chapter 3, p. 51). Digital systems used in medical imaging, while generally an improvement on medical radiographic film capture, will still be calibrated for low dose and short exposure times. Resolution may be sacrificed depending on the imaging application. Cost is also a guide. Small cheap digital scanners are available which are often recommended for installation into cabinet X-ray units. These convert the X-ray flux to a light image and then digitise this in a similar way to a flatbed document scanner, producing images with a rather low resolution for many applications. Figure 4.4 is a detail of the same doll as in Figure 4.3, showing all the materials from a metal pin through to lavender seed heads (see Chapter 3, pp. 28–30). It provides a comparison between digital images captured using a digital X-ray scanner of this type, an NTB EZ 240, and by digitising a high quality film image with the Agfa FS 50B industrial radiographic film scanner. Both images were made using the same model of cabinet X-ray unit; a Hewlett Packard Faxitron 43855 A. The quality of the digital radiograph (Figure 4.4a) is very poor compared with the digitised film image (Figure 4.4b). Unfortunately, at the very low kVs utilised in textile radiography this has produced an image with low contrast and dynamic range. Combined with the lower resolution, striations and other electronic ‘noise’, together with the superimposition of the image of the surface of the scanner, this has produced disappointing results of limited information value. Some digital systems have software that automatically pre-processes the captured image in a particular way before it is viewed. This is done to provide optimal images of a range of specific and predictable

features very quickly. For instance, the image in medical systems may be treated in different ways depending on whether bone or soft tissue is being radiographed. In an industrial setting, the operator may have to select different presets for welds or particular casting faults. These settings are very subject specific and none may give good results with images of textiles. As noted, systems which automatically archive the raw image are a better option as they allow full control of the digital image processing. This unprocessed image is usually saved in a proprietary format and cannot be altered but it can be viewed and manipulated with the image capture software’s digital imaging tools. Copies, manipulated in various ways to improve the visibility of the captured information, can then be saved or exported from the system. This ensures that an unprocessed image is always available for reference and can be explored with different image software at a future date.

Direct and computed radiography Direct radiography (DR) and computed radiography (CR) are both forms of indirect digital capture. DR employs a rigid flat panel in which the intermediate light image is formed by, for example, a caesium iodide scintillator and the photons of light are detected and stored as electrical charge, which is then converted into a digital signal. The panel takes the place of the film but is connected directly to a computer. CR is a two-part process in which a phosphor storage plate takes the place of the film. When X-rays hit electrons in the atoms of the phosphor coating on the imaging plate, energy is transferred to the electrons, forcing them into a higher orbit where they become trapped. The phosphor stores the radiographic image as trapped electrons until it is transferred to a plate scanner, or reader, where it is scanned with a red laser. The laser frees the trapped electrons and as they fall back to their original orbit, they emit the stored energy as blue light. These photons of blue light are detected and converted into the digital image. The plate is then completely erased by exposure to a very bright white light. It takes less than a minute from starting the scanning of the plate to viewing the images on the computer monitor. CR is more versatile than DR as the plates are easily substituted for film in existing radiography facilities. With careful handling, CR imaging plates can be read and erased at least three thousand times before replacement becomes an issue; they maintain their sensitivity throughout. The resolution of the captured

68

Textile X-radiography

(a)

(b)

Figure 4.4 Radiographic images from another area of the same object as in Figure 4.3, (a) captured using a low cost X-ray scanner (scanned by Steve Dobson, University of York and reproduced by permission), (b) a film radiograph digitised using an industrial radiographic film scanner. (Private collection; © Sonia O’Connor, University of Bradford.)

image depends on both the quality of the storage phosphor and the scanner with which the film is read.

CR and textile radiography CR is very well suited to the radiography of textiles because the imaging plates have a much greater dynamic range than radiographic film, are very sensitive to low energy X-rays and can produce images with a wider exposure latitude than film while maintaining good contrast (Figure 4.5). Only plates and plate readers designed for the highest definition

industrial applications should be used. Currently CR does not produce images with as high a resolution as films such as Afga Structurix D4 and D7 or Kodak Industrex MX,2 but the results are very acceptable for most applications, including visualising individual threads (Figure 4.5). Higher resolution plates are being developed but even the current plates, used in conjunction with a microfocus X-ray system, should, if required, produce images in which individual silk fibres would be distinguishable. The imaging plates come in a range of sizes and are usually used in a film cassette to protect them from damage and from bright light which will partially

Textile X-radiography and digital imaging 69

improvements in the ability to visualise components of a textile artefact but which would be impossible to achieve with conventional radiography. Negative transform

(a)

(b)

Figure 4.5 Radiographed stitches, (a) detail, digitised film image, (b) detail, CR (screens used without a cassette). (© Sonia O’Connor, University of Bradford.)

erase the image before it can be read. At the low kVs used for imaging textiles, the cassette will filter the beam, differentially absorbing the lowest kV X-rays and reducing the image contrast. In addition, features of the surface of the cassette will be superimposed over that of the textile further degrading the image quality. This can be overcome if the X-ray unit is housed in an area with low ambient light levels. A windowless room lit only by safe lights is ideal but not necessary. Indirect artificial or daylight entering through the partly opened door of the room will not cause undue problems as long as there is just enough light by which to work safely. The cassettes can be taken to the X-ray unit, the plates removed for the duration of the exposure and repacked in the cassette for insertion into the plate reader. The plate, once removed from the cassette, is almost as thin and flexible as film and can be inserted into narrow openings or supported against curved surfaces. CR plates require shorter exposures than film; IPX plates, manufactured by GE Inspection Technologies, are nearly twice as fast as Agfa Structurix D7 film and ten times as fast as Structurix D4. This makes it feasible to collect high contrast images at energies as low as 5 kV without excessively long exposure times.

Digital image processing Once a digital radiograph has been acquired, the advantages of storing the image in a digital medium open up a wide range of possibilities for digital image processing. There are a number of simple manipulations which can often produce significant

A digitised radiographic image will usually be displayed in the same way as a conventional radiograph, looking rather like a photographic negative (Figure 4.6b). In order to enhance white or light grey areas (low penetration), within a region of high penetration (black) it can be useful to create a positive image of the whole radiograph (Figure 4.6c). This is a simple procedure in most imaging software, completed in a single operation. Some digital capture software automatically performs this operation on export and necessitates a negative imaging operation to be performed in order to ‘flip’ the image back to one resembling a conventional radiograph. Histogram manipulation The information within an image is commonly displayed as a 2-D histogram (bar chart) with the X-axis representing the range of greys, or levels of contrast, and the Y-axis indicating the frequency of that particular grey level appearing in the whole image. Altering the histogram can make certain details easier to resolve or remove elements that obscure other details. Images with truncated or compressed histograms (Figure 4.7a) can sometimes be improved by stretching the histogram so that the full range of greys offered by the bit depth are employed to display the image (Figure 4.7b). Mosaicing Mosaicing film radiographs of large objects is made difficult by four factors: distortions in adjacent images which produces errors when they are overlapped, variations in tonal quality between images, the actual physical size of the resulting mosaic and the degradation of information through the successive stages of producing internegative prints and, finally, an overall photograph of the mosaic (see Chapter 3, pp. 44–5). With digital images, all these problems can be addressed resulting in the production of mosaics with better retention of information, more consistent tonal range, fewer distortions and which can be stored digitally in a very small space. While any software for producing photographic panoramas can be used for linking two or three radiographs together, superior results can be gained by using a dedicated image processing package. Figure 4.8a is a series of six radiographs of a chasuble. These were then linked together

70

Textile X-radiography

(a)

(b)

(c)

Figure 4.6 Embroidery detail from a chalice veil (see Chapter 19), (a) photograph, detail, (b) radiograph, detail, (c) the same image after performing a negative transform operation, creating a positive image and increasing the clarity of the features. (Private collection; © Sonia O’Connor, University of Bradford; reproduced by permission of James Spriggs.)

(a)

(b)

Figure 4.7 A radiograph of a detail from the waist cloth in Figure 4.1, (a) ‘as captured’ image with an insert showing a compressed histogram reflecting the lack of contrast in the image, (b) image after processing to improve contrast with histogram insert showing the wider spread of greyscale values. (Private collection; © Sonia O’Connor, University of Bradford, and Jason Maher; reproduced by permission of Rosemary Payne.)

in NIP3 which removed distortions and reduced variations in the general tonal values between the different exposures in the X-ray mosaic. Other digital image processing Many other manipulations can be performed that improve the clarity of specific features, such as false

colour imaging, stereo anaglyphs (see Chapter 3, p. 52), composing HDR images, etc., or remove unwanted features. It is always worth experimenting with some of the predefined ‘filters’ in an imaging package as they may enhance the visualisation of key features of interest. Figure 4.9 shows the result of applying The Art Media filter ‘Black Pencil’, from the Effects menu of Paint Shop Pro (version 8.10), with

Textile X-radiography and digital imaging 71

(a)

(b)

Figure 4.8 Mosaiced radiograph of a miniature chasuble (see page 47), (a) six individual radiographs, (b) completed mosaic. (TCC IM 17.33 Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford, and Jason Maher; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

increasing strength to the heel of the shoe used in the construction of the nursery rhyme toy – ‘The Old Woman in the Shoe’ (see Chapter 1, pp. 3–4 and Figure 1.1, p. 4). Although intended as an artistic effect, the result has been to define the edges of features helping trace the long tacking threads along the top edge of the shoe, the construction of the rear seam and details of the heel stiffener, all hidden below the silk covering. Effects like this may be directional so rotating the image through 90 degrees before applying the filter may make different features more prominent. Other manipulations require a deeper understanding of image processing. Fourier transforms or fast

Fourier transforms (FFT) are particularly useful for removing regularly spaced features. In Figure 4.10, a Fourier transform has been used to remove the image of two sheets of corrugated plastic, orientated at right angles, providing a rigid support for the embroidery but functioning to disrupt its radiographic image.

Summary The use of digital imaging, coupled with radiography, opens up a wide range of possibilities for the investigation of textiles and can also simplify the

72

Textile X-radiography

(a)

(c)

archiving and sharing of radiographic images. While hardware and software to capture and manipulate digital images are developing rapidly, they are not particularly difficult to use once armed with the knowledge of basic terminology and an understanding of best practice. If the initial costs of migrating to digital image capture are prohibitive or the

(b)

Figure 4.9 Radiograph of the shoe heel from ‘The Old Woman in the Shoe’ toy, (a) unenhanced, (b) with some edge enhancement, (c) with greater edge enhancement. (YORCM: BA2573; © Sonia O’Connor, University of Bradford; reproduced by permission of York Castle Museum, York Museums Trust.)

expertise lacking, it is always possible to contract out the digitisation and DIP to companies specialising in digital radiography. With film-based radiography becoming less used and supported by the radiographic industry, the move to digital radiography is mirroring the change in conventional photography. The necessity to become familiar with digital imaging

Textile X-radiography and digital imaging 73 3.

NIP is a program developed jointly by the National Gallery, London, and the University of Southampton which is particularly useful for producing seamless mosaics. See http://www.vips.ecs.soton.ac.uk

References

(a)

(b)

Figure 4.10 Detail of embroidery, (a) radiograph with the cell walls of the two supporting sheets of plastic visible as a grid, (b) radiograph after Fourier transform, most of the cell walls have been removed making the details of the embroidery easier to view. The change in contrast is purely to aid visualisation and not a by-product of Fourier transform. (WA1954.90; reproduced by permission of Ashmolean Museum, Oxford; digital manipulation by Jason Maher.)

in general, and digital radiography in particular, has never been greater but the time taken to acquire these new skills will pay dividends in the investigation of textiles and the dissemination of information, now and in the future.

Notes 1. 2.

These comparisons were done as part of a joint study with Andy Chopping, Head of Photography, Museum of London Archaeology Service. Tests imaging aluminium castings at about 30 to 60 kV have shown that the latest generation of phosphor screens, IPS and IPS2 from GE Inspection Technologies, will perform equally as well in terms of high resolution as these films. Personal communication, Philip Morris, Regional Sales Director (UK) GE Inspection Technologies.

Byers, F. R. (2003). National Institute of Standards and Technology. Special Publication 500-252: Care and Handling of CDs and DVDs – A Guide for Librarians and Archivists. National Institute of Standards and Technology. Gonzalez, R. C. and Woods, R. E. (2001). Digital Image Processing (2nd ed.). Prentice Hall. International Organisation for Standardisation (2003). Information and Documentation. The Dublin Core Metadata Element Set, ISO 15836. International Organisation for Standardisation. Lang, J. and Middleton, A. (2005). Radiography of Cultural Material (2nd ed.). Elsevier. O’Connor, S. and Maher, J. (2001). The digitisation of Xradiographs for dissemination, archiving and improved image interpretation. The Conservator, 25, 3–15. Okkalides, D. (1996). Perception of detail and greyscale range in X-ray fluoroscopy images captured with a personal computer and frame-grabber. European Journal of Radiology, 23(2), 149–158. Reinhard, E., Ward, G., Pattanaik, S. and Debevec, P. (2005). High Dynamic Range Imaging: Acquisition, Display and Image-Based Lighting. Morgan Kaufmann. Technical Advisory Service for Images (2002a). Choosing a File Format. Technical Advisory Service for Images. Technical Advisory Service for Images (2002b). Using CD-R and DVD-R for Digital Preservation. Technical Advisory Service for Images. Technical Advisory Service for Images (2003). New Digital Image File Formats. Technical Advisory Service for Images. Technical Advisory Service for Images (2004). Generic Image Workflow: TASI Recommended Best Practice for Digitisation Projects. Technical Advisory Service for Images.

5 Image interpretation Sonia O’Connor

Introduction In the case of industrial and medical radiography, the subjects being examined and the range of structures and defects that are likely to be revealed are well understood. Image interpretation is also greatly aided by comparison with standards and reference images whose details have been confirmed through sectioning of sample objects or post-mortem examination. The radiography of cultural material is a different matter because the subjects of such studies are often entirely hand-made ‘one-off ’ pieces and, even if mass-produced, show unique patterns of wear. They are diverse in type, material, construction and state of preservation. Image interpretation of cultural material is further hampered where the conclusions cannot be tested by more invasive examination of the object. There is developing literature which is useful in the diagnosis of particular features of, for instance, paintings, watermarks on paper, ceramics and archaeological metalwork; Lang and Middleton’s book (2005) has made a major contribution in this area. The interpretation of textile images is, in comparison, as yet only in its infancy. It may seem obvious to state that the correct and thorough interpretation of any radiographic image depends on having a knowledge of both radiography and the object type in question. However, it is far too often the case when cultural material is radiographed that the commissioner of the investigation is relying on the radiographer to explain what the image shows – and the radiographer has no detailed understanding of the object. This has been one of the main stumbling blocks to the realisation of the true potential of radiography as a tool for textile studies and conservation. The progress of this research project has been largely due to the radiography specialist and the textile specialist working together with a willingness to acquire an understanding of each other’s fields. 74

Developing interpretation skills is not difficult but takes a little time and is essential to all those commissioning or using radiographs in their work. Radiographs should not be taken, interpreted and put away, but made available for consultation during the study and conservation of a textile. Making observations from an image, developing an hypothesis about the textile, testing and revising these ideas through examination of the textile and the image will gradually refine image interpretation skills.

Negative images It can be very confusing to try to understand the information captured on a radiograph at first, particularly because, with radiographic films at least, the image is exactly opposite to that which might be expected. In our daily lives we have come to accept that light transmitted through an object made from a semi-transparent material will vary in intensity in relation to the thickness of the material. The thicker areas will absorb more of the light and will appear to be darker than the thinner areas, which may transmit virtually all the light and appear almost as white, or bright, as their surroundings. This is exactly the situation that occurs with X-rays and, if X-rays were visible, the variation in intensity in the thicker and thinner areas of the same object would seem familiar and more intelligible. As X-rays are not visible, film has traditionally been used to convert the photons of energy into an image that can be recorded and viewed, albeit as a negative image (see Chapter 2, p. 15). Radiographs are occasionally reproduced in a positive form, particularly for publications targeted at a non-specialist readership. Figure 5.1 is a series of radiographic images of a detail from a corded quilt oblong (see Chapter 22, Figure 22.11, p. 286). The stitching of the channels and cords they contain are

Image interpretation 75

(a)

(b)

(c)

(d)

Figure 5.1 Detail of a corded quilt oblong, (a) photograph, (b) radiograph, negative form, (c) radiograph, positive form, (d) radiograph, positive form after DIP. (YORCM: BA3058; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

evident in both the negative and positive radiographs (Figures 5.1b and c respectively) but it is perhaps in the positive version that information seems more accessible. In this image, the cords look solid and the unfilled and unembroidered areas stand out brightly. A positive photographic print can be made using the radiographic film image as the negative. However, for the purposes of image interpretation this is not viable as this is time-consuming, skilled work and there is inevitably some loss in image quality. However, once a radiograph is in digital

form, transforming the image from negative to positive or vice versa is a quick and easy procedure that involves no loss of information (see Chapter 4, p. 69). It can, in fact, make certain details more obvious to the eye so it is worth studying the digitised radiograph in negative and positive form. Enhancing images The benefits of digitising the radiographs and exploring them using image software cannot be

76

Textile X-radiography

overemphasised which is why Chapter 4 is devoted to digital radiographs. Digital image processing (DIP) in particular can help in image interpretation by enhancing the visibility of features of interest in a number of ways (see Chapter 4, pp. 69–73). This need not require a comprehensive understanding of digital image manipulation algorithms. In Figure 5.1d a predefined ‘Art Media’ filter called ‘Charcoal’ from the Effects menu of Paint Shop Pro (version 8.10) has been applied to the positive image of the corded quilt radiograph. This applies a highlight and a shadow to features with significant changes in greyscale intensity as if the image were a textured surface lit with a low-angled, racking light. The cords seem less prominent but finer detail, such as the stitching, the weave and creases in the backing textile stand out.

●

●

●

Interpretation basics ●

At the simplest level, the brightest features in the radiographic image are formed by the parts of the object that have transmitted the least X-rays. If the material and density of an object are unvarying, this is then purely a function of thickness. The complexities in image interpretation arise from variations in all these factors and in the superimposition of information from all depths throughout the object. Bright features easily distract the eye from the more subtle tones where the real interest may lie so, after the initial excitement, it is important to undertake a systematic survey of the image, working from the ‘knowns’ to the ‘unknowns’ with logic and reason. It will not always prove possible to interpret every last detail and it is often necessary to present probable alternative explanations. The less well the object type is understood, the more scope there will be for misinterpretation. Despite pressures to provide definite answers, it is better to present the evidence and the range of interpretations that it supports. Counterintuitively, it is a sign of competence to be able to say ‘I don’t understand fully’. There are some important fundamentals to image interpretation which should be followed whatever the subject or circumstances: ●

●

Examine the radiographs with the object to hand so that each aspect of the object or the image can be compared directly. View the radiograph in optimum conditions on a light box following the guidance in Chapter 2 (p. 21). The contrast differences in a textile image can be quite subtle, even when taken under the

best circumstances. A directional light source will allow the object to be illuminated adequately without flooding the image with reflected light. Magnification is essential just as it is in the visual inspection of the textile objects. A low power binocular microscope or loupe can be used. To prevent scratching through contact with the lens, protect the radiograph with a clear plastic sheet or sleeve. If the image of the object does not cover the entire film surface, check the uniformity of the image background. Variations in background density, due to factors such as poor film processing or differential X-ray attenuation by a film envelope or cassette will also affect the object’s image (see Figure 3.7, p. 36). Check the image of the image quality indicator (IQI) for contrast and resolution of detail. This will give an indication of the penetrating power of the beam used and the probability of particular materials or structures being visible in the image. Orientate the radiograph with the object by identifying the images of obvious visible features, like buttons, seams, tears, metal fittings and pins.

Image interpretation is facilitated by identifying features on the front and back of the object. Once these are eliminated the interpretation of the hidden features can begin in earnest. Image density variations The first step of interpretation is to note variations in image density and to try to explain these in terms of the visible detail of the object, such as stitches, seams or layers of additional material, for example pockets or areas of appliqué. Figure 5.2a is the central roundel from the chalice veil discussed in Chapter 17. The roundel, as with all the embroidered elements, has been cut from its original textile and remounted on to a similar support fabric, to which a backing fabric was added. Thus, around the remounted elements where there are only two layers of fabric, there is little attenuation of the X-ray beam and the image produced is relatively dark. The embroidered areas are thicker and are, therefore, lighter in the radiograph. The area of the roundel hatched with gold metal thread work is not covered in silk embroidery but even here the additional absorption produced by the layer of original fabric is obvious. However, the face and hands of the Madonna and the face, hands and leg of the Christ child stand out from this. These features have been painted and the different pigments are absorbing the beam to different extents. Where the

Image interpretation 77

(a)

(b)

Figure 5.2 Central roundel from a chalice veil, (a) photograph, (b) radiograph. (Private collection; © Sonia O’Connor, University of Bradford; reproduced by permission of James Spriggs.)

78

Textile X-radiography

paint is missing (or had not been applied) the image is darker than the immediate surroundings. This led to the conclusion that the original textile was missing in these areas and that the paint had been applied directly to the support fabric beneath. Even though alerted by the radiograph, this was not obvious from visual inspection of the piece. The closely cut edges of the textile had frayed away and were now largely confused and obscured by the embroidery and the stitches used in its remounting. That the paint was applied directly to the support fabrics was, however, confirmed by tracing the path of a narrow band of thick weft threads. These threads can just be seen on the forehead of the Madonna in Figure 3.1a (p. 24). In the radiograph of this detail, they stand out below the convoluted white lines of the metal thread work because they have been highlighted by the paint (see Figure 3.1b, p. 24). This irregularity in the weft was not found in the original textile of the roundel but was identified, under low magnification, across the width of the support fabric. Understanding the gross variations in image density of this piece has allowed original and later features to be distinguished. Eliminating external features With complex images it is sometimes helpful to produce a transparent overlay for the radiograph on which the features can be marked as they are identified. With digital images, basic image software can be used to give colours to specific features or to add lines and symbols directly to a copy of the image as the interpretation proceeds. This approach was used in the initial study of radiographs of a patchwork coverlet (Figure 5.3). Dated 1718, this is the earliest dated English patchwork coverlet (Fenwick Smith and Osler, 2003). The radiograph, Figure 5.3b, is a confusion of information from the different layers of the coverlet. Adding grey lines over the image of the seams visible in Figure 5.3a makes it is easier to concentrate on the underlying and concealed features such as seam allowances, tacking stitches, patches in the backing fabric and tears and folds in the paper templates (Figure 5.3c). A similar technique using different coloured lines was used to identify and trace the direction of the seams and stitching in the various layers of Mary Burnett’s quilt (see Chapter 22, Figure 22.8, p. 283).

Characteristic images Although a radiograph renders three-dimensional objects in two dimensions, it is possible to gain some idea of the number and nature of layers in a textile

object, to speculate on the shape of features revealed and even to secure identification for some materials. Context is important, of course, and there is no substitute for having a comprehensive knowledge of different types of textile objects. However, to say, for instance, that ‘this filling must be wool because these are always stuffed with wool’ is a dangerous approach. The most important things to grasp are the implications of the fact that radiography reveals all the features of an object at once – the difference between looking at something and looking through it. So, for instance, a line of running stitch identifiable on the surface of an object by its very intermittency becomes, in the radiograph, a continuous, often slightly sinuous, line. This is clearly illustrated by the lines of quilting stitches and the signature of Elizabeth Watson both executed in running stitch on the quilt discussed in Chapter 22 (p. 282 and Figure 22.9, p. 284). It is also necessary to understand what sort of images different geometric shapes will form and how different these will be if the object is solid or hollow. Essentially solid flat objects look much as would be expected. A circular sequin will produce a bright circular shape against its textile support, an ellipse if imaged at an angle to the X-ray beam and a thin line if caught on edge. The radiographs of more three-dimensional objects, however, can prove rather more difficult to interpret. It must also be remembered that the position of a feature, in relation to the X-ray beam centre and plane of the film, will affect the size and shape of the image it produces. Captured at the film’s surface a feature will be life size. The further it is from the film, the more its image will be magnified, and the further it is from the beam centre, the more spread and distorted. Reconstructing depth First, imagine a plastic cube viewed directly from above. When radiographed under the beam centre, it will appear as a square of uniform image density as the thickness is unvarying. The surface furthest from the film will be slightly magnified, as if the sides of the cube were all slightly diverging so a drop in image density may just be noticeable towards the edges of the square. Now imagine this as a hollow but thick-walled cube. The image of this will be a square of the same image density as before but with a smaller darker square set within it. The difference in size between the nested squares will be equal to the thickness of the wall. The smaller square will be lighter than the background density of the radiograph; the difference in

Image interpretation 79 Figure 5.3 Detail of a patchwork coverlet, (a) photograph (© The Quilter’s Guild of the British Isles), (b) radiograph, (c) radiograph with overlaid lines. (© Sonia O’Connor, University of Bradford; reproduced by permission of The Quilter’s Guild of the British Isles.)

(a)

(b)

image density being determined by the combined thickness of the upper and lower walls. The next stage is to imagine the solid cube is painted on its top or bottom surface with a thin, continuous layer of metal pigment-based paint. Even though this layer will attenuate the beam significantly,

(c)

it may not be obvious that it is there unless the image can be compared with that of a second, unpainted cube of the same dimensions. All that will be seen, however, is a difference in image density between the two cubes and such a discrepancy could have several different causes. It just may be that the cubes are made

80

Textile X-radiography

of different plastics or that the second cube is not a perfect cube, but is shorter in the vertical dimension so, although it has the same radiographic outline as the first cube, the beam is passing through less material. If, however, the cube is painted on one side, even a thin paint layer may be very obvious in the radiograph. The path of X-rays passing down the very edge of the cube, although diverging from the perpendicular, will be almost parallel to the sides of the cube particularly if the surface is close to the beam centre. This means that, although the layer may only be a few microns thick, the X-rays may encounter several millimetres of paint as they pass through it at a very shallow angle. The result is that the paint layer appears in the radiograph as a bright outline along the edge of the cube. This edge effect does not only occur with vertical surfaces or with paint but can also be seen to a lesser extent on other angled surfaces as long as the coating is substantially more radio-opaque than the substrate. An example of this is the painted and gilded carving of the eighteenth century lindenwood settee hidden beneath later upholstery which Gill revealed through radiography (see Chapter 12 and Figure 12.3, p. 179). All this is quite simple to grasp until the cube is turned at an angle to the X-ray beam. Now it is no longer of uniform thickness and the image density varies in all directions. It may be helpful to visualise this by turning a playing die between the fingers. When the face with one painted spot (or ‘pip’) is perpendicular to the beam, its image would show seven spots – two parallel rows of three circular spots (from the lower surface of the die) and one in the centre. The spots on the sides would appear as intermittent dashes. As the die is turned so that one corner is pointing vertically (along the imaginary axis of the X-ray beam), spots become ovals overlying each other while the die is thickest in the centre and gradually thinner and thinner towards the corners. These are the basic image features of objects which are tapered, sculptured, solid or hollow. It must not be forgotten, however, that if the X-ray beam energy is too low to penetrate the material, these threedimensional objects will be reproduced only as white shadows, so losing the density variations which help to distinguish them from flat objects. This sort of exercise can be usefully extended to a number of other shapes. A sphere, for instance, will produce the same radiographic image from whichever direction it is viewed. It will be paler at the centre, where it is thickest, and will gradually become darker towards its edges. It is also worth considering what the radiograph of a circular cross-section rod

would be like or that of a tube (see the feather quill, Figure 8.16a) and to reason out how these images might be distinguished from those produced by square or rectangular section rods. Twist will also change the image. Imagine a long, thin, flat strip wrapped in a helical manner around a solid core so that the edge of the strip lies up against the edge of the previous turn. With the core removed, the strip will produce a radiographic image of apparently interlocking triangles as the strip passes across itself time and again. This is the characteristic image of a metal wrapped thread (Figure 10.4, p. 154). These are all rigid shapes but they form the building blocks for understanding the more fluid forms of textile structures. A yarn is, after all, a flexible rod-like structure of circular cross-section which may be flattened and twisted, curved and folded, to form a woven fabric, a knitted garment or a lace trim. When two identical yarns cross each other, or a single yarn is folded back on itself, the beam attenuation will be doubled. Which yarn is above which will not be apparent although context and visual inspection of the textile may provide the clues to resolve these questions. For example, examine the radiographs of a corded quilt oblong in Figure 5.1. Where the cords of the quilt pass through the channels, they are compressed so that they are wider than they are thick and lie perpendicular to the beam. At the ends of the channels, the cord is brought out between the yarns of the coarsely woven backing and then reintroduced through this backing into the next channel. Here the image density of the cord is different because the X-rays are passing through a greater thickness of cord. This is not only because, unrestricted by the channels, the cord is rounder in cross-section and therefore thicker but also because the cord is making a change of direction as it enters and leaves the quilt, briefly becoming parallel to the beam direction. Identification of materials and structures Many of the materials encountered in textiles vary in composition or density as well as thickness. They might be denser at the surface which may explain, for instance, the brighter edge often observed on the images of extruded plastic elements. The material may vary internally in density, structure or composition. This can produce a range of effects, for instance: ●

a dappled or striped image, as in baleen (Figure 8.14b, p. 121),

Image interpretation 81 ●

●

a ribbed or fluted surface like porcupine quill (Figure 8.16b, p. 122), a speckled effect with more radio-opaque particles as is often seen in cardboard (Figure 5.8c).

Correctly identifying any material by the interpretation of its radiographic image is dependent on some knowledge of the chemical and structural nature of that material. Equally, to interpret properly the structure of a textile object requires an understanding of textile construction techniques and those used in the fabrication of substructures. For instance, in the study of furniture this would include upholstery and woodworking techniques. Even then it is not always possible to correctly predict the exact nature of the radiographic image a material may produce or to interpret the signs of use, damage, reuse and repair which together make up an object’s biography. A further complication to image interpretation of individual components concealed within objects is that they are not on their own. For example, the image of a doll’s stuffing will be a combination of the individual elements of the stuffing on which may be superimposed the image of its body and its clothes. It is not uncommon when identifying possible baleen stays through radiography to find the characteristics of this material overwritten with the stitching of a seam and the weave structures of the outer fabric, interlining and lining of the bodice. The most useful aid to developing an understanding of these nuances is to make a reference collection of radiographs of objects of known materials, construction and history. Reference images As part of this research, radiographs were taken of samples of a range of woven and non-woven fabrics, stitching and seam types, fillings and other associated materials from bone buttons, fish scales and pearls to baleen, wood and bamboo. Similarly, radiographs were made of partial or damaged objects that could be thoroughly investigated so that the identification of the materials recorded on the images could be confirmed. Many of these are illustrated in Part 2 of this book. These images proved an invaluable resource, forming the basis for the image interpretation of the textile objects from private and museum collections by allowing the component parts and construction techniques to be identified with some confidence even when the objects could not be investigated further by more intrusive techniques. The following are examples of reference material which proved very useful.

Similar volumes of filling material samples were packed in polythene bags and radiographed six at a time, using the same exposure parameters, to provide some amount of consistency so that rudimentary comparisons could be made between them (Figure 5.4). These images demonstrated the characteristics of individual elements, towards the edges of the bags, and of the filling en masse in the thicker areas. Comparing the images of each bag also allowed some estimation of relative radio-opacity to be made. Figure 5.5 is a group of overlapping squares of a canvas. When radiographed, these produce interference patterns characteristic of the number of layers and the angles at which they are crossing each other. The machine chain stitched hem in Figure 5.6 shows the same effect as is created in the bottom right of Figure 5.5 where two layers lie at a slight angle to each other. Similar effects are seen in the radiographs of some of the sample seams in Chapter 9, in particular Figure 9.19, p. 140. The vertical folds in Figure 5.6 have a double outline where both layers of textile are caught on edge rising up to and falling away from the direction of the beam. Understanding these images proved useful in detecting additional layers and construction details in items such as bodices, petticoats and quilts. Samples were also made from scraps of fabrics and other materials to try to explain or replicate the characteristics of structures observed in the radiographs of textile objects. For instance, building a model from plastic foam, copper wire and paper helped in the understanding of the finer details of the mechanism of the seventeenth century nutmeg tape measure discussed in Chapter 19 (p. 247 and Figure 19.7, p. 246). The form of the seam in Figure 9.25a was at first puzzling. Interpretation of this radiograph suggested a rather unusual form of machine sewn seam and this premise was confirmed by making a sample with the same radiographic signature (Figures 9.25b and c, p. 144).

Effect of exposure on image interpretation The duration of an exposure and the beam energy used will determine what components are captured in an X-ray image and this can greatly affect interpretation of that image. Exposures appropriate for particular components may produce no image of others. However, just because something is not visible does not necessarily mean that it is not there.

82

Textile X-radiography

(a)

(b)

Figure 5.4 Samples of filling materials, (a) photograph, fillings prepared for radiography, (b) radiograph of samples (clockwise from top right) sheep fleece, horsehair, coir, cotton sheet wadding, cotton wadding, cotton noil. (© Sonia O’Connor, University of Bradford.)

(a)

(b)

Figure 5.5 Canvas test samples, (a) photograph, (b) radiograph. (© Sonia O’Connor, University of Bradford.)

Image interpretation 83

The radiograph taken to image the mechanism in the walking doll, Figure 20.6a (see p. 256), has provided a wealth of detail about the metal components but, apart from the outline of the head and shoulder plate, nothing of the body or clothes of the doll are

Figure 5.6 Radiographs of folds in a simple hem. (© Sonia O’Connor, University of Bradford.)

visible. Yet their presence is not disputed because they are either visible to be seen on the doll or the doll would collapse without them, and this is clearly not the case (Figure 20.1e, p. 250). Figure 5.7 is a mixed-media embroidery from a private collection, made in raised embroidery technique on silk with appliquéd textile and metal threads. It is about 300 mm high by 350 mm wide, mounted on card and was one of a set of three embroideries of uncertain date but perhaps of continental origin. It depicts the biblical scene The Offering of Abigail (I Sam. 25), a story which was popular during the seventeenth century. The embroidery was in a poor state when examined prior to conservation but its condition allowed the materials used in the construction and padding of many of the features to be identified. These included silk fabrics, a linen interlining, felt, velvets, fur, leather, card, horsehair, metal braid and thread work, wool wadding and brown paper. This embroidery was radiographed both to acquire images of these materials in the context of an embroidery and as a demonstration of the effect of varying beam energy and exposure time on the visibility of information. The results of the exposures made at 30 kV with times decreasing from 8 to 0.5 minutes are presented in Table 5.1. Those marked with an asterisk are illustrated in Figures 5.8a to d. At the longer exposure only the metal components are visible; the corrosion indicates these to be silver. These do show differences in image density caused by

Figure 5.7 Embroidery The Offering of Abigail (TCC 2740.1), photograph before conservation. (Private collection; © Textile Conservation Centre.)

84

Textile X-radiography

Table 5.1 Results of radiographic exposure tests on the visibility of detail from the left side of the embroidery in Figure 5.7. Those marked with * are illustrated in Figure 5.8 kV

Time

Images details

30

8

Only metal braid and metal purl elements are clearly visible.

30

4*

Wrapped metal threads are visible and clearly distinguishable from the braid and metal purl because they have a lower image density (greyer).

30

2*

The metal components and the thickest, densest areas of padding (for instance, in David’s horse) are visible, as are the cream kid leather of David’s hand and the larger particles of mineral filler in the card mount.

30

1*

30

0.5*

The metal components can no longer be distinguished by image density. The smaller particles of mineral filler in the card (or possibly the backing paper) are apparent. More of the padding is visible including: ● the horsehair beneath the grass; ● details of the cotton wadding (with its characteristic inclusion of seed coat fragments) in the horse and the bodies of the figures; ● folded paper forming the hips and legs of the soldier. The shaft of the soldier’s spear, the card of the tree bole and bough and the black foundation thread of the chenille forming the leaves, grass and outline of the horse have appeared. The leather boots and the outline of the figure’s clothes are becoming visible. All the metal looks the same. The background is blotchy, perhaps due to variations in the backing card and/or unevenly distributed adhesive. The linen interlining is visible across the image and the turned edges of the fabrics on the back of the board are seen to the left. The full extent of the padding is shown, silk embroidery can be seen around the heads and there is a laying out stitch or perhaps an underdrawing visible in the foliage of the tree canopy. The detail of the clothing and the horse’s mane and tail are recorded.

differences in thickness and, perhaps, composition, allowing the braid, purl and wrapped threads to be distinguished. What it does not show is any hint of the materials and structures hidden beneath the surface of the embroidery. As the duration of the exposures is reduced, fewer and fewer X-rays are reaching the film through these metals and their image becomes brighter and brighter. In consequence, the differences in image density become more and more difficult to see. On the other hand, the more ephemeral components of the image are revealed although, interestingly, the very obvious light-green patch by the legs of the soldier in Figure 5.7 did not become visible even when the image contrast was increased by reducing the beam energy to 20 kV. Interpreting gaps The radiographs of this complex object also illustrate one of the major pitfalls when interpreting the image of concealed features – assuming that absence of evidence is evidence of absence. For instance, because the linen interlining has left no trace in Figure 5.8c, it would be very easy to examine the radiograph and, without evidence to the contrary

from the object itself, assume that there was nothing between the background silk of the sky and the board. In Figure 5.9, a detail from the radiograph of the shoe in Figure 3.14 (see p. 44), it almost seems that there is a gap in the insole and that the nails between the heel, the sole and insole are straining to keep the components together! Of course, because this is a shoe, this clearly cannot be the case – there must be a layer here which is very much more radio-lucent than those surrounding it. Without such familiarity with the context, a very different interpretation could have been made. Sometimes threads, which on their own are apparently too ephemeral to produce an image on a relatively low kV radiograph, become visible where they lie across a fabric because the beam attenuation the fabric produces is added to that of the thread. This is the case with the channel stitching in the radiograph of a corded quilt oblong in Figure 5.1. The stitches appear as individual light dashes beginning and ending at dark gaps in the underlying fabrics. The thread is, of course, continuous in this small, tightly worked backstitch. Where the thread is visible between the holes, the image is actually formed by three layers of the thread (as the stitch loops back on itself between

Image interpretation 85

(a)

(b)

(c)

(d)

Figure 5.8 Radiographs of a soldier and David on his horse from the embroidery of The Offering of Abigail (TCC 2740.1) before conservation, taken at 30 kV for, (a) 4 minutes, (b) 2 minutes, (c) 1 minute, (d) 0.5 minute. (Private collection; © Textile Conservation Centre.)

86

Textile X-radiography

Figure 5.9 Detail of radiograph of the shoe in Figure 3.14a (see p. 44) taken using lead screen intensifiers. (TCC 92.10 Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

each pair of holes before it progresses to emerge again from the next stitch hole along) and is superimposed on three layers of fabric (the top and bottom fabrics forming the channels and a backing fabric). The tension of these stitches has distorted the weave of the top and bottom fabrics producing quite large stitch holes. Here only one thickness of thread and the backing fabric are interposed between the X-ray source and the film and they have hardly attenuated the beam at all, producing the intermittent quality to the stitch line. Occasionally, quite substantial looking yarns, threads and fabrics may not produce any visible image on the radiograph, even when added to that of other

layers of fabric. It all depends on the relative radioopacity of the materials involved. Figure 5.10a is a detail of two lines of an embroidered braid applied to the bottom edge of a lightgreen, silk sateen nineteenth century bodice. Notice the overcast stitching along the edges of the braid which attaches it to the bodice. These stitches pass through both the silk and the cotton twill lining fabric. Figure 5.10b is a detail of the inside of the bodice, also showing an area of the lower edge, and Figure 5.10c is a radiograph of that same area. There is a steel stay held in a cotton tape casing in Figure 5.10b. The stay is attached along the line of a dart in the lining fabric, the cut edge of which is seen to the right of the casing. The outside of this part, the left front panel of the bodice, would not have been seen during wear as it was overlapped by the right front panel. Consequently, the braid is terminated at this point as is the green silk fabric. The stitches securing the top edge of the braiding can be seen coming through the cotton lining on the left and the even thicker thread lying below this, to the left of the stay casing, is used to form two loops on the braids for the hooks on the edge of the right front panel. To the right of the image, the vertical line of machine stitching holds the selvedge edge of the green silk. The turning of the silk is visible along the bottom edge of the bodice and the raw edges of this and the cotton lining are covered by a broad cotton tape. The radiograph of the bodice, Figure 5.10c, was taken at 15 kV to image the textile components and, unsurprisingly, the steel stay has not been penetrated. The image of the braids is highly detailed and that of the green silk is very bright, possibly because it is a weighted silk (Brooks et al., 1996). The selvedge edge of the silk is well defined as is the frayed edge of its turning, visible towards the bottom right of the radiograph. This turning produces a significant increase in the attenuation of the beam because the thickness of silk is doubled. The cotton tape casing of the stay is also visible against the silk fabric but the cotton clearly is not as radio-absorbent as the silk, because the increase in beam attenuation is much less than that produced by two layers of the silk. The machine stitching and the hand stitching on the casing and along the cut edge of the dart are all visible in the radiograph. However, other components of the bodice have left little or no trace in the radiograph (Figure 5.10c). The thread forming the loops for the hooks is just discernible. The path of the thread securing the braid can only be inferred by the indentation of the upper edge of the braid itself and the image of the cotton

Image interpretation 87 Figure 5.10 Details from the lower edge of a green silk bodice, (a) photograph of the braids on the exterior, (b) photograph of the interior showing the stitching at the end of the braids, (c) radiograph of the same area. (TCC 8.31a Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

(a)

(b)

(c)

88

Textile X-radiography

finishing tape certainly would have been overlooked if its existence was not already known. The bodice’s cotton lining fabric is similarly very difficult to spot. Beyond the selvedge edge of the silk and in the tear on that edge, nothing of the cotton twill can be seen. Where the cotton overlies the silk, the edge of the dart is visible because this marks a change from one layer of cotton fabric to three. Elsewhere the only hint that there is more than a single layer of silk is the interference between the weaves of this and the cotton fabric which is producing a moiré-like pattern. It is interesting to note that this pattern varies in its character to the right and left of the steel stay. This is because although there is a dart in the lining, there is no corresponding dart in the silk. While the directions of the warp and weft in the cotton lining change, those in the silk do not and so the pattern of interference alters. What these examples illustrate is that the less familiar and more complex a textile object is, the more important it is to be cautious about the image interpretation.

Image artefacts Finally, it is important to remember that not all the features that appear in a radiograph may relate to the object itself. Some may be image artefacts, mostly caused by poor film handling and processing (see Chapter 2, p. 20). Faults that are also visible in the background to the image of an object, such as uneven development, or image artefacts produced by the use of inappropriate or damaged film cassettes (see Chapter 3, p. 35 and Figure 3.7, p. 36), can be easily discounted. However, they can easily be misinterpreted where they occur only within the object area. Some image artefacts can be very common. Processed film is susceptible to mechanical damage whenever it is handled for viewing, unless it is protected by a clear plastic sleeve (see Chapter 2, pp. 21, 22). Surface scratching and scuffing has the appearance of fine dark lines when viewed in transmitted light. Mechanical damage, chemical splashes and contaminated fingerprints, drying marks and fungal or bacterial growths are unlikely to affect both emulsions of the film identically. By viewing each side of the film in reflected light at a low angle, it is often possible to detect these image artefacts as anomalies in the surface. Other image artefacts, such as those produced by creases and pin holes in metal foil filters, dents in

lead screen intensifiers or air bubbles trapped on the film surface during development, will not be evident in reflected light. These are more likely to be mistaken for real features of the object, particularly if they do not occur very commonly. Figure 5.11 shows examples of image artefacts found by trawling through approximately 300 radiographs of textiles from various sources. Figure 5.11a shows the effect of arrested image development where two overlapping films have become temporarily stuck together during development in a shallow dish. Dish development, even of single films, is not recommended as it so often leads to uneven development and scratching of the film (see Chapter 2, p. 20). Film emulsion is particularly tender when it is wet and can easily become detached from its plastic base if damaged mechanically. As a result of this, scratches that occur while the film is wet can be quite large and different in character to those formed on dry film. The damage in Figure 5.11b, the radiograph of a set of pattern woven ribbons, was probably caused by the metal hanger of another film, rather clumsily introduced into the wash bath after the film was fixed. Lines have been gouged in the emulsion on one surface of the film only but this has caused sufficient loss of image density that the scratch appears as a whiter line across the image of the ribbons. The emulsion from the lower scratch has not come away completely but is displaced and folded back on the film, producing an intermittent, dark edge to the right of the scratch. The emulsion from the upper scratch was completely detached but has become stuck to the surface of the film, a short distance to the left, where it could be mistaken for actual damage to the end of the ribbon. Figure 5.11c shows a black crescent with soft edges that become increasingly diffuse towards the points. This is pressure damage and its colour (black rather than white) indicates that it has occurred to the film after exposure and before processing rather than before exposure. The mark, although it looks the size and shape to have been made with a fingernail, is the result of the large sized film becoming bent or kinked as it was withdrawn from a tightfitting, flexible film cassette. The last image artefact, Figure 5.11d, is the result of a rare phenomenon and it took some time to arrive at a positive identification for this, partly because even quite experienced radiographers had not seen it and also because its form can be so variable. This radiograph shows a detail of a doll’s head made from wax on composition, with a crack

Image interpretation 89

(a)

(b)

(c)

(d)

Figure 5.11 Examples of image artefacts on textile radiographs from various sources, (a) arrested development revealing the edge of an overlapping film, (b) scratches in wet emulsion, (c) pressure mark caused by kinking film after exposure, (d) static electric discharge marks. (© Sonia O’Connor, University of Bradford.)

90

Textile X-radiography

running up to and through one eye. To the right of the eye is a feature that looks like a small black invertebrate, or sea creature, with many legs and long trailing filaments. To the left of the eye is a cluster of dark grey branched and pointed structures which give the appearance of the composition being crazed with cracks. Examination of the doll could not explain these image features and they could not be located on subsequent radiographs of the same area. It was eventually established that these are static electric discharge marks. In a relatively low humidity environment, even slight friction can create a static charge on the surface of the film emulsion, for instance when the film is loaded or unloaded from its cassette or envelope. If the charge is high enough it can discharge to earth, producing discharge artefacts, or fog, in the processed image. The development of skill in image interpretation will come with experience. Revisiting images can be

very humbling with the dawning of understanding of detail which had originally appeared insignificant.

References Brooks, M. M., O’Connor, S. and McDonnell, J. G. (1996). The application of low energy X-radiography in the investigation of degraded historic silk textiles. In Preprints of the 11th Triennial ICOM-CC Triennial Meeting Edinburgh, Scotland, 1–6 September 1996 (J. Bridgland, ed.), pp. 670–679, James & James. Fenwick Smith, T. and Osler, D. (2003). The 1718 silk patchwork coverlet: introduction. Quilt Studies. Special Issue:The 1718 Silk Patchwork Coverlet, 5, 24–30. Lang, J. and Middleton, A., eds (2005). Radiography of Cultural Material (2nd ed.). Elsevier.

6 Assessing the risks of X-radiography to textiles Sonia O’Connor, with a contribution on DNA by Jason Maher

Introduction

Colour

One of the most common questions asked is whether or not radiography is safe to use on textiles. Rumours abound regarding possible degradation of fibres, the changing of colours or even that radiography can interfere with carbon 14 dating. None of these rumours is based on reliable or testable observations or supported by any published experimental data. They seem to arise, in part, from confusion between radiographic imaging and X-ray fluorescence (XRF) spectroscopy and perhaps a misunderstanding of the interactions that occur when X-rays interact with matter. This chapter looks at the impact that X-rays have on materials and the true nature of the risk involved. Bombarding material with photons of electromagnetic energy does cause changes. These changes may be temporary, such as the heating effect that microwaves have on water, or permanent as in the sterilising effect of UV light on bacteria or the disruption of proteins by gamma radiation. In the degradation of cultural material, the most familiar changes are those triggered by interaction of materials with light – photodegradation. The permanent changes produced by electromagnetic radiation in organic materials are mostly a consequence of molecular bond disruption. Different types of bonds are disrupted at different energies, making it difficult to generalise. It therefore cannot be denied that irradiating materials with X-rays causes changes, either through interactions with the X-ray photons directly or indirectly through their ionising effect. What these changes are and whether or not they will have a detectable effect will depend on the material, the energy of the X-rays and the dose involved.

X-rays, being ionising radiation, can displace electrons from the atoms of the material through which they are passing. Depending on the trajectory of the X-ray photon, its energy and the material involved, several things may occur. Electrons may be temporarily knocked out of orbit or electrons already ‘trapped’ in higher orbits may be freed. Irradiation with high energy X-rays can cause electrons to be emitted from some materials – a property that is utilised to advantage when imaging with lead screen intensifiers, undertaking electron emission or beta backscatter radiography (see Chapter 3). These same reactions can cause temporary or permanent colour changes through crystal lattice deformation in a range of gemstones, including beryl, corundum and quartz varieties, and is used commercially to improve the colour of stones such as amethyst (Shigley, 2000). Pough and Rogers (1947) irradiated approximately fifty natural and synthetic gemstones with a high intensity, 50 kV beam in experiments ranging from minutes to hours to explore these effects. Although the exposures and doses involved in imaging textiles are considerably less than those employed by Pough and Rogers it would be wise to seek advice when radiographing textiles with specific gems. The colour of yarns and threads dyed with organic or synthetic dyes does not depend on crystal lattice deformation and radiography should not, therefore, cause a colour change by this mechanism.

Dating Bombarding atoms with X-rays of the energies used in the imaging of cultural materials does not have 91

92

Textile X-radiography

any effect on the nucleus of the atom. The energies involved cannot disrupt the bonding of the protons and neutrons to form or change isotopes and/or influence the rate of decay of isotopes. Using X-rays to image a textile cannot therefore interfere with carbon 14 dating, which depends on comparing ratios of the carbon isotopes 12, 13 and 14. There are dating techniques that depend on electrons in crystalline material becoming trapped out of their natural orbit through time. This is caused by interaction with natural high energy radiation, such as gamma rays from the lithosphere (the earth’s crust) or cosmic rays from space. Samples of the material are treated to free these electrons. The energy emitted by the electrons as they fall back into their natural orbit is detected and measured to gain an estimate of age. This is the basis of optically stimulated luminescence (OSL) dating used in geology and in archaeology to date the formation of sediments, especially quickly formed deposits such as blown cover sands. It is also the basis of thermoluminescence (TL) dating, which is used to determine the date of the firing of ceramics. Brothwell and Pollard (2005) review these and other dating methods that may be applied to textiles or materials associated with them. Neither of these dating techniques is relevant to the study of organic textiles, but TL dating could conceivably be used on associated materials; ceramic beads for instance. Relatively low doses of X-rays can cause problems with TL dating. Haustein et al. (2003) carried out tests on samples by passing them through the security control X-ray machines at Dresden Airport and found that the radiation did influence the TL signal. However, these machines operate at energies of 160 and 140 kV. It is not clear if the larger doses of very low energy X-rays employed in imaging textiles will have a detectable effect on the TL dates obtained from associated ceramics. As a precaution, samples for TL dating should be taken before radiography.

animal remains, particularly bone. The more damaged the DNA, the less likely it is that characteristic sequences will survive and this reduces the possibility of such studies being successful. There have been very few studies related to textile objects or items of dress. That by Borson et al. (1998) looks at the origin of the materials used in the construction of an Anasazi feather artefact. DNA extracted from the baleen stays in the eighteenth century stomacher (see Barbieri, Chapter 14) provides evidence of a matriarchal line of North Atlantic right whale, Eubalaena glacialis, now believed to be extinct (Eastop and McEwing, 2005). There are no published investigations that evaluate the effects of radiographic studies on DNA in historical textiles. That X-rays can damage DNA in individual animal cells is well understood (Östling and Johanson, 1984; Singh et al., 1988). The ionising effect of X-rays can damage DNA in two ways, directly and indirectly. Direct damage is caused by liberated electrons with specific energies that can disrupt one or both of the DNA stands or the bases linking them. Other electrons disrupt non-DNA molecules forming free-radicals which can in turn cause indirect damage to the DNA. While radiography will have an impact on DNA structure in textiles, the effects of such investigations will be negligible compared to the DNA fragmentation that will have occurred over time by natural processes (hydrolysis, oxidation, alkylation, etc.) and as a result of degradation shortly after death ( Johnson and Ferris, 2002; Burger et al., 1999; Lindahl, 1993). Fibre preparation and textile processing may also exacerbate this deterioration. Ideally, samples for DNA analysis should be taken before radiography but at the low doses required for textile imaging, it is unlikely that such investigation would constitute a significant factor in DNA degradation.

Organic textile fibres and dyes DNA The study of deoxyribonucleic acid (DNA) provides phylogenetic information (identification to species and varieties within a species), sex and possibly even specific matriarchal lineages in animals (using mitochondrial DNA). If future examination of textiles and associated materials involve analysis of DNA, then the potential damage done to DNA by radiography must be considered. There is a growing literature relating to ancient DNA (aDNA), primarily focused on that found in

The deterioration of organic textiles and dyes by long exposure to visible light is well understood, producing colour loss and alteration, chain length scission and loss of tensile strength in the organic molecules of the fibres (Tímár-Balázsy and Eastop, 1998: 16–19). The shorter wavelengths and higher energies at the bluer end of the spectrum are particularly implicated in this form of damage and the risks rise even further as the wavelengths shorten beyond the end of the visible spectrum into the ultraviolet (UV) region. On the electromagnetic spectrum the X-ray region

Assessing the risks of X-radiography to textiles 93

follows on from that of ultraviolet light (see Figure 2.1, p. 13) and the very lowest energy X-rays1 have many similarities with highest energy UV light, including the ability to ionise the material through which they pass. It seems sensible therefore to expect exposure to X-rays to be even more damaging to dyes and organic textile fibres than UV light, particularly as they can penetrate deeper into or through textiles, which might otherwise be opaque to light, displacing electrons and disrupting polymer bonds as they go. To protect textiles from photo-degradation overall light levels are reduced as low as is practical to view the textiles, UV filters are employed and the duration of exposure is minimised by the use of proximity switches, rotation of displays and dark storage. Does it seem reasonable then to put textiles at risk of damage by exposing them voluntarily to X-rays? It is reasonable to assume that the rate of deterioration from exposure to low energy X-ray photons is comparable to that caused by UV light and that the mechanism of that deterioration is similar. Undoubtedly continuous exposure to X-rays would produce devastating results but the dose of X-rays received by textiles during imaging is insignificant when compared to the amount of photons of light received while an object is displayed for an hour, a day or a week, or is being conserved at the bench. In addition radiography has been employed in the recording of watermarks and works of art on paper for decades without any reported problems. Similarly paintings on canvas, very comparable to painted textiles, have been radiographed for over fifty years with no observed deleterious effects (Hassall, 2005: 112).

X-ray analysis X-ray fluorescence (XRF) spectroscopy is an analytical technique used mainly in the study of metals and minerals. Usually the object is placed in a vacuum chamber, X-rays are directed at it and the spectrum of emissions produced by the object is detected. The energies and intensities of these emissions will be characteristic of the materials present at, or near, the surface of the object. XRF has been used in the study of textiles to analyse pigments, mordents, metal threads or weighting materials in silks (Brooks et al., 1996). In studies of cultural material, conventional energy dispersive XRF (EDXRF) is often favoured over other analytical techniques for being ‘non-destructive’. In many forms of spectroscopy it is first necessary for a sample of the material to be atomised. With XRF

the material being tested is not destroyed, although some amount of surface preparation may be required depending on the purpose of the analysis. When detection of all but the lowest atomic number elements (below silica) and only qualitative or semiquantitative data is required, the analysis does not necessarily need to be performed under vacuum. This means that large objects can be analysed using external XRF probes or even portable, hand-held equipment, avoiding the need to sample objects too big to fit into the vacuum chamber or objects that might be damaged under vacuum. Such EDXRF systems have been used widely over many years without any report of deterioration of the irradiated areas of the objects that were analysed. However, radiation damage has been observed with some forms of XRF that involve devices to concentrate the X-ray beam onto a small spot, enhancing the intensity of the X-ray beam by factors of up to several thousand (Mantler and Klikovits, 2004). Concern over this damage prompted Mantler and Klikovits to carry out experimental work to test the nondestructive nature of conventional EDXRF on samples of two modern office papers (one bleached without chlorine), a cotton and a silk textile (both described as ‘white industrial’). Discs of the materials were held in the specimen holder of the spectrometer and irradiated under a range of conditions. The results were quite dramatic. Textile and paper fibres samples exhibited mechanical damage, brittleness and permanent yellowing. After irradiation, the yellowing of some of the modern papers appeared to fade but scanning electron microscopy revealed irreversible mechanical decomposition of the fibres. The papers were the most sensitive to damage, followed by the cotton. Interestingly the silk was least affected. As the first indisputable traces of yellowing were produced after only 10 minutes’ irradiation at 20 kV, Mantler and Klikovits’ experiment would seem to have dire implications for the non-destructiveness of radiography. When using slow-speed, high definition, industrial X-ray film with, for instance, the Hewlett Packard Faxitron cabinet X-ray unit, it is not uncommon for the duration of exposures at 20 kV to last several minutes. If an exposure test is made or a series of exposures taken to capture components of an object with widely differing radio-opacity, the total duration of exposure can easily exceed 10 minutes. Additionally, the Faxitron and the Siemens SRS 303 AS spectrophotometer used by Mantler and Klikovits both have a beryllium window on the X-ray tube so are designed to generate beams with a high output of low kV X-rays.

94

Textile X-radiography

However, the damage produced in Mantler and Klikovits’ XRF experiment is not observed in radiographed textiles because there is a huge difference in the dose received per unit area of the object during exposures of the same duration. This is because the intensity of the beam reaching the object in the XRF and the X-ray machine are very different. Beam intensity is directly proportional to the tube mA and inversely proportional to the square of the distance between the focus of the X-rays and the object. In an XRF spectrophotometer the working distance between the X-ray focus and object is a matter of a few centimetres and the tube current that caused damage after 10 minutes was 100 mA. Typically X-ray imaging of textiles in a Faxitron is done at a working distance of at least 0.5 m and at 20 kV the tube current is approximately 3 mA. If all other factors were equal, on working distance alone the intensity of the Faxitron beam would be at least 100 times less intense than the XRF beam at the surface of the object. Add to this the difference in tube current and the intensity of the Faxitron beam is well over 3000 times less than that delivered by the XRF spectrophotometer in the experiments. Therefore the level of damage produced in the XRF experiment should not be equated with that of low kV radiographic imaging.

Testing radiographed silk samples In the absence of literature relevant to the effect of low energy radiography on textile fibres, an experiment was designed to detect the results of X-ray induced deterioration in both the short and long term. In this pilot project,2 silk was selected as the test fibre because of its proven sensitivity to light-induced degradation. Silk fabrics were irradiated at 15 kV for ten times the duration of that required for imaging. There was no observable change in the colour of the silks before or after X-raying. Samples were then artificially aged and subjected to mechanical testing to compare the tensile strength with that of unirradiated samples of the same silks. Four samples were chosen to represent a range of production treatments and physical conditions on the off-chance that these factors might affect the rate or trajectory of any deterioration caused by the X-rays. Three of the sample silks were eighteenth and nineteenth century fabrics in various states of preservation. One was dyed black and weighted with iron and tin. The other two were pink and puce and both were weighted with tin and silicon. The fourth silk was a recently made, unbleached, undyed, Habutai silk test fabric. Six replicate tests were performed for

each sample but the results of the mechanical tests showed that the X-rays produced no significant difference in the extension or breaking strength of the silk either immediately or after ageing. It was concluded that the low energies and brief exposures involved in radiographic imaging could be assumed to be safe for imaging silk textiles, causing at worst a negligible deleterious effect.3 It is proposed in the future to extend this work to other organic textile fabrics.

Summary Radiography is generally regarded as a non-destructive investigative technique fundamental to the study of cultural material, although irradiating any material with X-rays can produce a range of changes, both temporary and permanent. A proportion of the X-rays will pass through an object without having any effect at all but those that do interact can cause temporary heating, fluorescence, ionisation and electron emissions or initiate irreversible chemical changes such as breakages in molecular bonds. What changes occur, and whether they are temporary or permanent, will depend on the materials involved and the energy range of the X-ray spectrum, while the rate of change will depend on the intensity of the irradiation. The changes that undoubtedly occur during the radiographic imaging of textiles are at such a low level as to be undetectable in terms of alteration of their physical or chemical properties. Compared to the deterioration caused by heat, exposure to light and continual ‘natural’ background radiation, radiography can be regarded as a low risk procedure for textiles as for other cultural material.

Notes 1. 2. 3.

Up to about 30 kV these are termed Grenz rays from the German word Grenzstrahlen meaning borderline rays. A collaboration between the author and Dr Paul Garside of the Textile Conservation Centre, University of Southampton. The publication of the full experimental details and results of this work by O’Connor, Brooks and Garside is forthcoming.

References Borson, N., Berdan, F., Stark, E., States, J. and Wettstein, P. J. (1998). Origins of an Anasazi scarlet macaw feather artifact. American Antiquity, 63(1), 131–142.

Assessing the risks of X-radiography to textiles 95 Brooks, M. M., O’Connor, S. and McDonnell, J. G. (1996). The application of low energy X-radiography in the investigation of degraded historic silk textiles. In ICOM 11th Triennial Meeting Edinburgh 1–6 September 1996. Preprints ( J. Bridgland, ed.), pp. 670–679, James & James. Brothwell, D. R. and Pollard, A. M. (eds) (2005). Handbook of Archaeological Sciences. John Wiley. Burger, J., Hummel, S., Herrmann, B. and Henke, W. (1999). DNA preservation: a microsatellite-DNA study on ancient skeletal remains. Electrophoresis, 20, 1722–1728. Eastop, D. and McEwing, R. (2005). Informing textile and wildlife conservation: DNA analysis of baleen from an 18th-century garment found deliberately concealed in a building. In Scientific Analysis of Ancient and Historic Textiles. Postprints from the first Annual Conference of the AHRC Research Centre for Textile Conservation and Textile Studies, Southampton 13–15 July 2004 (R. Janaway and P. Wyeth, eds), pp. 161–167, Archetype. Hassall, C. (2005). Paintings. In Radiography of Cultural Material (2nd ed.) ( J. Lang and A. Middleton, eds), pp. 112–129, Elsevier. Haustein, M., Krbetschek, M. R. and Pernicka, E. (2003). Influence of radiation used by the security control at

airports on the TL signal of quartz. Ancient TL, 21(1), 7–11. Johnson, L. A. and Ferris, J. A. J. (2002). Analysis of postmortem DNA degradation by single-cell gel electrophoresis. Forensic Science International, 126, 43–47. Lindahl, T. (1993). Instability and decay of the primary structure of DNA. Nature, 362, 709–715. Mantler, M. and Klikovits, J. (2004). Analysis of art objects and other delicate samples: is XRF really nondestructive? Powder Diffraction, 19(1), 16–19. Östling, O. and Johanson, K. J. (1984). Microelectrophoretic study of radiation-induced DNA damages in individual mammalian cells. Biochemical and Biophysical Research Communication, 123, 291–298. Pough, R. H. and Rogers, T. H. (1947). Experiments in X-ray irradiation of gem stones. American Mineralogist, 32, 219–229. Shigley, J. E. (2000). Treated and synthetic gem materials. Current Science, 79(11), 1566–1571. Singh, N. P., McCoy, M. T., Tice, R. R. and Schneider, E. L. (1988). A simple technique for quantitation of low levels of DNA damage in individual cells. Experimental Cell Research, 175, 184–191. Tímár-Balázsy, A. and Eastop, D. (1998). Chemical Principles of Textile Conservation. Butterworth-Heinemann.

7 Radiation safety Graham Hart

Introduction Throughout our lives we are exposed to natural and man-made sources of ionising radiation. X-radiography of cultural material should not increase levels of exposure for the operator or for anyone else. The safety procedures put in place to prevent accidental exposure should be in compliance with the legislation of the relevant country. This chapter aims to explain the principles of risk reduction for radiography and how exposure is measured and monitored in order to conform with current legislation in the United Kingdom.

Myths and legends Myths and legends surround the use of ionising radiation, much as they do with other spheres of work. Some of the more popular ones are listed here and will be addressed in this section: ●

●

●

96

‘When I use the X-ray unit, the radiation carries on bouncing around the walls once I’ve finished’ – this is true, but only for tiny fractions of a second! For all practical purposes, the radiation ceases once the exposure has been terminated. ‘When I use the X-ray unit, the material I’ve been X-raying and/or the whole room becomes radioactive’ – this is only true when materials are placed within nuclear reactors or other intense, high energy sources of particulate radiation. It does not happen when X-raying any materials. ‘If I stand behind this curtain/person/wall so I can’t see the radiation source, I’ll be safe’ – it doesn’t matter whether you can see the radiation source or not, only whether you are adequately shielded from the exposure; see the discussion on scatter and shielding in ‘Practical radiation protection’ (pp. 100–103).

●

‘All radiation is dangerous and will give me cancer/make me sterile/etc.’ – this is unlikely to be true, but is as much a question of personal approach to risk as science. It is examined in more detail in ‘Radiation and risk’ (pp. 97–100).

Justification, optimisation and limitation The use of ionising radiation under any circumstance needs prior consideration in relation to three principles, enunciated by the International Commission on Radiological Protection (ICRP) in their Publication 60: 1990 Recommendations of the International Commission on Radiological Protection (1990) – the principles of justification, optimisation and limitation. The first principle ensures that due weight is given to the potential risks of techniques using ionising radiations in comparison with other techniques which might be used. Before using such techniques it is important to be clear that, for the particular use, ionising radiation represents the best way to obtain the information needed. The rest of this book should make that decision much clearer. The second principle, optimisation, makes it clear why everything possible should be done to ensure that radiation doses to those involved in the technique are negligible or as low as is reasonably practicable. For the types of X-ray technique being considered here, those radiation doses should always be insignificantly small providing people always follow sound radiation safety principles which are outlined in this chapter. The third principle, limitation, places boundaries on the radiation doses that people can receive although this should never be a problem when using these techniques to radiograph cultural material. Nevertheless, all countries have enacted legislation to set down those limits, above which exposures are only justified in

Radiation safety 97

extreme emergency situations. These principles have found their way into a European Commission Directive (1996) which, in turn, was the basis of legislation in the United Kingdom called Ionising Radiations Regulations (1999). Although the particular legislation described here applies strictly to the UK, similar legislation appears in all member states of the European Union (EU) and in many other countries.

United Kingdom Ionising Radiations Regulations The aim here is not to present an exhaustive description of these Regulations. The Approved Code of Practice Working with Ionising Radiations, published by the Health & Safety Executive (2000) is a clear and thorough explanation of the Regulations and their application. The key features of the Regulations will, however, be summarised here (Table 7.1) since it is important that those working with ionising radiation understand the law within which they work.

Radiation and risk The key question is ‘how dangerous is radiation?’ Unfortunately, as yet, there is no definitive answer.

Opinions within the scientific community, let alone outside it, vary widely but can be broadly categorised into four positions: ●

●

●

●

‘Small doses of radiation are good for you’ – the so-called ‘radiation hormesis’ theory. ‘Radiation, unless in large quantities, is unlikely to be a cause of cancer or other detriment’ – based on the view that the body’s repair mechanisms will deal with any detrimental effects arising from small doses of radiation. ‘All doses of radiation are potentially harmful, although the probability of harm depends on the radiation dose received’ – based on the so-called ‘linear no-threshold’ theory. ‘Small doses of radiation may be extremely harmful ’ – particularly for internal radiation exposures.

Any of the four positions may be correct and possibly more than one of them at the same time. It is also possible that it may never be fully understood which is the case, as assigning the level of harm from one cancercausing agent, when cancer appears naturally at relatively high levels within the population, is quite difficult. However, the legal view enshrined in the Ionising Radiations Regulations is based on the third position. This linear no-threshold theory gives rise to the sometimes-heard statement ‘there is no safe

Table 7.1 Summary of Working with Ionising Radiations Ionising Radiations Regulations 1999 (Health and Safety Executive, 2000) Topic

Explanation

Implications

Prior authorisation & notification

Novel work with ionising radiation not covered by generic authorisations, or work being done in new premises may need to be notified to the Health & Safety Executive (HSE) and authorised by them. Before new work with ionising radiation begins, the employer needs to carry out a thorough assessment of the risks that will be involved in the practice – • what is the nature of the hazard? • what are the likely radiation doses? • how can they be eliminated or controlled? • what might go wrong (contingency plans)?

This area is complex and a Radiation Protection Adviser (RPA) should always be consulted to see whether the work falls within the remit of the Regulations. Note – this does not mean that every new X-ray exposure needs notification to HSE. This is a key part of the Regulations, and also a requirement of other safety legislation. The risk assessment must be written down and the key findings known to those involved with the exposure as part of the Local Rules (see below).

Prior risk assessment

(Continued)

98

Textile X-radiography

Table 7.1 (Continued ) Topic

Explanation

Implications

Restriction of exposure

This places a duty on the employer to ensure that all employees have their radiation exposure restricted ‘so far as is reasonably practicable’. The necessary Personal Protective Equipment indicated by the risk assessment must be adequate and fit for purpose. Classified radiation workers can receive up to 20 mSv/year non-classified workers can receive up to 6 mSv/year. Members of the public accessing places of work can receive only 1 mSv/year (although it is expected that they should receive only 0.3 mSv/year from any one source). Employers using ionising radiation need to have access to an RPA whose duty it is to provide the essential advice on compliance with the Regulations. All those engaged or directly concerned with radiation work must have received training to understand the basic principles of radiation protection and knowledge of the techniques they will be using. Areas where ionising radiation is being used may need to be designated as ‘Controlled’ (if doses might exceed 6 mSv/year) or ‘Supervised’ (if doses might exceed 1 mSv/year). RPSs need to be appointed by the employer to ensure compliance with the Local Rules (written systems of work for designated areas). Equipment being used to deliver or control radiation exposures must be subject to regular maintenance and quality assurance to ensure that it remains fit for purpose. This makes it clear that employees and others involved with the radiation exposure must not misuse or interfere with ionising radiation, and must exercise reasonable care not to expose themselves or others to radiation more than is reasonably necessary. Incidents must be notified to the employer.

There is a need to demonstrate this is being achieved, i.e. through suitable personal monitoring.

Personal Protective Equipment (PPE)

Dose limits

Radiation Protection Advisers (RPAs) Information, Instruction & Training

Designation of Areas

Radiation Protection Supervisors (RPSs)/ Local Rules/ Contingency Plans Equipment

Duties of Care

It must comply with the PPE Regulations; ask the Radiation Protection Supervisor (RPS) and/or RPA. There are very few classified radiation workers in the UK. Most research workers receive a small fraction of the dose limit. A foetus is a member of the public, so pregnant radiation workers are restricted by the 1 mSv limit. The RPA should be involved in new practices to help with risk assessment and contingency planning. The training needs to be documented and kept up-to-date.

Designated areas need to be demarcated to ensure access is restricted if necessary.

The Local Rules must include the contingency plans, the area they cover and the key working instructions to maintain safety. Maintenance and quality assurance results need to be documented.

This also links back to risk assessment, and the adequate training for, and supervision of, the work.

Radiation safety 99

dose of radiation’. While this statement may be true, it is not helpful, since there is no ‘safe’ speed for a car, and no ‘safe’ portion of peanuts (as there are people allergic to peanuts within the general population). It does suggest, however, that since risk is related to radiation dose received, precautionary measures should be taken to reduce that radiation dose to levels that are as low as reasonably practicable. So what is the risk of harm from radiation? The current view of the ICRP in their Publication 60 (1990) is that, for a working population, the risk of developing fatal cancer from radiation exposure is of the order of 5% per sievert (Sv). This figure may be incorrect but it is the best estimate thus far. However, it requires some explanation – not just for the level of the effect itself but for how radiation doses are measured. Radiation doses can be recorded in a bewildering range of ways but the most common units and their explanations are shown in Table 7.2. Radiation doses to any one organ of the body are usually expressed in gray (Gy), whereas the ‘effective dose’ in sievert is used in an attempt to compare doses to different parts of the body, or for ‘whole body’ doses. It should be noted that the various tissue weighting factors are subject to change as research into radiation effects continues. Under the latest proposals some tissues change their weighting factor slightly, although these have not, as yet, been finally agreed or promulgated into new legislation as this process usually takes some years. The detrimental effects of radiation exposure are usually divided into two types – stochastic (meaning random) and deterministic. Stochastic effects appear to occur randomly (although debate continues about genetic predisposition) but the probability of occurrence seems to

depend on the radiation dose. The current view is that this dependence is approximately linear in nature. The severity of the effect is seemingly independent of dose. The principal stochastic detriment is cancer, although hereditary effects are also possible. As has already been mentioned, the current view of the ICRP is that the risk of developing fatal cancer from radiation exposure is of the order of 5% per sievert. The risk of developing non-fatal cancer is thought to be a further 1% per sievert. Deterministic effects are different in nature. They only seem to occur when high doses of radiation are received in a short period of time and each effect only seems to occur above a certain threshold dose, worsening with increasing dose. These effects mostly occur because the radiation is killing cells faster than the body can repair or replace them. Some of the various deterministic effects are given in Table 7.3. Although deterministic effects only appear at unusually high radiation doses, the presence of Table 7.3 Deterministic effects of radiation Dose threshold

Organ affected

Nature of effect

100 mSv

Foetus in utero

200 mSv

Male gonads

500 mSv

Bone marrow

1 Sv

Bone marrow

2 Gy 5 Sv

Skin Whole body

Some mental retardation noted. Temporary loss of fertility. Temporary depression of blood count. Total loss of bloodforming cells. Radiation burns. Sufficient to kill 50% of population.

Table 7.2 Common units of radiation Unit

Name

Explanation

Gray (Gy)

Absorbed dose (D) Equivalent dose (H)

Sievert (Sv)

Effective dose (E )

The energy absorbed by the material (1 Gy 1 J/kg). The absorbed dose weighted for the various types of radiation, since they deposit their energy differently (H wr.D), where wr is the radiation weighting factor (which is 1 for X-rays – the lowest – but can be up to 20 for other types of radiation). The equivalent dose weighted for its effect on different types of tissue, since tissues vary in their response to radiation (E Σ wt.D), where wt is the tissue weighting factor (ranging from 0.25 for the gonads down to 0.01 for the skin).

100

Textile X-radiography

stochastic effects means that radiation exposure carries a risk of detrimental effect even at low doses. The fatal cancer risk of 5%.Sv1 can be converted to a risk ratio (5% a risk of 1:20). However, a sievert is a large dose of radiation, and radiation exposures are normally thought of in millisievert (mSv 103 sievert) or microsievert (μSv 106 sievert). If the linear nothreshold hypothesis is correct, the fatal cancer risk from 1 mSv is therefore 1:20 000.

Putting risk in perspective This risk needs to be put into some perspective. For example, everyone is exposed to varying amounts of ‘natural’ background radiation as part of daily life. Background radiation mostly occurs from a variety of entirely natural sources, and thus varies considerably from place to place. In the UK, a recent Health Protection Agency report (Watson et al., 2005) shows that the average background radiation dose is approximately 2.7 mSv/ year which derives from the following sources: ●

●

● ●

●

Around 15% from medical exposures with another 1% from other man-made sources. Around 13% from cosmic radiation at ground level (although this increases slightly if we fly). Around 12% from terrestrial gamma radiation. Around 10% from naturally occurring radioactive substances in our food, drink and our bodies. Around 50% from natural radioactive substances in the air, principally radon gas.

Within the UK, the radiation dose varies by more than a factor of 5 on a county-wide basis, largely due to variations in the concentrations of radon gas resulting from differences in the underlying geology. Even within the UK, the variation in radon levels means that ‘individual annual doses vary from about 0.3 mSv to a few hundred millisieverts in homes with the highest radon levels’ (Watson et al., 2005: 18). In the USA, the average annual radiation dose is around 3.6 mSv (National Council on Radiation Protection, 1987). This is slightly higher than the UK, principally due to increased average radon doses and a higher component from medical exposures. If other countries are included, however, the variation becomes much larger still. This is demonstrated clearly in the report of the United Nations Scientific Committee on the Effects of Ionizing Radiation (2000). Data from around the world presented in this report shows that even average radon concentrations vary from country to country by much more than a factor of ten,

and terrestrial gamma radiation varies by well over a factor of one thousand. So, given that the average background radiation dose in the UK is approximately 2.7 mSv/year, this should translate into a fatal cancer risk of approximately 1:7500. However, other aspects of our daily life also carry risks. For example, the data of the USA’s National Safety Council for 2002 gives a risk of death from all transport accidents as 1:⬃6000 (or 1.25:7500), with similar figures likely for the UK. The lifetime fatal cancer risk in the UK from all causes is currently approximately 26% (or 1950:7500), of which up to one third is thought to be caused by smoking (Cancer Research UK). Now it is not entirely reasonable to compare risks taken ‘voluntarily’ with those that might be encountered in our working lives. Nevertheless, there is a clear need to place what is currently believed to be the risk of radiation exposure in context. For all the above reasons, it is best to avoid radiation exposure where possible. Hence the need to justify practices that will result in exposures and to ensure that radiation doses, where they are justified, are kept to levels that are as low as reasonably practicable – the principles of optimisation and limitation.

Practical radiation protection There are three key practical principles that need to be followed to restrict exposure to ionising radiation – time, distance and shielding. When considering each of these principles, it is necessary to understand whether protection is being sought against the so-called ‘primary’ radiation (the radiation being emitted by the X-ray machine itself) or the so-called ‘secondary’ radiation, caused by the X-rays being scattered by any object in the path of the X-ray beam. Primary radiation usually has a known X-ray spectrum, whereas the spectrum of scattered radiation is usually unknown and changes as it encounters other objects. The intensity of scattered radiation is usually between 105 and 106 of the intensity of the primary beam, and often has a strong angular dependence, depending on the object doing the scattering. Typically five times more radiation is scattered through 150° relative to the direction of the primary beam than is scattered through 30°, i.e. there is much more backscatter than forward scattered radiation. The three practical protection principles are explained in Table 7.4. These principles must be applied to the practical situations surrounding each exposure encountered.

Radiation safety 101 Table 7.4 Practical protection principles Principle

Effect

Explanation

Time

Radiation dose received is proportional to the time of exposure.

Distance

Radiation dose is inversely proportional to the square of the distance from the source.

Shielding

Variable, depending on the energy of the source and shielding material.

Straightforward – the more time spent near a radiation source, the greater the radiation dose received. This is true for small physical sources, such as the primary radiation beam, but is not true for scattered radiation which tends to be more of an extended source that falls as 1/d rather than 1/d2 (where d equals distance). Even so, distance is a highly useful dose reduction technique, and operators should try to maximise reasonable distance whenever open radiation beams are used. Shielding increases with the density of the material used – thus materials such as lead, concrete, stone, steel and brick are to be preferred. As the energy of the radiation increases, the shielding value of any material decreases, i.e. 1 mm lead transmits 2 106 of an X-ray beam at 40 kVp, 4 104 at 70 kVp and 102 at 125 kVp.

From a safety perspective, the simplest situations occur when exposures take place within a closed unit, such as the Faxitron cabinet X-ray unit (see Figure 3.2a, p. 27). Systems like this are designed with sufficient shielding to ensure that all the radiation is contained within the cabinet. Providing the unit is subject to regular maintenance and routine monitoring to ensure its safety features and interlocks remain intact, everywhere outside the cabinet is always at background radiation levels. The situation in a dedicated X-ray room is very similar. The room should be designed such that the public dose constraint of 0.3 mSv/year cannot be exceeded by anyone outside the room. Some rooms are designed so that the X-ray tube and object are within the room and the control panel outside, or else such that the operators stand behind a panel protected with lead sheet and lead glass (as in all fixed-installation medical X-ray exposures). In either situation, providing the room has been designed and checked by a suitable Radiation Protection Adviser (RPA), as required by local regulations, such as the UK Ionising Radiations Regulations 1999, the operators should be in a position where no radiation dose of any significance will be recorded. Mechanisms of protection need to be most carefully monitored in situations where the radiation exposures are less controlled, i.e. where exposure to open radiation beams may occur, such as when a mobile unit is brought to an object in a workroom or gallery. In these situations, the RPA, Radiation Protection Supervisor (RPS) and users must always

be involved in assessing how the three principles best restrict radiation exposures.

Radiation monitoring There are two different types of radiation monitoring – area monitoring and personal monitoring – each using substantially different instrumentation. Personal radiation monitors are at their most useful for open beam radiation work, where the badge is likely to provide a good estimate of the radiation dose the wearer has received. They are at their least useful when the user is working with closed systems, where a good area monitoring system should largely obviate the need for personal monitoring. Area monitoring Area monitoring needs to be done on a routine basis when using contained X-ray units. Figure 7.1 shows a typical instrument for area monitoring. Monitoring should always be done carefully and according to a consistent monitoring plan. All monitors used need to be traceably calibrated annually and maintained to ensure they remain in a state where useful measurements can be made. Users should always switch the monitor onto the ‘battery check’ setting and leave it there for at least five seconds. The needle should be stable and in or above the indicated area. Any other response demonstrates

102

Textile X-radiography

that the battery needs changing. Only once the battery status has been established should the switch be moved to a measurement position. Most area monitors have an audible output. It is therefore much easier to use the eye to steer the monitor and the audio output to check for any response above background levels. The dial only needs to be checked if a positive response is detected. Results should always be recorded. Avoid statements like ‘not above background’ as this might be misleading if the background level is slowly increasing. It is always preferable to record real numbers. Personal monitoring

Figure 7.1 A typical radiation monitor used for assessing leakage radiation or scattered radiation doses from X-ray units. This is the RO-10 Model Ion Chamber manufactured by Southern Scientific. (Reproduced with kind permission of Southern Scientific Ltd.)

Personal radiation monitoring is only legally required for classified radiation workers, when it has to be carried out by a dosimetry service approved by the Health & Safety Executive (HSE). Nevertheless, it is common practice for those involved with open beam work to use personal radiation monitors as a way of demonstrating that radiation doses are indeed as low as reasonably practicable. It is important for the RPA to assess with the RPS whether personal monitoring will provide useful information and is thus to be recommended and, if so, to ensure that it will adequately record the energies and levels of radiation exposure likely to be encountered by the wearer. Figure 7.2a shows a typical body monitoring ‘badge’ using optically stimulated luminescence

Figure 7.2 Typical monitoring devices, (a) Luxel dosemeter body monitoring ‘badge’, (b) ‘ring’ TLD used for assessment of hand/finger radiation doses. (Reproduced with kind permission of Landauer UK.)

Radiation safety 103

these requirements ensures there will be no significant additional risk from exposure to ionising radiation and thus X-radiography can be safely used as an investigative tool for furthering the understanding of cultural material.

Acronyms HSE ICRP

Health & Safety Executive International Commission on Radiological Protection IRR Ionising Radiations Regulations PPE Personal Protective Equipment RPA Radiation Protection Adviser RPS Radiation Protection Supervisor UNSCEAR United Nations Scientific Committee on the Effects of Atomic Radiation Figure 7.3 A typical electronic personal dosemeter. (Reproduced with kind permission of Thermo Electron Ltd.)

(OSL) although it is also possible for personal radiation monitors to use film or a system called thermoluminescence (TLD). A ‘ring’ TLD can be worn at the base of the finger and used for assessment of extremity exposures (Figure 7.2b). Both types are worn for a given period (usually one to three months) and then returned for ‘reading’ by the issuer of the badge. The results are therefore only available some time after the exposure has occurred. Figure 7.3 shows one example of an electronic personal dosemeter. This type of instrument can be read directly and may also be downloaded to a computerised system. As these do not have to be returned for reading, they provide an instant assessment of radiation dose, usually with integral alarms which may be set to warn the user of higher than acceptable radiation levels. This type of monitor tends to be quite expensive.

Conclusion Everyone involved in X-radiography must conform to the local rules for the facility being used. These local rules will be designed to meet current legislation and compliance is a legal obligation. Meeting

References Cancer Research UK. http://www.cancerresearchuk.org/ aboutcancer/statistics/statsmisc/pdfs/cancerstats_ mortality.pdf EC Council Directive (1996). Laying down Basic Safety Standards for the Protection of the Health of Workers and the General Public against the Dangers arising from Ionising Radiation. 96/29/Euratom, 13 May. http://europa.eu .int/comm/energy/radioprotection/doc/legislation/ 9629_en.pdf Health & Safety Executive (2000). Working with Ionising Radiations. Ionising Radiations Regulations 1999, Approved Code of Practice and Guidance L121. HSE Books. International Commission on Radiological Protection (1990). ICRP Publication 60. 1990 Recommendations of the International Commission on Radiological Protection, Annals of the ICRP, 21, 1–3. National Council on Radiation Protection (1987). Ionizing Radiation Exposure of the Population of the United State. Report 93. NCRP. The Stationery Office (1999). Ionising Radiations Regulations. Statutory Instrument No. 3232. HMSO. United Nations Scientific Committee on the Effects of Atomic Radiation (2000). Sources and Effects of Ionizing Radiation. UNSCEAR. Watson, S. J., Jones, A. L., Oatway, W. B. and Hughes, J. S. (2005). Ionising Radiation Exposure of the UK Population: 2005 Review. Health Protection Agency – Radiation Protection Division.

This page intentionally left blank

Part 2 Exploring the X-radiographic features of textile objects Sonia O’Connor and Mary M. Brooks

This section presents a radiographic survey of different textile materials, structures, construction and condition. The aim is to intrigue and inspire as well as provide some basis for standard comparisons. Although this survey is wide-reaching, it was not possible here to cover every possible element or feature which may be encountered in relation to textile and dress. Radiographs are clearly not the sole means

of identification but they can provide corroborative evidence – or challenge preconceived thinking. The radiographic images provided here are intended to help identify features. Conservators and curators will need to develop their own atlas of photographs and radiographic images to reflect the nature of their own collections.

This page intentionally left blank

8 Materials

Every textile and dress curator and conservator is well aware that textile artefacts are rarely found without non-textile materials forming part of the construction or the decoration. This chapter explores the radiography of these materials including fibres, fillings and a range of non-textile materials.

Fibres It is possible to differentiate some fibres in their radiographic images. This brocatelle fragment (Figures 8.1a, b and c) from 1730 to 1760 has a blue silk warp and a yellow silk and a blue linen weft. The linen, being more radio-opaque, shows up more distinctly. In some cases, the way in which the fibre is spun and woven – as well as the nature of the finished artefact – effects how easily such information can be perceived. For example, in the radiograph of an unfinished patchwork coverlet worked over paper templates, the patches made of cotton fabric can easily be distinguished from those made of wool fabric (Figures 8.2a, b and c). These different absorption patterns reflect differences in fibre composition, structures and the different atomic numbers of chemicals absorbed onto the fibres such as mordants, dyes, weightings and other processing residues – or indeed soiling. Fibre characteristics seen through radiography, such as coarseness, level of organisation, crimp and degree of radioopacity, are useful for identification of fibres used for fillings. However, it must be remembered that there can be great variations within a single example, and, of course, fibres may be blended. Different species and different prepar-ation processes result in different fibre characteristics on the radiograph; for an example of this phenomenon, see the discussion of the eighteenth century silk stocking (see Chapter 16, p. 221).

The characteristic crimp in sheep’s wool shows up clearly in radiographs. Depending on the coarseness of the wool, it is possible to pick out individual fibres (Figures 8.3a and b). Silk from the domestic silk moth (Bombyx mori) appears smooth, dense and more flowing; individual fibres cannot be picked out (Figures 8.3c and d). This sample has been prepared for spinning and therefore consists of aligned straight fibres. Silk’s characteristic lustre is lost in the radiograph but, when viewed under magnification, soft white ‘spots’ become visible which may be traces of sericin ‘gum’. In contrast, the straight fibres in linen (Linum usitatissimum) appear coarser, harder-edged and stiffer (Figures 8.3e and f). The characteristic natural twist in cotton (Gossypium hirsutum) means the fibre retains some ‘loft’ and does not display the level of organisation seen in the silk and linen radiographs (Figures 8.3g and h). Its fineness means that individual fibres cannot be distinguished and the result is an overall impression of varying densities with only a hint of fibre direction. This cotton sample has not been fully processed and a lot of vegetable debris remains mixed with the fibres. More unusual fibres can also be identified through their radiographic images. The solid smooth outlines of the translucent white horsehair weft visible on the photograph can be differentiated from the cotton warp in this upholstery fabric (Figures 8.4a and b). Similar mapping can be undertaken with man-made fibres. These fibres absorb X-rays differently from natural fibres according to their chemical composition and cross-section. It is important to note that, as the maximum resolution of a scanner is approached, its ability to resolve individual fibres will be comprised. The resolutions of the images reproduced here are limited by the process of digitising the radiographs using a dedicated radiographic scanner (see Chapter 4, pp. 64–6), but finer detail may be seen in the films when viewed on a light box with a lens or low power microscope. 107

108

(a)

(c)

Exploring the X-radiographic features of textile objects

(b)

Figure 8.1 Brocatelle fragment 1730–1760, (a) detail of obverse, (b) detail of reverse, (c) radiograph of same area. (Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

Materials 109

(a)

(c)

Fillings An amazing variety of materials have been used as fillings for quilts, garments, dolls, toys, upholstered furniture and car seats. Radiography can aid in their identification without the need for excision thus

(b)

Figure 8.2 Detail of patches with paper templates in an unfinished patchwork; the patch in the upper left corner is white cotton and the lower right is red wool, (a) detail of right face, (b) detail of reverse, (c) radiograph from the same area. (YORCM: BA3009; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

enabling greater understanding of the artefact as well as possibly aiding dating and, perhaps, pinpointing places of origin. It also provides corroborative evidence which can help clarify sometimes confusing terminology. The series of radiographs in Figures 8.5, 8.9 and 9.11 were taken using identical 15 kV

110

Exploring the X-radiographic features of textile objects

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Figure 8.3 Fibre samples all to the same scale. Wool fibres, (a) photograph, (b) radiograph; silk fibres, (c) photograph, (d) radiograph; linen fibres, (e) photograph, (f ) radiograph; cotton fibres, (g) photograph, (h) radiograph. (© Sonia O’Connor, University of Bradford.)

Materials 111

(a)

(b)

Figure 8.4 Woven horsehair fabric Ardennais, John Boyd Textiles, UK, (a) photograph, (b) radiograph. (© Sonia O’Connor, University of Bradford.)

exposures and similar volumes of filling materials (except for Figures 8.9d and f, where smaller samples were used) to produce comparable information. Coir Coir fibre is obtained from the seeds of coconut palms (Cocos nucifera). After processing, it may be used as fillings for mattresses and, when rubberised, car seating (Royal Botanic Gardens). Comparing the radiograph of coir with horsehair is instructive (Figures 8.5a and b and Figures 8.5e and f ). Although they appear similar, in this sample the coir fibres seem to have both a larger diameter and a greater variation in diameter than the horsehair. Coir does not have the semi-circular curls and wire-like characteristics of horsehair. Cotton Colby (1972) notes that the fluffy fibres from the seed heads of cotton plants (Gossypium hirsutum) were used as fillings from the early fifteenth century. Until the nineteenth century, cotton fillings used in Britain were imported from India. They were then manufactured in Manchester using a shorter staple cotton than was used for spinning. By the mid-nineteenth century, quilts made in the north of England tended to use cotton wadding as it was easily available. Cotton was more rarely used in quilts made in Wales and southern England. However, some quilters in the north of England changed their practice in the early

twentieth century when the Rural Industries Bureau provided wool to County Durham quilters who were producing quilts commercially to generate income during a depression (Osler, 1987). This radiograph of a cotton boll (Figures 8.6a, b and c) shows lines of seeds set in the long cotton fibre hairs, each seed having a dense halo of shorter fibres around the seed coat. The characteristics of the individual fibres can barely be resolved in this image because their diameter is too fine but the halos around the seeds can be seen. Rae et al. (1995: 219) note that, in historic quilts which have been washed, cotton wadding tends to clump. Along with remains of cotton seed debris, this clumping may be visible in transmitted light. The bright dots of seed coat debris, sometimes with a halo of short fibres, and the clumping also show up well on radiographs, as in Mary Burnett’s many-layered quilt (Figure 22.3d, p. 279). Both the samples here are modern cotton waddings (Figures 8.3g and 8.5c). Quality can vary considerably and some can be heavily contaminated with plant fragments and other debris (Figure 8.5d). The quality of the more homogeneous cotton noil appears closer to the wadding which is often seen in radiographs of earlier quilts (Figure 8.3h). Cotton wadding for quilting was originally sold in pound packets but then became available in lengths coated with a papery skin (Osler, 1987); this may show up on radiographs where it is creased (Figures 8.5c and d).

112

Exploring the X-radiographic features of textile objects

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Figure 8.5 Filling samples all to the same scale. Coir sample, (a) photograph, (b) radiograph; modern cotton wadding with paper skin, (c) photograph, (d) radiograph; curled horsehair sample, (e) photograph, (f ) radiograph; kapok sample, (g) photograph, (h) radiograph. (© Sonia O’Connor, University of Bradford.)

Materials 113 Figure 8.6 Cotton boll, (a) photograph, (b) radiograph, (c) detail showing halo of fibres around seed. (© Sonia O’Connor, University of Bradford.)

Feathers Birds have several different types of feathers. These are constructed around a rigid central rachis which is hollow at the base (quill). Barbs branch out on either side of the rachis and are held together by hooked barbules forming the vane of flight feathers. Down feathers, characterised by fine fluffy barbs without hooks, form the undercoating of water birds such as geese, swans or ducks and are highly valued for their insulating properties. More expensive items, such as this 1870–1880s ‘Patent Purified Russian Down Skirt’ (Figures 8.7a and b) made by McLintocks, are filled with down alone.

In less expensive items, down is mixed with processed curled or crushed feathers. This radiograph shows that the filling of a 1950s eiderdown, visible even through multiple fabric layers, is a combination of down and short sections of chopped feather rachis, sometimes appearing as tubes (Figure 8.8a). These features can be seen more easily in radiographs of areas where the filling has thinned and settled (Figure 8.8b). Horsehair As well as being used in woven fabrics and judges’ wigs, horsehair is a traditional upholstery filling and was even used as a resilient, fire retarding stuffing for train seats. Darbyshire (2003: 80) notes that some

114

Exploring the X-radiographic features of textile objects

(a)

(b)

Figure 8.7 McLintocks ‘Patent Purified Russian Down Skirt’ 1870–1880s, (a) photograph, detail, (b) radiograph of the same area. (YORCM: BA5713; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museum Trust, York Castle Museum.)

(a)

(b)

Figure 8.8 1950s eiderdown, (a) radiograph through full thickness of eiderdown, (b) radiograph towards the edge where the filling is thinner. (Private collection; © Sonia O’Connor, University of Bradford.)

early nineteenth century wax headed dolls had leather bodies stuffed with horsehair. For use as an upholstery filling, the tail hair was plaited into long ropes which were boiled and allowed to dry for several months to set in the characteristic curl; this can be seen in Figures 8.5e and f. In contrast, smooth uncurled horsehair was used for the weft in woven structures (Figures 8.4a and b).

Kapok Kapok comes from the seed pod of the kapok tree (Ceiba pentandra), sometimes called the Java-cotton or silk-cotton tree. The seed hairs contain air reservoirs making them light and buoyant. The fibres are similar in diameter to cotton although lacking cotton’s distinct collapsed twist. Kapok was used for life jackets, upholstery and insulation but has now largely

Materials 115

been superseded by synthetic materials. From the 1920s, kapok was used in toys and teddy bears instead of wood-wool and sawdust as it was considered more hygienic (Cockrill, 2001: 6; Fawdry and Fawdrey, 1993: 88). When processed, it can, like cotton, contain seeds and debris but appears to clump less than cotton waddings. On the radiograph, it appears very similar to cotton (Figures 8.5g and h). Polyester Polyester (polyethyleneterephthalate) is a condensation polymerisation of ethylene glycol with terephthalic acid. First produced in the 1950s, it is widely used in toys, clothing and quilting. Unbonded polyester fibre, needle punched polyester and thermally or adhesive bonded polyester wadding (batting) are available. The radiograph of the unbonded fibres contrasts with that of the wadding (Figures 8.9a, b, c and d). The unbonded polyester is a mass of crimped but disassociated fibres without any obvious direction of organisation interspersed with thread-like structures of fibres lying parallel to each other and more or less crimped. The brighter and better defined lines in the radiograph therefore do not indicate variations in fibre diameter but variations in the organisation of the fibre mass. The wadding structure is very open but there are accumulations of material where the fibres cross and are bonded, either through heat or by adhesive, to form the chaotic web. This gives the radiograph a rather speckled and rigid appearance contrasting with the flowing lines of the unbonded polyester. Rubber Foam rubber was originally made from latex whipped with air but later foams, often used for upholstery and fillings for toys, were made from isocyanate or styrene-butadiene rubber. Today, such foams are usually polyurethane. This late 1950s home-made cushion has an outer cover decorated with letters and an inner bag for the filling of torn pieces of foam rubber (Figures 8.10a and b). The cushion has the characteristic smell of degrading rubber. At least two different foams are present, the majority being khaki-coloured while most of the remainder are lighter grey with a smaller pore size. A few fragments are now rigid, brittle and, in some cases, compressed and almost ashy. This seems to account for the variations in the radiographic image of the stuffing (Figures 8.10c). The majority of the fragments attenuate the beam very little but some crumbs are distinctly more radio-opaque; perhaps

these are from the grey foam. The compressed crumbs are the densest of all; these show as almost splinter-like bright fragments. Soya bean fibre Soya beans (Glycine max.) can be used to form a regenerated protein fibre. In the 1940s, Henry Ford was experimenting with rubberised soya bean fibre for use as upholstery filling in car seats (Brooks, 2005). Nowadays, soya bean fibre, usually including PVA, is being used commercially as a woven fibre and is available as a loose fibre for craft work and quilting. The radiograph of a contemporaneous soya bean fibre sold ready for spinning shows it to be a featureless fibre with some degree of organisation and no debris (Figures 8.9e and f ). Straw Straw is the term usually given to the hollow stems of cultivated grasses such as wheat. However, the bulk of the ‘straw’ used in the stuffing of toys and dolls may, in fact, be the long strap-like leaves of grasses commonly known as hay. This sample of hay does have a small component of grass stems (Figure 8.9g) but, as its radiograph shows, these are no more prominent than the long ribbon-like images of the leaves (Figure 8.9h). The sides of the leaves are quite bright which may be caused by edges rolling up as they dry. The overall impression of the radiograph is that of pairs of lines. This can give an impression similar to the hollow tubes of the chopped feather fragments (Figure 8.8b). Feathers clearly have other distinct characteristics which can be used to confirm their identification. Even when caught on edge, these leaves do not produce the very bright single line seen in the wood-wool radiographs (Figures 8.11b and d). ‘Straw’ does not have the curl of wood-wool although the leaves may be rolled, folded or twisted depending on how they were manipulated before use as a stuffing. For a radiograph of a doll stuffed with ‘straw’, see Figure 20.4b, p. 254. Wood-wool This fine grade of wood shavings, also known as ‘excelsior’, was used as a stuffing for toys and teddy bears. It was pushed into the pre-sewn bodies of teddy bears using a stout wooden stick (Cockrill, 2001: 6).1 Comparing these samples of 1950s wood-wool and new spruce wood-wool shows how the contemporary example has more sheen and curl whereas the older sample has lost its lustre and the strips seem flatter and shorter with broken ends; presumably all these

116

Exploring the X-radiographic features of textile objects

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Figure 8.9 Filling samples all to the same scale. Polyester fibre, (a) photograph, (b) radiograph; polyester wadding, (c) photograph, (d) radiograph; soya bean fibre, (e) photograph, (f ) radiograph; ‘straw’ sample, (g) photograph, (h) radiograph. (© Sonia O’Connor, University of Bradford.)

Materials 117

(a)

(c)

changes are due to aging, dehydration and compression (Figures 8.11a and c). These subtle differences are reflected in the radiographs although it would not be safe to use them as indicators of age as the variations in the width of the ribbons when they were produced might change such characteristics. The radiographs show stiff ribbon-like structures caught flat and on edge, the bright lines being formed where the X-rays have passed through from one edge to the other (Figures 8.11b and d). This 1960s teddy (Figures 8.12a and b) is tightly packed with wood-wool, an old-fashioned filling which contrasts with its modern man-made ‘fur’. The radiograph shows the orientation of excelsior ‘balls’ in the foot, a clear indication of the stuffing method. Sawdust may also appear as a stuffing for bears, dolls and toys (see Figure 11.12, p. 172).

(b)

Figure 8.10 Cushion, (a) corner detail, (b) sample of foam rubber filling, (c) radiograph of corner. (Private collection; © Sonia O’Connor, University of Bradford.)

Wool Osler (1987: 14–15) notes that both the earliest recorded and the earliest surviving wadded quilts made in Britain had wool fillings. This gives a warm filling although a great deal of preparation is required. As wool can ‘beard’ (i.e. pass through the outer layers of fabric), modern quilters tend to use a layer of muslin between a wool wadding and the top cover. This might show up on a radiograph as another weave system (see Chapter 5, pp. 81, 87–8). Radiographs of unprepared and unwashed sheep fleece will pick up not only the natural tufting but the highly radio-absorbent lanolin which shows up very brightly (Figures 8.11e and f); this contrasts with the characteristic image of treated sheep wool (see p. 107 and Figures 8.3a and b). In contrast to sheep

118

Exploring the X-radiographic features of textile objects

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Figure 8.11 Filling samples all to the same scale. Wood-wool 1950s sample, (a) photograph, (b) radiograph; woodwool fresh sample, (c) photograph, (d) radiograph; sheep fleece sample with lanolin, (e) photograph, (f) radiograph; mohair sample, (g) photograph, (h) radiograph. (© Sonia O’Connor, University of Bradford.)

Materials 119

(a)

(b)

Figure 8.12 1960s Teddy bear, (a) photograph, (b) radiograph showing excelsior balls in foot. (Private collection; © Sonia O’Connor, University of Bradford.)

wool, the radiograph of mohair from angora goats shows long, straight fibres (Figures 8.11g and h). Recycled materials Radiography may also reveal the presence of fillings made from recycled materials. Osler (1987) describes some quilts as textile middens containing everything from old woollen garments, blankets and flannelette sheets to reused net curtains. Mary Burnett’s quilt contains an earlier quilt reused as a filling (Figure 22.3, p. 279). Mary Ranby remembers that ‘old, worn, white or cream blankets were used in my home province in Canada to fill quilts during and after the war’ (Ranby, 1998: 4). Cut up shirts and socks were also used (Rae et al., 1995: 219). Today, Bangladeshi pictorial kantha quilts have rags under a layer of cotton fabric as the filling. The radiograph of a worn koala bear, appropriately known as ‘Scruffy Bear’, shows his head and legs are filled with wood-wool (see above; Figures 8.11a, b, c and d). However, his body is stuffed with woodwool around a central ball of rolled-up waste threads, some of wool and some of other fibres (Figures 8.13a, b and c). The bright atypical geometric shape, lower left, is a cardboard washer.

Supports, stays and substructures This section explores the radiography of materials used as supports, stays and substructures in corsets, bodices, wigs and bonnets as well as toys; see Chapter 10 for a discussion of materials used in surface decoration. Baleen Baleen comes from the sheets of keratinaceous material attached to the upper jaws of baleen whales which facilitate filter feeding. These sheets have a tripartite structure with flat plates of horny material to the front and back and a central layer of longitudinal tubules set in a protein matrix (O’Connor, 1987). Considerable processing, including splitting the sheets and removing the tubules, was necessary to produce flexible but firm stays for corsets. Radiographs of baleen often show longitudinal variations in densities that reflect the grain in the material although this may be masked by textile layers in some examples. In Figure 8.14, the radiograph shows these characteristic striations in a baleen stay despite the fact that it is held in a casing mounted onto a multi-layered

120

Exploring the X-radiographic features of textile objects

(a)

(c)

bodice. This stay is cut straight across the top and has a drilled hole (Figure 8.14b). The stay stitching can be seen passing through this hole in both the photograph and the radiograph. In contrast the stay in Figure 9.24b, p. 144, also secured within a casing, is not so obviously striated. It has trimmed and shaped ends but no hole. This bodice also has a baleen stay which has been incorporated in a dart simply by stitching across the dart to stop the stay sliding out of position. Radiography can also provide information about construction; in the stomacher discussed by Barbieri, a baleen stay became caught during insertion into the pre-sewn channelling and so the side peeled back (see Chapter 14 and Figure 14.6, p. 208).

(b)

Figure 8.13 ‘Scruffy Bear’, (a) photograph, (b) radiograph, (c) photograph of mixed filling. (Private collection; © Sonia O’Connor, University of Bradford.)

Steel stays The mounting method for steel stays shown in Figure 8.15a is similar for that of baleen stays (Figure 8.14a). Unless staining caused by the metal is visible, the two can be hard to distinguish without using a magnet or intrusive examination. When taking exposures intended to give good radiographic images of textile components, baleen transmits some of the X-ray beam so its image appears as light grey but steel stays do not so their image is bright white. Figure 8.15b therefore establishes that this stay with an off-centre stitching hole is made of steel. The staining seen in Figure 8.15a is sometimes incorrectly identified as blood. It is, in fact, corrosion of the steel,

Materials 121

(a)

(b)

Figure 8.14 Baleen stay in casing, (a) photograph, (b) radiograph. (© Sonia O’Connor, University of Bradford.)

(a)

(b)

(c)

Figure 8.15 Steel stay in casing, (a) photograph, (b) radiograph taken at 15 kV, (c) radiograph taken at 120 kV with lead screen intensification. (© Sonia O’Connor, University of Bradford.)

122

Exploring the X-radiographic features of textile objects

probably caused by perspiration. The radiograph shows irregularities in the area of the staining along the edge of the steel where there is an accumulation of rust (Figure 8.15b). In Figure 8.15c, taken at high kV using lead screen intensifiers (see Chapter 3, pp. 41–3), the pitting of the metal formed during this corrosion process is quite clear as is the non-ferrous metal caps on each end of the stay. Revealing such features not only allows the conditions of the components to be assessed more accurately but presents the opportunity to develop typologies which may help in the refining of dates and origins, perhaps even to specific manufacturers. Feather boning Feather boning was introduced as a substitute for baleen stays in the early 1880s (Helvenston et al., 1996). Like other natural substances such as porcupine quills, feather quills have a distinct radiographic image so it should therefore be possible to distinguish feather boning from baleen stays using a radiograph (Figure 8.16a and b).

(a)

(b)

Figure 8.16 Radiograph of(b) (a) feather quills, (b) porcupine quills. (© Sonia O’Connor, University of Bradford.)

Cane As well as being used in corsets, cane occurs in headgear. Both the cane and different thickness of metal wire could be imaged in the 1840–1850 ruched silk bonnet shown in Figure 1.3, p. 7 (Leath, 1995). Breaks, overlaps and turns in the wire could be seen (Figure 1.3). Paper and card The presence of paper or card in an artefact is often first revealed by speckling of the radiographic image due to mineral fillers; care must be taken not to confuse this with images of soil particles trapped in textile layers. Papers without such inclusions may be very difficult to detect although they will be attenuating the beam so reducing the overall image contrast. It may only be where the paper is creased or torn that its presence becomes apparent. Cut edges in paper have a continuous outline and folds are delineated with bright edges where the paper turns at an angle to the beam and the X-rays pass through a greater thickness. Creases and folds in textiles can be differentiated from those in papers as alterations in the direction of the textile structure will be visible in the former but not in the latter. This paper template in a hexagon patchwork has very unevenly sized particles of mineral filler (Figure 8.17). Note

Figure 8.17 Radiograph of paper template in situ in hexagon patchwork. (YORCM: BA3009; © Sonia O’Connor, University of Bradford. Reproduced by permission of York Museums Trust, York Castle Museum.)

Materials 123

the cut edges, creases, pin holes and tears and how the paper has torn away from the seams of the patchwork and become folded over on itself. Cardboard appears in many textile artefacts including dolls, toys and embroideries (see Chapter 20 and Chapter 5, pp. 81, 83–4). It was used as a support inside the velvet-covered brim of the late nineteenth/ early twentieth century hat in Figure 9.20, p. 141. The hard edges of the cardboard are visible in the radiograph at the outer edge of the brim and where the inner edge meets the junction with the crown.

Miscellaneous materials Other materials that might appear as part of the substructure of textile objects include reed, rattan and bone (see Chapter 1 and Figure 1.1, p. 4) as well as man-made materials. Plastics can vary greatly in density and radio-absorbency. Chlorinated plastics, for instance, will attenuate the beam far more than those made solely from carbon, hydrogen, nitrogen

and oxygen. The addition of mineral fillers can also greatly change the radio-opacity of these materials. Parts of plants and flowers may appear unexpectedly in textile artefacts; sometimes their presence reflects either use or context. Radiography revealed the unsuspected – and unexplained – presence of a plant stem fragment inside a chalice veil (see Chapter 17, Figure 17.6, p. 230). In other cases, they are an integral part of the artefact as with the lavender bag doll (see Chapter 3, p. 28, Figure 3.3, p. 29) and a perfume sachet containing dried rose buds (Figure 8.18). Other materials found as fillings may include bran or even tea as in the tea dolls made by the Innu people of Quebec and Labrador (Figures 8.19a and b). The filling material of the torso and head has a dappled appearance in the radiograph which might be tea but the arms and legs are filled with rolled-up bundles of textiles (Figures 8.19c and d). Radiography revealed the remains of seeds inside the Swinegate purse. Whether these are an accidental inclusion or reflect the use of the bag is unclear (see Chapter 3 and Figure 3.16d, pp. 48–9).

Figure 8.18 Radiograph, perfume sachet containing dried roses. (Private collection; © Sonia O’Connor, University of Bradford.)

124

Exploring the X-radiographic features of textile objects

(a)

(b)

(c)

(d)

Figure 8.19 Tea doll, (a) photograph, front view, (b) radiograph, (c) radiograph detail, chest, (d) radiograph detail, leg. (Private collection; © Sonia O’Connor, University of Bradford.)

Sonia O’Connor and Mary M. Brooks 125

(a)

(b)

Figure 8.20 Natural leather (left) and synthetic leather (right), (a) photograph, (b) radiograph. (© Sonia O’Connor, University of Bradford.)

Leather often appears as part of textile artefacts. Radiography may be used to distinguish natural from synthetic leather as in these samples (Figure 8.20). Although on the surface both have a convincing grain pattern, the radiograph of the ‘Genuine Leather’ reveals natural variation in the underlying structure where the imitation leather has a less variable texture and the layer of reinforcing open weave textile sandwiched within it is revealed.

Note 1.

For an explanation of this process, see the website of the traditional bear manufacturers HermannSpielwaren: http://www.hermann.de/fertigung/ fer_st3e.htm

References Brooks, M. M. (2005). Soya bean protein fibres – past, present and future. In Biodegradable and Sustainable Fibres (R. Blackburn, ed.), pp. 398–440, Woodhead. Cockrill, P. (2001). Teddy Bears and Soft Toys. Shire Publications.

Colby, A. (1972). Quilting. Batsford. Darbyshire, D., ed. (2003). Encyclopedia of Toys and Dolls. Quantum Books. Fawdrey, K. and Fawdrey, M. (1993). Pollock’s History of English Dolls and Toys. (Reprint of 1979 edition.) Promotional Reprint Company. Helvenston, S., Berryman, V. and Damstra, C. (1996). The legacy of Warren Featherbone. Michigan History Magazine, Sept/Oct. Available on http://www.fabrics. net/joan905.asp (accessed 19 July 2006). Leath, K. (1995). The Investigation and Treatment of a MultiMedia Bonnet, c.1840–1850. TCC 1943. (Unpublished Diploma Report, Textile Conservation Centre.) O’Connor, S. (1987). Identification of osseous and keratinaceous materials at York. In Archaeological Bone, Antler and Ivory. Occasional Paper 5 (K. Starling and D. Watkinson, eds), pp. 9–12, United Kingdom Institute for Conservation. Osler, D. (1987). Traditional British Quilts. Batsford. Rae, J., Tucker, M., Travis, D., Adams, P., Long, B., O’Connor, D. and Fenwick Smith, T. (1995). Quilt Treasures of Great Britain. Quilters’ Guild and Rutledge Hill Press. Ranby, R. (1998). Blanket approval (letter). The Quilter, Summer, 4. Royal Botanic Gardens. What is Coir? See http://www. rbgkew.org.uk/ksheets/coir.html (accessed 18 July 2006).

9 Threads, fabrics and construction techniques

This chapter explores the radiography of threads, fabrics and construction techniques.

Yarns and threads The structure of spun and plied yarns and threads can be seen on radiographs in terms of the number of constituent elements and whether, for example, they are plied (Figures 9.1a and b). The characteristics of the fibres are still expressed, for instance the dark green crewel embroidery wool has a very open texture and the individual wool fibres are visible in the radiograph detail (Figure 9.1c). Special effects, spinning irregularities, slubs and overspinning can be identified and mapped. This modern man-made fibre thread is Z plied and comprises two thin Z wrapped metal threads and a thicker thread which consists of floss wrapped in an S direction around a central core (Figures 9.2a and b). In the radiograph, the elements of the thread are distinguishable, the floss producing a distinct outer layer to the core of the thick thread. Interestingly, the metal wrapped thread has produced a very similar image density to the yarn’s organic components. However, as the whole of a yarn’s structure is seen on the radiograph identifying the Z or S direction is not possible so it is necessary to see the yarn itself to establish the spin and ply direction. Joining yarns and threads by knotting shows up clearly on radiographs as the thickness through which the X-rays pass is usually at least trebled in a knot. Splicing is a yarn joining technique seen in early Egyptian textiles. The linen was probably spliced together to form a rove before spinning (GrangerTaylor and Quirke, 2003). This fragment of linen mummy wrapping (Figures 9.3a and b) does not appear to show evidence of splicing but in the radiograph spliced yarns become evident. 126

Radiography is also a useful method – and sometimes the only method – by which the presence of internal stitching threads and their structure can be traced; for example, threads used for internal tacking stitches or concealed quilting may become visible (see Chapter 22). The critical factor is the relative radio-opacity of the thread and the surrounding materials. There are examples where threads can be seen on the artefact but not on the radiograph (Figure 5.10, p. 87). Others show up with startling brilliance such as the weighted silk repair thread in ‘The Old Woman in the Shoe’ toy (see Chapter 1 and Figure 4.9a, p. 72).

Cords and plaits Radiographs of cords and plaits are characteristic of their structures and methods of production. They are therefore helpful in identifying examples of such techniques which appear on radiographs but are actually contained or concealed. Figures 9.4a and b show a ‘cabled’ warp on an Egyptian archaeological fragment. Figures 9.5a and b are of three-stranded plaits on a thirteenth century purse excavated in Swinegate, York (see Chapter 3, pp. 48–9).1 This purse was made of alumtawed leather (now missing) with a silk lining, tassels and warpfaced piping. The piped edges of this purse were woven and attached in one process (Figures 9.5c and d). The weft thread passing through the side of the purse and the lining and back out to the piping can be seen as short loops on the radiograph.

Woven textile structures Radiography is an excellent method to document weaves, particularly those of hidden internal fabrics. Both surface and hidden weave features can be seen with new clarity. Details such as the ground weaves

Threads, fabrics and construction techniques 127

(a)

(c)

in pile fabrics, intersections between different weave types, regular and irregular variations in thread thickness and thread count, broken or missing threads and weaving faults all become visible. However, it is necessary to be cautious in interpreting this evidence.

(b)

Figure 9.1 Penelope crewel wool thread (2 ply S) made by William Briggs & Co., a blue Filo-Floss silk thread made by Pearsall’s and a yellow silk thread, possibly Filoselle, (a) photograph, (b) radiograph, (c) radiograph, detail of wool thread. (© Sonia O’Connor, University of Bradford.)

Plain weaves are generally straightforward to identify. This radiograph of a patch from a coverlet (Figure 9.6 and also Figures 8.2a, b and c, p. 109) shows a single layer of plain weave fabric in the centre. The number of layers in the turnings can be identified. Moiré effects can sometimes be produced

128

Exploring the X-radiographic features of textile objects

(a)

(b)

Figure 9.2 Modern thread in man-made fibre, (a) photograph, (b) radiograph. (© Sonia O’Connor, University of Bradford.) (a)

(b)

Figure 9.3 Twelfth Dynasty Egyptian linen mummy wrappings. The arrows identify the spliced threads, (a) photograph, detail, (b) radiograph, detail. (Department of Archaeological Sciences Teaching Collection, Catalogue No. 22 Egyptian sample N28; © Sonia O’Connor, University of Bradford; reproduced by permission of the University of Bradford.)

Threads, fabrics and construction techniques 129

(a)

(b)

Figure 9.4 ‘Cabled’ warp on a Twenty-first to Twenty-fifth Dynasty Egyptian linen mummy wrapping fragment, (a) photograph, detail, (b) radiograph, detail. (Department of Archaeological Sciences Teaching Collection, Catalogue No. 19b Egyptian sample T23; © Sonia O’Connor, University of Bradford; reproduced by permission of the University of Bradford.)

when two similar plain weaves with slightly different thread counts overlay each other (see Chapter 5, pp. 87–8). Even in plain weaves, one system may be largely obscured by the other as with weft or warpfaced plain weaves. Warp direction does need to be taken into account so that bias-cut fabrics are identified. It is also important to consider the relative radio-opacity of warp and weft threads. For example, twill weaves have a characteristic diagonal surface patterning which may be observed to some extent in the radiograph (Figures 9.7a and b). This may be obscured if the warp gives a brighter image than the weft (Figures 9.8a and b). This results in a more linear effect on the radiograph. Satin weaves cannot usually be distinguished from plain weaves although they give a ‘hazier’ image on the radiograph as some yarns are crossing other yarns for longer distances before being incorporated back into the main structure (Figures 9.9a and b). The

same effect can be seen in this self-patterned satin (Figures 9.10a and b). The following examples demonstrate the range and degree of information which can be obtained from radiographs of weaves. Figure 9.11a is a detail of a plain weave fringed Egyptian archaeological textile with a hemmed edge. The other two edges are cut. The radiograph, Figure 9.11b, shows details which might not have been noticed – or could have been misinterpreted – on the piece itself or on a photograph: ● ●

● ●

variations in image density in individual threads irregular banding across the fabric due to the spacing of the vertical threads. These are more closely packed towards the fringe. more regularly spaced horizontal threads the full extent of areas of damaged and broken threads

130

Exploring the X-radiographic features of textile objects

(a)

(b)

(c)

(d)

Figure 9.5 Details of a thirteenth century leather and silk drawstring purse, (a) photograph, plait, (b) radiograph, plait, (c) photograph, piping, (d) radiograph, piping. (YAT 1989.28 ctx 3204 sf 267; © Sonia O’Connor, University of Bradford; reproduced by permission York Archaeological Trust.)

Threads, fabrics and construction techniques 131

Figure 9.6 Radiograph of patch with paper template from a coverlet. (YORCM: BA3009; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

(a)

(b)

Figure 9.7 White twill weave silk fabric, (a) photograph; the warp and weft elements can be seen separately in the cut away sections, (b) radiograph; arrow indicates warp direction. (Sample No. 26; © Sonia O’Connor, University of Bradford.)

132

Exploring the X-radiographic features of textile objects

(a)

(b)

Figure 9.8 Lavender twill weave silk fabric, (a) photograph; the warp and weft elements can be seen separately in the cut away sections, (b) radiograph. (Sample No. 12; © Sonia O’Connor, University of Bradford.)

(a)

(b)

Figure 9.9 Brown silk satin weave fabric, (a) photograph; the warp and weft elements can be seen separately in the cut away sections, (b) radiograph; arrow indicates warp direction. (Sample No. 17; © Sonia O’Connor, University of Bradford.)

Threads, fabrics and construction techniques 133

(a)

(b)

Figure 9.10 White floral self-patterned satin weave ribbon, (a) photograph; the warp and weft elements can be seen separately in the cut away sections, (b) radiograph; arrow indicates warp direction. (Sample No. 29; © Sonia O’Connor, University of Bradford.) ●

●

a weaving fault (marked with an arrow in Figure 9.11b) where a weft thread has changed direction before reaching the edge of the piece (Figures 9.11c and d) strands from the fringe which, instead of being folded back on themselves toward the right, have been dragged to the left by the passing weft thread. They are secured in position by the next weft thread and loop back to the edge in the following shed. This confirms both that this is a wovenin fringe and that it is a selvedge.

Radiographs of complex weaves may provide additional insights into the various components of the structure. The plain weave ground of this velvet fabric (Figure 9.12) can be seen clearly. The red, green and undyed silk stripe which forms part of this plain weave wool medieval fragment (Figures 9.13a and b) is very distinct in the radiograph. The highly spun wool yarns are single rather than plied threads; note their characteristic twist and the consequent distortion that it gives to

the weave in the radiograph (Figure 9.13c). In the area of the stripe where different horizontal (probably weft) threads have been introduced, the weave changes so the vertical (probably warp) threads are paired. These paired threads are obscured in the photograph but are clear on the radiograph. The extra brightness associated with this horizontal stripe reflects the denser packing of these thicker threads although this absorption could also be influenced by dyes and their mordants. The presence of silk is unlikely to produce increased attenuation of the beam as its lower sulphur content makes it less radioopaque than wool. The dark colour and fragility of this degraded fulled medieval woollen fabric (Figures 9.14a and b) made it so difficult to determine its character that it was originally catalogued as felt but further study showed it to be a woven structure. The radiograph allowed the plain weave to be documented and revealed the details of the roughly sewn puckered seam and a totally unsuspected stripe, forming a ‘V’ across the seam. This stripe was woven in a similar

134

Exploring the X-radiographic features of textile objects

(a)

(b)

(c)

(d)

Figure 9.11 Details of Twelfth Dynasty Egyptian fringed linen mummy wrapping fragment, (a) photograph of corner, (b) radiograph of corner; arrow indicates weaving fault, (c) close-up photograph of weaving fault, (d) radiograph of weaving fault. (University of Bradford, Department of Archaeological Sciences Teaching Collection, Catalogue No. 16 Textile Sample No. 18; © Sonia O’Connor, University of Bradford.)

manner to the stripe in the fabric shown in Figures 9.13a and b. Creases visible on the photograph show up as white lines on the radiograph. Interesting phenomena appear in the radiographs of the silks (Figures 9.7, 9.8 and 9.9). A great deal more variation in the radiographic density of the textile appears than might be expected from visual inspection. There are various possibilities to be considered here.

This could relate to the differences in the thread such as diameter, spin, mordants, dye and beating variations in the packing of the thread into the weave structure. Sometimes these variations in the warp direction are in repeating bands. The radiograph may be highlighting variations in the warp which result from the dressing of the loom. In the weft direction, the variation in image density is often more frequent and irregular.

Threads, fabrics and construction techniques 135

Figure 9.12 Radiograph of modern woven velvet. (© Sonia O’Connor, University of Bradford.)

Van de Wetering (2000: 99–100) made similar observations when seeking methods to establish the warp direction of Rembrandt’s canvases using evidence from radiographs. The radiograph of the twill fabric has bands of lighter and darker greys; in the warp direction, these appear as blocks whereas in the weft direction they are narrower and more irregular (Figures 9.7a and b). These variations in density cannot be accounted for by variations in the thread count. They therefore relate to the properties of the threads themselves. In contrast, the banding in both directions in the brown satin weave fabric is so similar that, without other evidence such as a selvedge, it would not be safe to deduce the warp and weft directions on the evidence of the radiograph alone (Figures 9.9a and b). The white floral self-patterned satin ribbon presents a rather different situation. The radiographic density in the warp varies almost from thread to thread without a discernible pattern (Figures 9.10a and b). In the weft direction, however,

the variations in radiographic density give the appearance of an almost undulating surface where the grey shades vary from dark through mid-tones to light, and back to dark, in a regular manner. However, the sample was tensioned so that it lay flat against the film during radiography. Under microscopic examination of the image, it becomes clear that this phenomenon is also not due to variations in the individual threads but is caused by the spacing of the threads – and hence represents the beating pattern. Surprisingly, this variation in thread spacing is not obvious on the photograph and so the radiograph provides complementary information to that gained from averaged thread counts. The images of these silks provide a visual signature for the different fabrics and may open a new area of study into single-layered structures. It may, for instance, lead to research into the relationship of such patterns to different thread production techniques, weaving practices and equipment or give insights into the quality of woven fabrics. Another possibility is to use these visual signatures to identify the same fabric in different artefacts or reconstruct the layout of pattern pieces. The work that van de Wetering (2000) has done on Rembrandt’s canvases and Mottern et al. (1980) have done on the Turin Shroud are indicators of the potential here. Changes in weave structure do appear as regular repetitions on radiographs even if exact design features cannot be discerned. It is important to track such repetition so these features are not confused with damage (Figure 9.10). Features such as brocade elements worked in different colours may be evident on a radiograph allowing the passing of the threads on the back to be seen (Figures 9.15a and b). The potential of selvedges as a means of understanding both textiles and dress construction is amply demonstrated by the very varied articles contained in Lisières et Brodures (Cousin et al., 2000). Hidden selvedges may be characterised and recorded through radiography. Cut edges with fraying threads contrast with woven selvedges. The selvedge edges in the seams in Figure 9.16 are quite different from each other. That on the right is bright and hard due to the presence of additional warp threads while, in that on the left, it is possible to see each weft thread being brought to the edge and turning back into the next shed. The radiograph of the selvedge on this modern velvet highlights the final stabilising warp (Figure 9.12). This evidence of the variety of selvedge types could provide a further basis for the dating and provenance of textiles.

136

Exploring the X-radiographic features of textile objects

(a)

(c)

Radiography can also aid in condition assessment and highlights irregularities which are otherwise not easily visible. Weaving flaws can become more evident on radiographs and can be more easily mapped. The weaving flaw which helps establish the weft direction in the fringed Egyptian archaeological fragment discussed above can be seen far more clearly in the radiograph than in the photograph (Figures 9.11c and d). A horizontal flaw in the twill weave of the silk sample in Figure 9.7a can barely be seen in this photograph. However, its characteristics, including a related vertical thread, can be easily analysed in the radiograph (Figure 9.7b).

(b)

Figure 9.13 Medieval wool and silk fragment, (a) photograph, (b) radiograph, (c) radiograph; detail of colour stripe (YAT 1976.13 ctx 2573 sf 1384; © Sonia O’Connor, University of Bradford; reproduced by permission of York Archaeological Trust.)

Non-woven structures Radiographing non-woven structures can be equally rewarding. Radiographs of felt look very much like those of paper with random orientation of the fibres, reasonable consistency in overall image density but with variation in fibre density and length depending on the fleece type. If the felt has been reinforced with a layer of a woven fabric, this may well show depending on the relative radio-opacities of the fabric’s yarns and the surrounding felt matrix. Radiographing non-woven textile structures, such as sprang or knotless netting, may not reveal much

Threads, fabrics and construction techniques 137

(a)

(b)

Figure 9.14 Degraded fulled medieval woollen fabric, (a) photograph, (b) radiograph. (University of Bradford, Department of Archaeological Sciences Teaching Collection, Catalogue No. 16 Textile Sample No. 12; © Sonia O’Connor, University of Bradford.)

138

Exploring the X-radiographic features of textile objects

more than can be seen by eye except for carefully hidden joins and knots. However, it can prove useful should yarns which appear to be the same turn out to have contrasting radiographic signatures. This would allow the movement of specific single threads to be traced. In addition, shadows are lost in

(a)

(b)

Figure 9.15 Woven braid in an ecclesiastical style, (a) photograph, (b) radiograph. (Private collection; © Sonia O’Connor, University of Bradford.)

(a)

Figure 9.16 Radiograph of seams in two different layers of fabric in a chalice veil. (Private collection; © Sonia O’Connor, University of Bradford; reproduced by permission of James Spriggs.)

(b)

Figure 9.17 Detail of bobbin lace, (a) photograph, (b) radiograph. (Private collection; © Sonia O’Connor, University of Bradford.)

Threads, fabrics and construction techniques 139

radiographs so the resulting image may be easier to interpret. Even in the case of a thread which is visible, such as a thicker gimp in a lace, radiography facilitates tracking its path both across and through the structure (Figures 9.17a and b). On radiographs of knitted structures, many features become visible. This sample has a single rib at the bottom and a double rib at the top. The main

(a)

body of the sample is stocking stitch knitted showing various techniques for increasing and decreasing and a band of moss stitch (Figure 9.18a). On the radiograph of this knitted sample, joins appear as short, bright, sinuous lines (Figures 9.18b and c). They are strongly contrasted in the areas of stocking stitch but are less evident in areas of more complex patterning such as the double ribbing in the

(b)

(c)

Figure 9.18 Knitted sample in acrylic fibre, (a) photograph, (b) radiograph, (c) radiograph detail. (© Sonia O’Connor, University of Bradford.)

140

Exploring the X-radiographic features of textile objects

lower left-hand corner of Figure 9.18b. The cast-on edge at the bottom of the sample contrasts with the much denser structure of the cast-off edge. The single and double ribbed sections are characteristically different; the variations in image density in the double rib being more marked. The band of moss stitching is apparent with a slight diagonal element. As might be expected from the photograph, some of the increasing techniques are very obvious in the radiograph because of the holes this produces in the knitted structure. In the radiograph, the different decreasing techniques can be distinguished from each other and are more obvious than in the photograph. This makes radiography a useful tool when recording such shaping features, particularly when the knitting has a piled or napped finish as such detail is less obscured in the radiograph (Figure 9.18c). The knitted sides are characterised by their firm, regular and dense appearance, probably because they tend to curl slightly towards the back. This enables them to be differentiated from cut edges.

(a)

(b)

Construction techniques Radiography is an excellent tool for identifying and recording construction details such as seams, stitching, internal linings and support as well as non-stitching methods and materials which cannot otherwise be seen. Seams It is hard to separate seams and stitches but for the present purposes, seams will be examined first followed by the stitches used to create them. Information about seam types, turnings and their finishing can be obtained from radiographs. These may be indicators of the quality of construction and could possibly help establish whether an artefact was made domestically or professionally. Unsuspected seams may also become visible, giving insights into construction and/or repairs. The radiograph of the medieval woollen fulled fragment discussed above

(c)

Figure 9.19 Radiograph, hand sewn seam samples, (a) flat seam, (b) flat seam with finished edges, (c) French seam. (© Sonia O’Connor, University of Bradford.)

Threads, fabrics and construction techniques 141

revealed a hitherto unsuspected roughly constructed seam, possibly using a linen thread (Figure 9.14b). Mapping seams in individual fabric patches in an early patchwork coverlet was made easier by radiography (Brooks and O’Connor, 2005; 172–174). Unsuspected seams in the paper templates were also recorded. Hidden seams in the chalice veil were a clue to the history of reuse concealed within it; see Chapter 17. In the case of the musette, understanding the seams was critical to understanding the nature of the object and its provenance; see Chapter 27. In order to assess the evidence of seams on historic artefacts, a sampler was made using a cotton thread on plain undyed cotton fabric to show a range of commonly used seams and stitches. Figure 9.19a is a radiograph of a flat seam with cut, pressed, unfinished turnings while the flat seam in Figure 9.19b has one pressed edge finished with a running hemstitch and the other finished with blanket stitch. A French seam produces a more complex image (Figure 9.19c). The same cotton thread was used throughout but it is less visible on the right in this seam because here it is one out of six components and so the amount of difference of absorption is a much smaller than where it is used on the left.

(a)

(b)

A range of seams were identified in the historic textiles. The flat seam in this hat brim (Figure 9.20a) is invisible on the exterior because it is obscured by the pile of the bias-cut velvet fabric.2 The seam has pressed turnings, one cut and one with a selvedge (Figure 9.20b). The line of stitching crossing the seam is a pintuck which encircles the brim (Figures 9.20b and c). Figures 9.16 and 9.21a show flat seams with the turnings pressed to one side while a flat and fell seam (de Dillmont, n.d.: 6), with the turnings pressed to the left, turned under again and stitched down, is captured in Figure 9.21b. The seams at the edge of the hexagonal patch are overcast butted seams (Figure 9.6). Stitches Radiographic images allow different constructional stitches to be identified; see Chapter 22 for a discussion of stitching in quilts and Chapter 23 for stitching in shoes revealed by radiography. In some cases, it is possible to suggest that variations in the regularity of hand stitching and different methods of starting and finishing off sewing threads allows different makers to be distinguished (see Chapter 11, p. 169).

(c)

Figure 9.20 Machine sewn seams in the brim of the Cupar hat, (a) photograph of brim, (b) radiograph of brim, (c) detail of radiograph, pintuck. (Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

142

Exploring the X-radiographic features of textile objects

It is important to remember that elements of a stitch from both sides of the fabric appear to be lying together in a radiograph. The same stitch may look quite different depending on where it is in the textile and the angle from which it is being viewed. For example, when a running stitch is viewed from above, it appears as a continuous line on a radiograph; see Chapter 22, Figure 22.9, p. 284 and Figure 22.10, p. 285. However, if a radiograph is taken of a running stitch in a seam which has had the turnings opened and pressed back, the stitching is seen on the radiograph as a sinuous line linking the two fabrics (Figure 9.19a). Figure 9.16 shows two examples of backstitch in different layers of the same object. In the seam on the left side, the backstitch is very tight and neatly worked so the part of the stitch where the thread is looped back on itself has the three layers of thread superimposed on each other in the radiograph. The stitching of the seam on the right is very different. The stitches are larger, looser and less carefully worked. The threads in the loops mainly do not lie on top of each other to the same extent. This produces an image

(a)

which is very different in character to the left-hand seam which looks more like a row of bright dots. The right-hand edge of the turning in Figure 9.19b is finished with a line of blanket stitch. Knots and stitches used for securing threads at the start and end of a line of stitching are normally hidden within a textile artefact. Mapping them on radiographs can provide invaluable information about the

(b)

Figure 9.21 Radiographs of machine lock stitched seams, (a) flat with the turnings pressed to one side, (b) flat and fell. (© Sonia O’Connor, University of Bradford.)

Figure 9.22 Machine stitching, (a) chain stitch, (b) lock stitch, (c) lock stitch with too tight bobbin thread, (d) lock stitch with too loose bobbin thread. (Jason Maher after Dillmont (n.d.): 26–27.)

Threads, fabrics and construction techniques 143

direction in which stitching was carried out. Differences in the way a thread is secured and finished can also suggest different hands at work. The radiographic image of the tacking stitches in one patch from a nineteenth century quilt shows the level of information that may be obtained, including the direction in which the stitching was worked, the evenness of the stitches, the knot used at the start and the overcasting used at the finish (Figures 8.2c, p. 109 and 9.6). Machine stitching is recognisable on radiographs not only because of its regularity but because of the ‘punched’ appearance of the needle holes. Different machine stitches can be distinguished from each other, the most common being chain stitch and lock stitch (Dillmont, n.d.: 26–27).3 The less secure chain stitch was a single thread looped back on itself (Figure 9.22a). This was used on a variety of sewing machines from 1795 onwards with varying degrees of success. The more effective lock stitch, two interlocked threads, was developed in the 1830s and 1840s (Figure 9.22b) although chain stitch machines continued to be produced (Head, 2004: 13–14). However, care may be needed to distinguish the radiographic images

(a)

(b)

of these two stitches. Machine stitches are not always entirely regular in their spacing, for instance when passing over a bulky seam and other irregularities may result from mechanical problems. Differences in the tension between the upper and lower thread of the lock stitch can change the character of the stitch altogether (Figures 9.22c and d). Machine chain stitch is an asymmetric stitch (Figure 9.22a). A line of chain stitching may look like lock stitch on the upper surface (Figure 9.23a) but the lower face is very different showing a series of loops with each new loop emerging through the centre of the previous one (Figure 9.23b). When radiographed from above, the loops in the chain stitch dominate the image (Figure 9.23c). When seen from the side, securing a seam with the edges turned back, the asymmetry is evident. In this radiograph, the loops are to the right side of the seam as the single thread passes through – and links – the two fabric pieces (Figure 9.23d) thus identifying the side from which the seam was sewn. In a properly tensioned lock stitch, the interlocking point lies in the centre between the two threads.

(c)

(d)

Figure 9.23 Machine chain stitch, (a) photograph, top view, (b) photograph, bottom view, (c) radiograph from above, (d) radiograph from the side. (© Sonia O’Connor, University of Bradford.)

144

Exploring the X-radiographic features of textile objects

However, when the tension is not equal, the interlocking point becomes displaced to one or other side of the seam (Figures 9.22c and d). Figure 9.24a is a radiograph of a line of lock stitching in two layers of fabric imaged from above. In some places the two threads lie absolutely on top of each other but elsewhere they are laterally displaced so both threads are visible. Figure 9.24b shows a flat seam from above so that the lock stitch is seen from the side; the upper and lower threads that form the stitch can be seen on both sides of the seam as they pass through – and link – the two fabric pieces. Figure 9.24c is also a lock stitch seam but the fabrics that have been joined are no longer in continuous contact with each other. Although they are still holding, the stitches have been pulled and extended. This may be due to tension created by the sheer weight of the piece. This sort of distortion has been observed in large pieces such as quilts and patchwork coverlets which have been displayed vertically. Radiography therefore provides a method for documenting and monitoring both construction methods and display stresses in stitches and seams. In Figure 9.25a, the thread tension was uneven so the interlocking point is to the right of the seam. In this radiograph, there is a bright edge to each fabric on either side of the dark line indicating the join. These bright edges are bounded by the stitching and therefore have a scalloped appearance. This seam is

not a flat seam but is apparently the result of improvisation with machine sewing substituting for what would normally have been a butted overcast hand stitched seam. First, the edges of the two fabrics were turned back and tacked to give a finished edge before they were joined together; a line of tacking stitches can be seen to the left of the seam. Next, the seam was made by placing the two fabrics face to face and stitching just inside the turned edges through the four layers of fabric. When the seam was opened and pressed, a small raised ridge formed on the underside of the seam and this has given rise to the bright edges. This seam type is illustrated by the worked sample in Figure 9.25b which shows the ridge in the underside. On the right side, this seam is indistinguishable from a normal flat seam (Figure 9.25c). With this seaming technique, unless the join is particularly bulky, it may not be distinguishable from a flat seam without the use of radiography.

(a)

(a) (b)

(b)

(c) (c)

Figure 9.24 Machine lock stitch, (a) radiograph, from above, (b) radiograph, from the side, (c) radiograph, from the side of seam under stress. (© Sonia O’Connor, University of Bradford.)

Figure 9.25 (a) Radiograph of lock stitched seam, with interlocking point displaced to the right; photographs of seam sample replicating features of this seam (b) from below, (c) from above. (© Sonia O’Connor, University of Bradford.)

Threads, fabrics and construction techniques 145

Pins, thread and needle holes Pairs of holes separated by a short distance suggest the use of pins. This radiograph of a bodice shows where the pins were placed during the construction of the darts (Figure 9.26a). Threads hidden within an artefact, such as tacking threads, tailors’ marks, gusset reinforcements and so on, may be revealed by radiography. Where the threads are not themselves visible in the radiograph, their passage may sometimes still be inferred through the presence of regular distortions in the fabric. These result from both the puncturing of the woven structure by the needle and the displacement of warps and wefts caused by tension from the sewing thread. Radiography can also record alterations when these distortions persist after the removal of the thread (Figures 9.23d and 9.26b). Radiography could be very useful in documenting this type of constructional information if it is feared that such evidence might be lost, for example during a wet cleaning process.

(a)

Caution is nevertheless necessary where visual verification is not possible. Distortions may indicate the presence of threads which have not formed a significant image on the radiograph because they are much more radio-lucent than the fabrics which surround them rather than threads which are actually missing. The stitching on a carefully worked Indonesian reverse appliqué panel can be seen with the naked eye but do not appear on the radiograph (Figure 3.8, p. 37).

Other construction methods and materials It is also feasible to detect the remains of adhesives and possibly stiffening solutions. This Chinese painted silk furnishing fragment (Figure 9.27a) was protected from the light on the right side so retains its original sheen and colour while the left side has suffered photo degradation and soiling. Not only does the radiograph aid in the mapping of the remains of stitching threads,

(b)

Figure 9.26 Radiograph, (a) pairs of pin holes on either side of a dart, (b) lines of puncture holes left after the removal of machine stitching. (© Sonia O’Connor, University of Bradford.)

146

Exploring the X-radiographic features of textile objects

(a)

(b)

Figure 9.27 Chinese painted silk furnishing fragment, (a) photograph, (b) radiograph. (Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

Figure 9.28 Radiograph, woman’s Edwardian cape collar and front facing. (Private collection; © Sonia O’Connor, University of Bradford.)

Threads, fabrics and construction techniques 147

stitch holes and tack holes but two types of possible adhesive residues can also be identified (Figure 9.27b). Fabric linings and support structures Radiography can be particularly useful in identifying and understanding important but hidden interlinings and support structures. For example, the plain weave interfacing in the collar of a black wool Edwardian woman’s lace-trimmed cape and another interfacing along the front opening are revealed (Figure 9.28). The fine ribbed silk cover of this early nineteenth century kid boot (Figures 9.29a and b) does not show

in the radiograph. Instead the plain weave visible throughout is the lining. The extent of the toe puff and associated zigzag stitching is also made visible. This 1863–1869 Swiss bodice (Figures 9.30a to d) is a good exemplar of the amount of information which can be extracted through radiography. The green silk panels are mounted on a sturdy twill fabric and trimmed with hand-made lace (Figure 9.30a). The stays are stitched into casings on all the major vertical seams. All these elements, and many more details, were recorded by the radiograph (Figures 9.30b and c). For instance, the information gained in just a small area of the junction between

(a)

(b)

Figure 9.29 Early nineteenth century silk covered kid boot, (a) photograph, (b) radiograph; detail of toe puff. (TCC 2840; © Sonia O’Connor, University of Bradford; reproduced by permission of Sherborne Museum.)

148

Exploring the X-radiographic features of textile objects

the bodice and the tabs (Figure 9.30c and d) included: ● the cord in the piping which has the image characteristics of a three strand, cabled cord ● a variety of sewing stitches, including the point where a thread was fastened off by repeated oversewing and another one secured with a knot

(b)

(a)

(c)

the graduation of the turned edge within the waist seam ● the end of the lace which is cut and not turned within the seam. These details could not otherwise be determined because this seam is covered with a tape. Taken together, this information enhances understanding of nineteenth century dress construction techniques. ●

(d)

Figure 9.30 Swiss bodice, (a) photograph, (b) radiograph of proper left front, (c) photograph of lace and piping, (d) radiograph of same detail. (YORCM: TC694; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

Threads, fabrics and construction techniques 149

Notes 1. 2.

3.

Personal communication, Penelope Walton Rogers, The Anglo-Saxon Laboratory, 19 June 2006 and 23 June 2006. This hat from Cupar, Fife, is part of the Deliberately Concealed Garments Project and now forms part of the Karen Finch Reference Collection, Textile Conservation Centre, University of Southampton; see http://search.concealedgarments.org/results.jsp?view =detail&pos=1&id=1497 For an animation showing how these stitches are constructed see http://home.howstuffworks. com/ sewing-machine.htm

References Brooks, M. M. and O’Connor, S. A. (2005). New insights into textiles. The potential of X-radiography as an investigative technique. In Scientific Analysis of Ancient &

Historic Textiles, Informing Preservation, Display and Interpretation. Post-prints of the AHRB Research Centre for Textile Conservation & Textile Studies, 13–15 July 2004 (R. Janaway and P.Wyeth, eds), pp. 168–176, Archetype Press. Cousin, F., Desrosiers, S., Geirnaert, D. and Pellegrin, N. (2000). Lisières et bordures. Journées d’études de L’Association Française pour l’Etude du Textile. Editions Les Gorgones, Tramess et Savoirs. De Dillmont, T., (n.d.). Encyclopaedia of Needlework. DMC Library. Granger-Taylor, H. and Quirke, S. (2003). Textile production and clothing. Digital Egypt for Universities. http://www.digitalegypt.ucl.ac.uk/textil/tools.html (accessed 19 June 2006). Head, C. (2004). Old Sewing Machines. Shire Publications. Mottern, R. W., London, J. R. and Morris, R. A. (1980). Radiographic examination of the Shroud of Turin – a preliminary report. Materials Evaluation, 38(12), 39–44. Van de Wetering, E. (2000). Rembrandt. The Painter at Work (1st paperback edition). University of Amsterdam/ Donside.

10 Surface decoration

The range of painted, printed, embroidered and appliqué decoration which may be used to embellish textile is vast, in terms of both materials and techniques. Many of the case studies in Part 3 discuss such embellishment. This part seeks to demonstrate how radiography can enable additional insights and provide comparative material.

Painted and printed textiles As different pigments and dyes attenuate the radiographic beam to different extents, radiography is a helpful tool for distinguishing painted and printed textiles where relatively radio-opaque materials have been used. Some printing techniques may use pigments similar to those used on painted textiles but with different methods of application. Broadly speaking, the image on a printed textile is created by controlled contact between the colouring agent on a block or a roller that transfers the colourant to the surface of the textile. Unless there is excessive colourant on the printing surface, the transfer of colour generally only sits on, or is absorbed by, the ‘high’ points of the woven structure. In painting, the colourant is applied by spreading, using either brushstrokes or stippling through a stencil. This pushes the colourant into the interstices of the weave as well as covering the ‘high’ points. Even when a fabric is sized, the surface topography persists so paint will still collect more deeply between the threads than upon them. These differences in the distribution of the colourant may be evident on the radiographs, allowing different techniques to be distinguished. Some techniques, such as indigo pencilling, may result in more ambiguous images. Radiography is universally used for studying paintings on canvas (Hassell, 2005: 112). It is potentially just as useful in the study and recording of painted textiles. In the example of an embroidered and painted image of the Madonna (Chapter 3, Figure 3.1a and b, p. 24; Chapter 17), the flesh tones are very similar to 150

the colour of the underlying silk and the original delicacy of the features is obscured by grime. The radiograph delineates the painted flesh tones, including the distribution of specific paint layers, much more clearly than the photograph. Figure 10.1a is a photograph of a painted flower on a Chinese silk upholstery fabric. The boundaries of the flower petals are visible on the reverse as diffuse pink lines (Figure 10.1b). In the radiograph, different pigments have attenuated the beam to different extents (Figure 10.1c). A magnified detail of the radiograph shows that the paint pigment sits largely in the interstices of the woven textile, producing an almost honeycomb effect which contrasts with the unpainted background (Figure 10.1d). Interestingly, the edges of the petals within the flower shown in Figure 10.1c are less radio-absorbent than the petals themselves and so appear on the radiograph as darker lines suggesting that a resist technique was used to define the edges of the petals. If the outlines of the petals were added on top of the purple, these lines would appear lighter than the background as they would cause additional beam attenuation. The evidence of the radiograph suggests that the opposite is the case and that the purple paint is absent under the outlining. The differences between a painted and a printed image can be seen by comparing Figure 10.1d with Figure 10.2c. The red patterned outer fabric of the ‘Patent Purified Russian Down Skirt’ produces a radiographic image (see Figure 8.8, p. 114). The pattern of stylised flowers and leaves appears to have been created by discharging the red ground to leave white areas, some of which were then overprinted with blue and cream. Only the cream dye is visible in the radiograph; the white ‘daisy’ centres and the blue dye are not distinguishable from the red ground. A similar effect can be seen on the radiograph of both the red ground fabrics forming the outer covers of Mary Burnett’s quilt (see Figure 22.3, p. 279). Evidence of printing is also revealed in the radiograph of a stylised red and green ‘seed head’ motif on a silk fabric in a 1718 English patchwork coverlet (see Figure 6 in

Surface decoration 151

(a)

(b)

(c)

(d)

Figure 10.1 Flower detail from a Chinese painted silk, (a) photograph, obverse, (b) photograph, reverse, (c) radiograph, (d) magnified radiograph. (Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

152

(a)

(c)

Exploring the X-radiographic features of textile objects

(b)

Figure 10.2 McLintocks ‘Patent Purified Russian Down Skirt’ 1870–1880s, (a) photograph, printed design, (b) radiograph, printed design, (c) radiograph, detail of printed design. (YORCM: BA5713; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

Surface decoration 153

(a)

(b)

Figure 10.3 Two wrapped ‘gold’ threads and a wrapped silver thread, twentieth century, (a) photograph, (b) radiograph. (Private collection; © Sonia O’Connor, University of Bradford.)

Brooks and O’Connor (2005: 174) and Figure 5vi in O’Connor (2002/3: 25)). The green does not show up but the very radio-opaque red has produced a clear image which shows that a printing technique was used.

Underdrawing Depending on the medium used, underdrawing in embroidery and other decorative techniques may be revealed by radiography where it is relatively radioabsorbent in comparison to other materials used either in the original artefact or in later mounting and conservation methods.

Appliqué and embroidery Radiography can supplement photography as an aid to understanding and interpreting appliqué and embroidery techniques by showing stitching which is hidden by top fabrics and/or linings, threads moving from one motif to another and the starting and

finishing points. Elizabeth Watson’s 1819 quilt has both appliqué and embroidery. The radiography revealed otherwise invisible information about seams and turnings and enhanced understanding of the stitching (see Figure 22.9, p. 284). In the case of the embroidered chalice veil, different stitch types appear quite distinctly on the radiograph (see Figures 17.3a and b, p. 228). However, the radiographic results can be very unexpected and look unlike the actual stitches because the threads, whether above or below the fabric, are rendered identically and are superimposed on each other. In other instances, one element of a complex feature may not show up at all. In the radiographic image of the cartouche surrounding the couple in the pastoral scene (Figure 19.3, p. 241), the underlying element predominates. There may also be variations within areas of the same stitch. Of the different stitches in the chalice veil, variations in the image density of the long and short stitching on the radiograph relate to the different colours used and may largely be due to differential absorption by the dyes and mordants. In several of the

154

Exploring the X-radiographic features of textile objects

seventeenth century embroideries examined in Chapter 19, the most prominent image in the radiograph was produced by black threads. Black threads were also especially dominant in the radiograph of The Proclamation of Solomon (Figure 19.4, p. 242). Tracing the location and number of threads moving from one motif to another can help to indicate whether the priority was economical use of precious threads or creating the piece as quickly as possible. With sufficient comparative data, it may be possible to hazard a guess as to whether the maker was a professional or a skilled amateur. Knots within the embroidered areas often appear as diffuse light ‘dots’; these can be seen in the radiograph of the embroidered flowers of the chalice veil (Figure 17.3, p. 228). Knots can be distinguished from particulate matter such as soiling caught in weaves or linings as this tends to be more radio-absorbent, producing very bright sharp-edged and often angular images. Mapping the occurrence and location of such knots made visible through radiography can allow thread length and direction of working to be established.

Metal Threads Radiography also reveals many of the supports used in raised work embroidery such as metal wires and pins and wooden moulds; see Figure 19.2, p. 240. Metal threads used in both woven textiles and embroidery have been the subject of extensive study (Bacchus, 1999), often using energy dispersive X-ray fluorescence spectroscopy; for example, Darrah (1987), Hacke et al. (2005), Járo (1984) and Járo et al. (2000). When carried out using appropriate techniques (see Chapter 3, p. 46), radiography can aid documentation and interpretation as well as providing information about condition. Figure 10.3a shows two different ‘gold’ threads and a silver thread. Note the difference in the image density between the radiographs of these wrapped threads (Figure 10.3b). Threads of different metals, different thread types and constructions, joins within them and the embroidery techniques used can be mapped; see Chapter 18. Although the structure of these threads can be imaged using a normal X-ray unit, this radiograph was taken with a micro-focus X-ray unit which allows a larger-scale image to be reproduced (Figure 10.4). Caution is necessary when interpreting such images. Although the image appears to show that the ‘Jap gold’ wrapped thread is broken, it is only the metal filament which is discontinuous. The fibre core is actually continuous but does not show up at

Figure 10.4 Micro-focus radiograph of ‘Jap gold’ wrapped thread taken at X-Tek Systems Ltd. (Private collection; © Sonia O’Connor, University of Bradford.)

the beam energies used to image the metal filament. The central thread appears to have a solid core but this is actually produced by the overlapping of two separate wrapped threads. The following examples show more of the potential of this technique for metal threads and embroidery. Radiography has established the presence and structure of unsuspected, partially concealed and tarnished metal threads in a modern ribbon (Figure 10.5). In Figure 10.6, coiled metal wire (purl) and flat coiled metal strip wrapped in silk thread can clearly be distinguished around the date embroidered in seed pearls in a seventeenth century embroidery showing The Sacrifice of Isaac (Figure 10.6 and see Chapter 19, pp. 238–240; Brooks, 2004: 36–39). Complex metal threads constructed from both strips of metal and metal wire in a seventeenth century embroidery showing The Proclamation of Solomon are revealed by radiography (Figures 10.7a and b; Brooks, 2004: 48–49). Many different types of

Surface decoration 155

(a)

(b)

Figure 10.6 Radiograph of metal threads in The Sacrifice of Isaac. (WA OA.414; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation by Sonia O’Connor, University of Bradford.)

Figure 10.5 Modern ribbon with partially concealed wrapped metal threads, (a) photograph, (b) radiograph. (Private collection; © Sonia O’Connor, University of Bradford.)

(a)

(b)

Figure 10.7 Metal threads in The Proclamation of Solomon, (a) photograph, detail of rock pool, (b) radiograph, detail of rock. (WA 1947.191.313; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

156

Exploring the X-radiographic features of textile objects

(a)

(b)

Figure 10.8 Detail of Indian hanging, twentieth century, (a) photograph, (b) radiograph. (Private collection; © Sonia O’Connor, University of Bradford.)

metal threads of remarkably regular construction and tension can also be distinguished in the radiographs of the miniature single-sided chasuble which was probably a commercial sample (see Figure 3.15, p. 47). The weft in this Indian hanging (Figures 10.8a and b) looks like a traditional wrapped thread on a central core. However, the radiograph shows that this metal filament has so little absorbency that the underlying plastic mesh can be seen, not only between individual threads but also through them, when imaged at only 15 kV. The metal wrapping may be little more than a mono-molecular layer on a thin plastic strip. In contrast, the metal wrapped thread used for the fringe has the sort of absorbency normally associated with traditional metal wrapped threads. It is worth noting that radiographing pieces which have overall metal thread work may not always be useful as their image masks that of the less radioopaque structures and substructures completely.

Unusual materials used for surface decoration The many unusual materials used for surface decoration have individual radiographic characteristics. Even when these materials can be viewed visually, using radiographic images can enhance understanding by aiding identification as well as enabling mapping and recording. Radiography can help establish the nature of decorative ‘stones’ such as those which appear in the rocks in Figures 10.9a and b in the embroidery Charity (Brooks, 2004: 62–63). Diamonds are very radiolucent because they are entirely carbon (atomic number 12) and so can be readily distinguished from faux diamonds and paste which appear relatively radio-opaque. Other gemstones will absorb X-rays to a greater or lesser extent depending on their major constituents and the trace elements which influence their colouration. Glass is also used in place of gemstones and again additions including colourants will

Surface decoration 157

(a)

(b)

influence their radio-opacity. Lead glass will be very radio-opaque. Sequins or spangles were originally made from coiled wire cut down one side to produce individual links which were then beaten to form disc-like sequins. Coiled and stretched wire, beaten to make a decorative edging, and individual sequins can both be seen in radiographs of the hanging pockets; see Figure 18.6b, p. 236. These sequins are usually not perfectly circular and have an indent on one edge where the butt join runs through to the stitching

Figure 10.9 Detail of embroidered pool with fish and rocks in Charity, (a) photograph, (b) radiograph. (WA 1975.13; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

hole. This is generally off-centre. The spectacular raised work fish in Charity is covered with such sequins (Figures 10.9a and b; Brooks, 2004: 62–63). The radiograph makes it easier to see that these have larger stitching holes and where the sequins are vulnerable to loss as the joins have become worn. Note also the overlapping joins in the metal filaments of the wrapped metal threads in the top row of the waves and the fish’s tail. This radiograph of an Edwardian bodice was taken at 15 kV and shows a detail of the lace, beads and

158

Exploring the X-radiographic features of textile objects

(a)

(b)

Figure 10.10 Detail of bodice decorated with lace, beads and sequins, (a) photograph, (b) radiograph. (Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

sequins together with baleen from an underlying stay (Figures 10.10a and b). The pink, green and gold coloured sequins were stamped from a copper alloy sheet and coated with either a clear or coloured lacquer. Some have developed a cracquelure and, in some places, exhibit both green and white corrosion products. The sequins are attenuating the beam as a result of their metal substrate. Sequins have been made from many materials including fish scales. Gelatine sequins were made from the early nineteenth century onwards. As gelatine alone would not attenuate X-rays to any extent, the density of an image of such sequins would depend on their metallic content. Later sequins were made of plastics with different colourants and metallic content including aluminium, so again their radiographic image would depend on their metal content (McCormack, 1995). When radiographed at 15 kV, the faceted pearlised semi-translucent sequins on the 1950s test doll hardly produce any image at all in comparison to the glass beads (Figures 10.11a and b). In contrast to the glass, these sequins have no metal content.

Fish scales were also used as decoration by Victorian embroiderers. In their 1882 Dictionary of Needlework, Caulfeild and Saward (1989: 207–208) give detailed instructions on how to prepare scales, including using needles to make holes and tinting them with Damar varnish. Alongside the silk and ribbon embroidery, this velvet reticule is ornamented with fish scales (Figure 10.12a). It may seem as if these scales have been artificially cut and grooved but this is actually the natural lobing along the anterior edge (Figure 10.12b; Maitland, 1972: 75). This characteristic enabled the scales to be identified as perch (Perca fluviatilis), a native British fish.1 The scales are a keratin-like protein impregnated with calciumbased salts so their ability to transmit X-rays is not dissimilar to baleen. This radiograph, taken at 15 kV to investigate the textile components of the reticule, also produced good images of the fish scales, even though there were three to four overlapping scales in places. The concealed stitching holes in each layer of the scales forming this elaborate flower were revealed (Figure 10.12c).

Surface decoration 159

(a)

(b)

Figure 10.11 Sequins on a 1950s doll, (a) photograph, (b) radiograph. (Private collection; © Sonia O’Connor, University of Bradford.)

Pearls are concretions of calcium carbonate in an organic matrix produced by molluscs with a pearly iridescent surface called nacre. The visual characteristics of various natural and artificial forms of pearl are described by Pedersen (2004). Gemmologists use radiographs to view the internal structure of a pearl to identify whether it is, for instance, natural, cultured (bead-nucleated), mantle-tissue-nucleated or keshi, a by-product of the freshwater culturing process (Roskin, 1998). The radiographic image of natural pearls usually shows concentric rings of varying density whereas cultured pearls have a definite interface between the nacre layer and the nucleus (Matlins, 1996; Lu and Shigley, 2000). The embroidered date 167(3) on The Sacrifice of Isaac uses natural pearls (Brooks, 2004: 36–39). The radiograph highlights the fact that some of the pearls are loose and some are misplaced. Seeing the direction of the drilled holes helps in reading this date (Figure 10.6). Imitation pearls, such as wax-filled glass, solid glass, plastic or mother-of-pearl beads, have distinctly different radiographic images (Figures 10.13a and b). If it is necessary to identify the specific type of pearl present on a textile, it would be preferable to use low-energy

micro-focus radiography because this allows a higher image resolution to be obtained. See Chapter 5 (pp. 83–5) for a discussion of the interpretation of a mixed-media raised work embroidery depicting The offering of Abigail which incorporates many of the materials and techniques reviewed here.

Note 1.

Personal communication, Andrew Jones, University of Bradford, 5 October 2004.

References Bacchus, H. (1999). Developing Guidelines for the Documentation, Characterisation and Analysis of Metal Threads. (Unpublished Diploma report, Textile Conservation Centre.) Brooks, M. M. (2004). English Embroideries of the Sixteenth and Seventeenth Centuries in the Collection of the Ashmolean Museum. Ashmolean Museum/Jonathan Horne. Brooks, M. M. and O’Connor, S. A. (2005). New insights into textiles. The potential of X-radiography as an

160

Exploring the X-radiographic features of textile objects Figure 10.12 Velvet reticule with fish scale embroidery, nineteenth century, (a) photograph, (b) photograph, detail, (c) radiograph, detail. (YORCM: BP625 © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

(a)

(b)

(c)

Surface decoration 161

(a)

(b)

Figure 10.13 Man-made pearls, from left to right: glass with pearly coating; coated wax-filled glass spheres; a Majorcan ‘pearl’; bead-nucleated pearls; semi-transparent pearlescent glass beads, (a) photograph, (b) radiograph. (Private collections; © Sonia O’Connor, University of Bradford.)

investigative technique. In Scientific Analysis of Ancient & Historic Textiles, Informing Preservation, Display and Interpretation. Post-prints of the AHRB Research Centre for Textile Conservation & Textile Studies, 13–15 July 2004 (R. Janaway and P.Wyeth, eds), pp. 168–76, Archetype Press. Caulfeild, S. and Saward, B. (1989). Dictionary of Needlework (reprint of 2nd ed. 1882). Blaketon Hall.

Darrah, J. A. (1987). Metal threads and filaments. In Recent Advances in the Conservation and Analysis of Artifacts (J. Black, ed.), pp. 211–221, Summer School Press. Hacke, A-M., Carr, M. C. and Brown, A. (2005). Characterisation of metal threads in Renaissance tapestries. In Scientific Analysis of Ancient and Historic Textiles (R. Janaway and P. Wyeth, eds), pp. 71–78, Archetype Publications.

162

Exploring the X-radiographic features of textile objects

Hassell, C. (2005). Paintings. In Radiography of Cultural Materials (2nd ed.) (J. Lang and A. Middleton, eds), pp. 112–129, Elsevier. Járo, M. (1984). The technological and analytical examination of metal threads on old textiles. In Fourth International Restorer Seminar, Veszprém, Hungary, 2–10 July 1983 (A. Balázsy, ed.), pp. 253–264, Központi Muzeumi Igazgatóság. Járo, M., Gál, T. and Tóth, A. (2000). The characterization and deterioration of modern metallic threads. Studies in Conservation, 45(2), 95–105. Lu, T. and Shigley, J. E. (2000). Nondestructive Testing for Identifying Natural, Synthetic, Treated, and Imitation Gem Materials. American Society for Nondestructive Testing. Available on http://www.asnt.org/publications/ materialseval/solution/oct00solution/oct00sol.htm (accessed 1 August 2006).

Maitland, P. S. (1972). A Key to the Freshwater Fishes of the British Isles. Scientific Publication no. 27. Freshwater Biological Association. Matlins, A. L. (1996). The Pearl Book. GemStone Press. McCormack, D. (1995). Spangle is a Synonym for Sequin. Available on http://www.thefanzine.com/sections. php?scolumns&id19&aarticles (accessed 1 August 2006). O’Connor, D. (2000/3). The dress show: a study of the fabrics in the 1718 silk patchwork coverlet. Quilt Studies, 4/5, 74, 91, 93. Pedersen, M. C. (2004). Gem and Ornamental Materials of Organic Origin. Elsevier. Roskin, G. (1998). AGTS’s new lab. Gem identification goes high-tech. Jewelers Circular Keystone, October. Available on http://www.jckgroup.com/article/CA6249742. html?stt001 (accessed 1 August 2006).

11 Makers and making, degradation and repair

This chapter explores how radiography may provide evidence about makers and making, degradation, including that resulting from use, wear or storage, and the results of reuse, alterations and modifications which could be categorised as repair or conservation.

Makers and making Radiography reveals not only unintentionally obscured details but those which were not intended to be seen. It may provide unprecedented insights into the methods – and perhaps the mindset – of the maker. Was economy, speed, precision or a display of ‘virtue’ the maker’s most important goal? The driving force for some makers might be the proverb ‘What the eye doesn’t see, the heart can’t grieve over’ while for others the quality of their stitching might be influenced by the belief that their work would be judged by an all-seeing god. Radiography exposes both approaches. Characteristic stitching patterns can be identified so establishing where more than one person’s handiwork is present. Hand stitched and machine sewn seams can be distinguished even when a seam is not externally visible so clarifying manufacturing methods and, possibly, suggesting a date range. For example, the regularity of the threads moving between design elements on a miniature single-sided chasuble sample confirmed this to be machine embroidery (Figure 3.15, p. 47). Quilts are an interesting case (see Chapter 22). They may be made by a single maker with careful attention to detail, not to say devotion, for sale or for specific occasions such as a marriage. Some quilts are signed with a single name, as in the case of Elizabeth Watson’s quilt (see Figure 22.9, p. 284). This does not necessarily mean that all the work was carried out by one person and a full radiographic survey would be an excellent tool for identifying and mapping different makers’ hands. Figure 5.3 (see p. 79) is a detail from an early eighteenth century English patchwork coverlet (see Chapter 5, p. 78). This is signed with the initials ‘E H’ in patchwork on the front and in

embroidery on the back. In its radiographic study, the hidden tacking stitches attaching the silk pieces to the paper templates became visible. The stitching techniques in adjacent areas vary, suggesting more than one individual was involved in its construction (Brooks and O’Connor, 2005: 172–174, Figure 6). Other quilts were clearly made by a group of people, either working together on the piece or making sections separately as their contribution to the whole. Although only a section of the silk log cabin quilt was radiographed, this was sufficient to reveal differences in stitching technique and quality of work suggesting the involvement of more than one person (see Chapter 22, p. 278, Figure 22.5, p. 280). Radiographing the whole of the small cord quilted oblong allowed the change in the cording technique to be recognised and mapped. This is clearly a second maker completing work left unfinished by the first (see Chapter 22, pp. 282-3, Figure 22.11, p. 286). Sometimes radiography raises questions which are still be to be answered. Most of the baleen stays revealed during this research project had flat or shaped ends, with or without stitch holes for attachment, and were consistent in having straight sides (Figure 11.1a). One exception was the baleen stay in the Swiss bodice discussed in Chapter 9 (pp. 147–148, Figure 9.30). The radiograph shows that this had been trimmed straight across the top and bottom but its sides had been very unevenly cut. Comparing Figure 11.1a with Figure 11.1b shows the variable width of this stay which is unusually narrow in relation to the width of its casing. Is this an example of poor quality manufacture? This seems to contrast oddly with the quality and attention to detail shown elsewhere in the construction of the bodice. Could this be evidence of a professional or a home dressmaker trimming a larger stay which was previously used elsewhere? However, such evidence can be ambiguous and it is important to evaluate data and deductions drawn from radiographs in the context of evidence from the artefact and its context. 163

164

Exploring the X-radiographic features of textile objects

Degradation Radiography can help gauge the condition of substructures in upholstered furniture, whether this is deterioration due to corrosion, rot or woodworm. This piece of furniture (Figure 11.2) was presented to a company for X-ray because the owner

(a)

(b)

Figure 11.1 Radiographs, (a) trimmed baleen stay, (b) unevenly trimmed baleen stay from Swiss bodice. (YORCM: TC694; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

(a)

had heard the chair squeaking. Two radiographs were taken with a short time delay between them. The images showed not only the presence of a tropical woodworm but that it was still active. Between taking the two images, it had moved position and the shape of the cavity had been altered (Figures 11.2a and b).1 Radiography also showed the trails of insect damage in the cardboard brim of the Cupar hat (Figure 11.3 and Chapter 9, Note 2, p. 149, Figure 9.20, p. 141). The larger oval is its flight hole. This damage is completely hidden by the fabric’s pile. As ever, it is important to compare evidence from the radiograph with evidence from the artefact. The brightness on this radiograph would be hard to explain except in terms of the presence of a much denser material (Figures 11.4a and b). In fact, this archaeological silk has been burnt along the edge and the damaged area of once molten silk has produced a very dense residue. Radiography makes it easier to track the passage of threads which are now lost. Bleached or dark coloured threads are often the first to be lost due to their increased vulnerability resulting from manufacturing processes. The stitches used to outline the painted features on the Madonna’s face on the support fabric of the chalice veil are lost but the stitch holes

(b)

Figure 11.2a and b Chair, radiograph showing the movement of a woodworm. (Private collection; © Peter Baucham, Argos Inspection, Washington, Tyne & Wear.)

Makers and making, degradation and repair 165

Figure 11.3 Insect damage in card inside Cupar hat brim, radiograph (Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

(a)

(b)

Figure 11.4 Archaeological silk with burnt edge, (a) photograph, (b) radiograph. (YAT 1989.28 ctx 3187 sf 470; © Sonia O’Connor, University of Bradford; reproduced by permission of York Archaeological Trust.)

166

Exploring the X-radiographic features of textile objects

remain (see Figures 3.1a and b, p. 24 and Chapter 17). In the case of the original embroidery, the very whitest silk threads are mostly missing from the flowers (Figure 17.2b, p. 227). Radiography enabled their erstwhile presence to be mapped, including stitch holes obscured by other stitches. This seventeenth century embroidered book cover (Figure 11.5a) includes a depiction of The Conversion of Saul. The laidwork metal thread scroll above the fallen Saul contains very fragmentary remains of black silk thread which probably once read ‘Saul, Saul, why persecutest thou me?’. Unfortunately, the radiograph could not give any indication of any remaining threads which might be concealed below the metal threads due to the high absorption of the X-ray beam by the latter (Figure 11.5b). Nevertheless, the needle punctures through the closely laid metal wrapped threads are visible, substantiating the presence of lost stitching but not enabling any text to be read. The stitched areas contrast with those areas of the scroll which did not carry text where the metal threads show hardly any damage or disturbance. This radiograph of a nineteenth century silk necktie maps the complex damage present which is probably due to a range of manufacturing processes including weighting (Figure 11.6). It also provides a vivid demonstration of how folding during storage exacerbates damage in an already weakened fabric.

Use and wear Radiography can help map the condition of specific threads. Abrasion to the surfaces of areas of raised metal thread work in elaborately embroidered hanging wall pockets could be easily mapped using radiography and the pattern of loss in the sequins charted (see Figure 18.2, p. 233). Detached fragments of metal thread can also be located and identified; see Conti and Aldrovandi’s discussion of the pieces of metal thread caught in the turnings of a chasuble (see Chapter 13, p. 190). It proved a useful tool for mapping the extent and distribution of damage in the stocking (see Figure 16.4a, p. 220) and could also be used for charting friction damage to the surface of textiles such as pilling. Stallybrass notes that, ‘In the language of nineteenth century clothes-makers and repairers, the wrinkles in the elbows of a jacket or sleeve were called “memories”. These wrinkles recorded the body that had inhabited the garment. They memorialised the interaction, the mutual constitution, of person

and thing’ (1998: 196). Radiography can highlight and map such ‘memories’. The interaction between human wearer and bodice is dramatically memorialised in this radiograph (Figure 11.7a). The stress creases associated with the hooks are visible in both the photograph and the radiograph. However, the radiograph shows that the baleen stay behind the hooks on the front fastening has broken into three fragments under the stress of bending during wear (Figure 11.7b). The first break is by the fifth hook from the bottom and the second break is below the ninth hook. The detached top fragment with its unused stitching hole has slid down and dropped behind the middle fragment. This has risen up slightly producing a gap at about the level of the fifth hook. It might have been possible to deduce some of this information from careful handling and manipulation but it is highly unlikely that the location of the upper fragment would have been discovered. Details such as the presence of the stitching hole would also not have been deduced even by the use of bright transmitted light because of the number of layers in the front fastening. Although photography is ideal for recording wear on the base of the soles of shoes, it is often more difficult to document the condition of the insides of soles where they would have been in contact with the wearer’s feet. Furthermore, although photography of archaeological shoe soles may make compression quite obvious, wear may be obscured by surface deposits and general degradation of the surface resulting from burial. Pin marks from the shoe’s attachment to the last during manufacture may show up clearly but radiographic information can sometimes be hindered by the presence of leather, wood or metal material in the heels or as stiffeners in soles. However, radiography has the advantage of recording features on both sides of the soles simultaneously while also giving a more accurate image of thinning. This is effectively illustrated by the radiograph of this well-worn leather sole from a late fourteenth century shoe excavated from waterlogged deposits (left, Figure 11.8).2 The leather had darkened so it was very difficult to determine the extent and distribution of wear and the stitch holes were obscured. Radiography showed the thinning of the sole produced by loss of material and the extent of wear around the hole produced by the big toe and the ball of the foot. Damage at the heel is also dramatically evident. Detail of the stitch holes and the holes for the tunnel stitching became visible. Radiography is also a useful tool for imaging the condition of inner soles in boots and shoes with a high vamp as these

Makers and making, degradation and repair 167 Figure 11.5 Detail from embroidered book cover The Conversion of Saul, (a) photograph, (b) metal thread scroll, radiograph detail. (WA 1947.191.318; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

(a)

(b)

168

Exploring the X-radiographic features of textile objects

may be impossible to see or photograph in their entirety. The radiograph of the sole in the silk and kid boot (see figure 9.29, p. 147) shows slight traces of wear at the toe (middle, Figure 11.8). The shoe illustrating the nursery rhyme ‘There was an old woman who lived in a shoe’ is filled with dolls so it is not possible to explore the inside (see Figure 1.1, p. 4). However, this radiograph appears to show that the toe contains some packing material (right, Figure 11.8).

Reuse, repair and conservation

Figure 11.6 Nineteenth century silk tie, radiograph. (Private collection; © Sonia O’Connor, University of Bradford.)

(a)

Objects that have had a long life may well show evidence of the work of more than one hand, not only in the initial making but in subsequent reuse, alterations and repair. Mary Burnett’s quilt is an amazing example of a textile which has been modified and remade by subsequent generations, giving it an extended working life (see Chapter 22 and Figure 22.3, p. 279). The case study of an early stomacher is an excellent example of how radiography can help understand the successive episodes of repair and reuse as it moved down the social scale (see Chapter 14). This c. 1870–1876 silk bodice has had a second life as fancy dress and has had numerous alterations and repairs to these alterations (Figure 11.9a).

(b)

Figure 11.7 Detail of bodice front fastening, (a) photograph, (b) radiograph, detail. (Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

Makers and making, degradation and repair 169

Figure 11.8 Radiograph showing three shoe soles, from left to right: fourteenth century sole from Monk Street, Tutbury (© Sonia O’Connor, University of Bradford; reproduced by permission of Birmingham Archaeology); early nineteenth century silk covered boot (TCC 2840; © Sonia O’Connor, University of Bradford; reproduced by permission of Sherborne Museum); ‘The Old Woman in the Shoe’ (YORCM: AA2576; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

Radiography revealed the extent of the damage hidden by the patch. As well as repair stitching, losses where the original lining was torn when the hook was lost can be seen (Figures 11.9b and c). It is important to remember that radiographic evidence needs to be contextualised for an object to be understood as fully as possible. Radiography established that the seams concealed in the new facing and backing of the chalice veil were distinct not only from the original seams but also from each other (see Figures 17.3a and b, p. 228). Those in these later fabrics were clearly worked by different hands although other data from different aspects of the artefact will be needed to establish the sequence of remounting and repair and which stages, if any, are contemporaneous. Kite discusses the use of radiographs to locate previous repairs such as wires holding on a teddy bear’s head (see Chapter 21,

p. 269 and Figure 21.5, p. 270). Radiography also provided data about unusual stuffings inside teddy bears (see Figures 21.2, p. 267 and 8.13, p. 120). In such cases, it would be worth reviewing the condition of the seams to see if these had been opened and resewn in order to establish whether this is the original stuffing or a later addition. What appears to be an oblong repair area to the metal thread embroidery becomes visible on the left thigh of Charity on the hanging pockets (see Figure 18.5, p. 136). It may be possible to image specific threads used for repairs. The repair thread in ‘The Old Woman in the Shoe’ was dramatically visible in the radiograph, suggesting this is a weighted thread (see Chapter 1, pp. 3–4 and Figure 4.9, p. 72). The eighteenth century knitted stocking discussed in Chapter 16 had been cut down and may also have been used for fancy dress.

170

Exploring the X-radiographic features of textile objects Figure 11.9 Silk bodice, detail of front showing wear and repair, (a) photograph, obverse, (b) photograph detail of patch, reverse, (c) radiograph, detail. (Private collection; © Sonia O’Connor, University of Bradford.)

(a)

(b)

Radiography shows the details of the darning technique used to secure runs and holes very clearly (Figure 11.10). This darn was partially successful in securing the stocking stitch but two separate knitted stitches have continued to run. The fine overstitched repairs in the eighteenth century cord quilted oblong were easier to document by radiography (see Figure 22.11, p. 286). Radiography may also be useful to establish treatments which have been undertaken while an artefact is in the care of a museum collection.

(c)

In this sample (Figure 11.11), the radiograph shows that a stitched support intended to stabilise damage has had an adverse impact upon the weave structure being conserved due to excessive tension in the stitching. Canadian Conservation Institute scientists have used radiography, in conjunction with other techniques, to identify and track the location of pesticides with heavy metal residues and arsenic on artefacts containing fur, feather and leather as well as textiles (Sirois and Sansoucy, 2001).

Makers and making, degradation and repair 171

Figure 11.11 Radiograph of tear with stitched silk net support. (© Sonia O’Connor, University of Bradford.)

Summary

Figure 11.10 Darn in eighteenth century knitted stocking, radiograph (YORCM: 1175–76 362/41; © Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

Figure 11.12a is one of a pair of small cloth dolls (c. 120 mm tall), of unknown provenance and date, with embroidered detail and filled with sawdust. The fabric knees of the doll had split with consequent loss of the sawdust filling. Tightly rolled strips of a similarly coloured textile were inserted into each knee to fill the gaps and to provide a solid foundation for the repair (Figure 11.12b).3 In the radiograph, Figure 11.12c, the rolled textile contrasts with the speckled image of the sawdust and provides a record of the extent and fit of each repair which could not have been gained in any other way.

The introduction showed that using radiography in the investigation of textiles is not new. Part 2 demonstrates that a systematic approach to the acquisition and interpretation of images can hugely increase the benefits of radiography in the understanding of textiles. The new evidence generated through radiography has answered many questions but – inevitably and excitingly – has provided a different perspective which inspires many new areas of research. Even radiographing one part of an artefact, carefully selected, has moved understanding forward, both in developing the techniques used and in gaining insights into the textile. It is hoped that this exploration encourages those working with textiles to regard radiography in the same way that it is regarded in fields such as the conservation and study of paintings and archaeological materials – a fundamental tool for understanding textile artefacts and for planning appropriate conservation strategies.

Notes 1. 2.

Radiography by Peter Baucham, Argos Inspection, Washington, Tyne & Wear. Excavated in 2005 by the Birmingham Archaeology Unit. Identification by Gareth Williams, British Museum.

172

Exploring the X-radiographic features of textile objects

(b)

(a)

(c)

Figure 11.12 Cloth doll with sawdust filling, (a) photograph, (b) photograph, detail of knees after conservation, (c) radiograph of knees after conservation. (Karen Finch Reference Collection, Textile Conservation Centre; © Sonia O’Connor, University of Bradford; reproduced by permission of the Textile Conservation Centre, University of Southampton.)

3.

Personal communication, Tom Bilson, Head of Digital Media, Courtauld Institute of Art.

References Brooks, M. M. and O’Connor, S. A. (2005). New insights into textiles. The potential of X-radiography as an investigative technique. In Scientific Analysis of Ancient & Historic Textiles, Informing Preservation, Display and

Interpretation. Post-prints of the AHRB Research Centre for Textile Conservation & Textile Studies, 13–15 July 2004 (R. Janaway and P. Wyeth, eds), pp. 168–176, Archetype Press. Sirois, P. J. and Sansoucy, G. (2001). Analysis of museum objects for hazardous pesticide residues: a guide to techniques. Collection Forum, 17(1–2), 49–66. Stallybrass, P. (1998). Marx’s coat. In Border Fetishisms: Material Objects in Unstable Places (P. Spyer, ed.), pp. 183–120, Routledge.

Part 3 Case studies Introduction

Mary M. Brooks and Sonia O’Connor The following case studies are intended to demonstrate the benefits of radiography in relation to a wide range of textiles and dress. The scope of the textiles is deliberately wide ranging in terms of type and time – from the pre-historical, historical to the modern. Archaeological and ethnographic textiles, sacred and secular dress, furniture, quilts, embroideries, painted canvas, dolls and teddy bears are all featured. Susanna Conti and Alfredo Aldrovandi’s extended paper is the result of extensive research applying radiography to both large and small textile artefacts over a period of years at the Opificio delle Pietre Dure, Florence. Kate Gill’s detailed exploration of the use of different types of radiographic equipment shows that effective radiographic examination of upholstered furniture can take place both within the conservation laboratory and in situ in historic houses. Other papers provide comparative data resulting from the radiography of similar groups of artefacts, such as the study of archaeological and historic shoes by Elizabeth Peacock, Sarah Howard and Robert Holmes and the review of nineteenth and early twentieth century dolls by Mary Brooks,

Sonia O’Connor and Josie Sheppard. Louise Bacon’s review of the use of radiography in relation to multi-media ethnographic artefacts including textiles sounds a significant note of warning about the possible ethical implications of revealing hidden elements which were intended to remain unseen. Shorter papers focus on the radiography of a single artefact such as Sylvie François’ investigation of an unusual musical instrument and David Starley and Fiona Cahill’s study of a mannequin. Other papers, such as Gabriella Barbierie’s study of a reused and altered stomacher, show how radiography can reveal the previous social life of artefacts. The importance of outreach to the public is not forgotten and Clare Bowyer’s paper describes how radiographs can function in a museum exhibition to interpret garments and shoes to visitors. These case studies, in which the authors generously share their knowledge and expertise, will provide inspiration for further textile radiography which will benefit both conservation practice and textile studies with the ultimate goal of enabling greater public appreciation of the world’s textile heritage.

This page intentionally left blank

12 Evaluating X-radiography as a tool for examining upholstered furniture Kathryn Gill

Introduction This chapter focuses on the role of radiography within upholstery conservation. It highlights some of the practical challenges of carrying out X-radiography of upholstered seat furniture and illustrates how information revealed by radiography has increased understanding of the configuration and range of materials concealed within upholstery structures. Two case studies involve the use of onsite radiography facilities within a museum where the furniture was transported within the same building, either from display or storage, to the radiography facility within the conservation departments. The final case study concerns the radiography of a piece of furniture displayed in a building with no radiography facilities. Rather than transporting the fragile artefact to an X-ray facility, a mobile X-ray unit was transported to the site.

Practical challenges to the radiography of historic upholstered seat furniture Radiography of historic upholstered furniture is a particularly challenging task for several reasons. The objects are sometimes large, always three dimensional, often unusually shaped and inevitably made from a wide range of materials. More often than not, they have unstable frames and fragile elements, including decorative surfaces, such as gilding and paint, and degraded textiles and collapsing understructures. Consequently, no matter what the set-up, type of equipment, level of expertise or experience, the degree of access and consequently the information revealed will be more limited in comparison to, for instance, smaller two-dimensional objects composed

of fewer materials in good condition. For example, many sections, such as across the width of a seat, are too large to be usefully X-rayed. The exposed X-ray film would be too complicated to interpret because the image of the elements furthest from the X-ray film would be magnified, distorted and insufficiently sharp to record detail, and would overlie the image of elements nearer the film. Therefore, radiography is limited to ‘thin’ areas to capture as much useful information from the layers as possible. Preferred angles tend to be from the inside to the outside face of a back unit; through an arm unit or perhaps through a corner of a seat taken at a 45° angle (Buck, 1991). In order to achieve these set-ups, the chair, the X-ray head and the film need to be manoeuvred into a particular alignment to capture the desired information: such set-ups are challenges in themselves. Unfortunately, further limitations are likely if, as is the case with many historic museum pieces, the furniture is too unstable to put in any position other than standing on all feet. Once a suitable alignment had been achieved, the next challenging task is determining the optimum exposure setting. Exposure setting has to accommodate the three dimensionality of these artefacts, the wide range of organic and inorganic materials likely to be present and the varying densities, volumes and atomic weights of individual materials and layers. Inevitably, the whole process has an element of ‘trial and error’ about it and several different exposure settings are usually necessary to achieve optimum results. In spite of this seemingly overwhelming list of obstacles, information captured on film by radiography of these challenging artefacts has proved on many occasions to have been a useful contribution to the understanding of an object as the following case studies indicate. 175

176

Case studies

Radiography for documentation: case study of an eighteenth century upholstered chair This case study, the documentation of an eighteenth century upholstered easy chair from the Brooklyn Museum of Art, New York, has been selected for two reasons (Figure 12.1a). The first is that it highlights well some of the practical challenges of undertaking radiography of upholstered furniture, in particular that of gaining access. The second is that it demonstrates the role radiography played in the examination process, including helping to understand the configuration and range of materials concealed in the upholstery layers. The primary reason for undertaking a detailed examination of the chair was to gain a better understanding of the materials and construction techniques used in its original manufacture. Details of this examination and the outcomes are discussed elsewhere in an illustrated report (Gill and Doyal, 2001). One of the main challenging aspects of this documentation project was gaining access. None of the materials could be removed to reveal the concealed inner layers; however, a number of individual layers (albeit small areas) could be observed through holes and partially detached areas. A set of basic examination and documentation tools, including camera, tape measure, magnifying lens, fine probes, spatulas and tweezers, were used to assist two upholstery conservators in the investigation and recording process. Full advantage was taken of the institution’s on-site X-ray facility and in-house objects conservator/ radiographer because of limited access to the understructure materials. Radiography was selected for its potential to shed some light on the number of tack lines, additional tack holes, frame joints and timber profiles and provided additional evidence to support and confirm what was revealed/assumed by other aspects of the examination. It was hoped that the developed X-ray films would also reveal how many layers of filling covers there were in a particular area, primarily by the location and distribution of tacks and other metal fasteners. Due to its fragile condition, the chair had to be positioned standing on all feet on top of a mobile bench throughout the X-ray procedure. Views taken included the arm/arm support from outer to inner side (Figures 12.1b and 12.1c), the front seat rail from top face to underside of seat and the back rail.1 As expected, radiography revealed construction joints in the wood frame, and the location, shape and number of concealed metal fasteners (Figure 12.1c).

Some individual textile and filling layers were also recorded. The image of the arm revealed a concealed hand-made pin, webbing, including details of its weave, a plain warp-faced weave and the faint texturing of the curled hair filling. The shape and location of light shadowing in wood frame areas adjacent to hand-forged iron tacks suggested tack holes with traces of corrosion products. Prior to examination, the materials observed were thought to be original to the frame. Radiography confirmed this view. However, the evidence revealed on the arm indicates that some of the original tacks are now missing. This suggests that the original materials may have been repaired or altered at some point after manufacture.

Radiography as a complement to photographic evidence: case study of the Seehof Suite This case study concerns a German rococo upholstered suite of side chairs, armchairs and settees with painted and gilded frames from the Metropolitan Museum of Art, New York (Figure 12.2). The outside backs of the settees were plain, while the side chairs were elaborately carved with a design of leafy tendrils poking through a trellis. This case study has been selected as it is a good example of how useful radiography was in assisting a team of conservators and curators in considering whether this particular suite of furniture should be reinterpreted for display in a museum setting following the discovery of new documentary evidence about its original appearance. The X-ray images also highlight the importance of interpreting the results next to the object and ideally with other evidence/information which suggests what the X-ray might be revealing. As was the case with many items of upholstered furniture in museum collections, although the eighteenth century painted silk panels of the suite were known not to be original to the frames, it was reasonable to assume that the frames had always been upholstered. However, new documentary evidence suggested that the upholstery covering the inner backs of the entire suite concealed elaborate polychromed and gilded carved panels not unlike those evident on the back of the side chairs. In the light of this fascinating information, deciding how the suite should be interpreted for display became a curatorial priority and was largely dependent upon the existence of the totally concealed carved panels. A conservation priority was to determine whether or not these panels were present, and if

Evaluating X-radiography as a tool for examining upholstered furniture 177 Figure 12.1 Easy chair with red stamped and watered wool top cover edged with cord bound in a red wool and silk woven tape, (a) side view, (b) detail of the proper right arm and arm support, (c) radiograph of the same section revealing hand-forged tacks, various filling cover layers, a large nail securing the arm rest to the arm support, warpfaced webbing and curled hair filling. (The Brooklyn Museum of Art, 32.38. Henry S. Batterman Fund; Carll H. DeSilver Fund; Maria L. Emmons Fund and Charles S. Smith Memorial Fund.)

(a)

(b)

(c)

178

Case studies

Figure 12.2 Painted and gilded lindenwood settee, one of a pair, German, c. 1763–1764. Shown before 1987 treatment which involved the removal of the painted blue silk top covers, under upholstery and close nailing. (The Metropolitan Museum of Art, The Lesley and Emma Sheafer Collection, Bequest of Emma A. Sheafer, 1973; 1974.356.121. Photograph by Kathryn Gill, Objects Conservation. Photograph © The Metropolitan Museum of Art.)

they were present, to determine their condition. One way was to temporarily release part of the upholstery. However, not being able to reattach the upholstery without causing damage to the fragile painted silk panels and the delicate surfaces of the surrounding frame was considered too great a risk. This approach was also considered too invasive at this stage of the investigation. Other less invasive options were therefore considered. Consequently, a decision was made to X-ray one chair and one settee (Figure 12.2) from the suite to test the hypothesis that the upholstery concealed the decorative surface illustrated in the newly revealed photographic evidence taken in the nineteenth century (Kisluk-Grosheide, 1990). In 1987 the furniture was X-rayed with an industrial unit at the museum’s in-house facility.2

Interpretation of the X-ray images The X-ray images taken of sections of the back did not reveal any upholstery fillings or filling covers concealing the back panels although some of the painting on the silk covers was recorded. However, since information on upholstery fillings was not important to this stage of the decision-making process further exposures with different times/kV/mA were not undertaken.

Most importantly, radiography of one of the settees confirmed that the carving on the inner backs was largely intact as shown in Figure 12.3a. The nineteenth century black and white photographs of current upholstery and the carved outer backs of the side chairs proved critical in helping to fully interpret the radiographs. The square grid surrounded by leaves visible in the radiograph matched the carved wood elements visible in the photographs. The grid and other main features have been highlighted in the line drawing in Figure 12.3b. The clear way in which the concealed trellis and leaves and the exposed edges of the frame showed up in the radiograph suggested the presence of lead-based pigment covering the carving and a lead component in the gesso under the gilding. The size, shape and location of the white splodges suggested that they were part of the painted silk layers covering both the inner and outer back of the settee and probably produced by a lead-based ground or pigment applied under the flower heads (Figure 12.2). Two lines of small white circles matched the single row of the dome headed nails securing to the wood frame the edges of both painted silk covers. The small ‘tadpole’-shaped white marks located within the leaf shapes looked remarkably like tacks. The most likely function of these tacks would be to secure tensioned twine over the upholstery filling to the carved wood surface. Should this be the case, although no evidence was visible on the radiograph, the tacks would have caused physical damage to the carved surface. In addition, lack of definition of carving adjacent to the gilded frame suggested that elements of the decorative surface and perhaps some of the carving were missing in these areas. The information gained from the radiograph was considerable. The process of interpreting the radiographs helped both in assessing and evaluating the pros and cons of all treatment options being considered and in anticipating the nature and extent of treatment challenges that lay ahead should a decision be made to permanently remove the later upholstery. The information not revealed on the radiographs was as influential as what they revealed. For example, it highlighted what could not be known unless the upholstery was removed (e.g. the level of intactness or overall condition of the concealed elements). It assisted in weighing up worst and best case scenarios in regard to the anticipated condition of the carving. The information also helped determine the next stages of the investigation. This involved releasing and lifting up the upholstery covering the corner of one of the four side chairs

Evaluating X-radiography as a tool for examining upholstered furniture 179

(a)

(b)

(c)

Key: Edge of settee frame

Lead-bearing pigments in painted silk covering the settee

Row of close nailing edging outer back panel

Painted and gilded carved tendrils and trellis

Row of close nailing edging inner back panel visible in Figure 5

Tacks securing under upholstery to front of carved panel

Figure 12.3 Back of painted and gilded lindenwood settee, one of a pair, German, c. 1763–1764, (a) radiograph taken before removal of the upholstery showing nailing, lead-bearing pigment and tacks, (b) line drawing highlighting features visible in the radiograph, (c) detail showing the upholstery filling in the process of removal; note areas of loss around the perimeter of the carved sections. (The Metropolitan Museum of Art, The Lesley and Emma Sheafer Collection, Bequest of Emma A. Sheafer, 1973; 1974.356.121. Photographs and line drawing by Kathryn Gill, Objects Conservation. Photograph © The Metropolitan Museum of Art.)

180

Case studies

to access the surface of the carving. Damage appeared to be restricted to the perimeter of the panel, the rest appeared to be in good condition. Based on this evidence, a decision was made to remove the upholstery panels and filling from the entire suite. Figure 12.3c shows the carved inner back of one of the settees emerging as the upholstery was removed. The areas of loss are clearly visible. Full details of this project appear in illustrated articles by one of the museum’s curators, KislukGrosheide (1990), and by the conservation team (Gill et al., 1990).

What is not revealed by radiography: case study of the Audley End settee One of the more challenging aspects of radiography of upholstery structures occurs during the interpretation of the radiographic images. The technique is used as a tool to help reveal information in otherwise inaccessible areas and often serves as an aid to looking more objectively at the upholstery and frame. As shown in the preceding case studies, information revealed on a radiographic image can be helpful and informative. However, the following case study has been selected as it reveals how much could be passed by completely unnoticed as it might not be recorded in the radiograph to start with, or be anticipated being there in the first place. In order to evaluate effectively what may not be revealed using this technique, radiography was undertaken on a piece of furniture, specially selected as all concealed upholstery elements were known beforehand. The focus of this case study is an eighteenth century settee from Audley End, one of English Heritage’s properties in Essex. A detail of the settee is illustrated in Figure 12.4a. The settee was selected because the materials and structure of every aspect of this piece of furniture, including concealed areas, were known as they had been fully documented a number of years ago by an experienced upholstery conservator during major conservation treatment of the frame. To access the frame all the upholstery structure materials had to be temporarily removed (Gill and Eastop, 1997). Consequently, during the course of conservation treatment every single element of the multilayered understructure and materials were fully revealed, thoroughly examined and clearly documented before being reconcealed between the frame structure and top covers following conservation. The documentation was extensive and included written descriptions, photographic imagery and line drawings.

The information and knowledge gained during the conservation treatment proved invaluable when interpreting the images revealed by the radiography of the settee, especially when assessing what information was there. Figure 12.4b shows the same view of the settee as Figure 12.4a captured as a radiograph.3 The X-ray was taken after conservation treatment. This particular section was selected as it was the ‘thinnest’ point along the seat rail where it was most likely that the profile of the individual filling cover materials and the remaining upholstery elements covering the front seat rail could be captured. It was angled in such a way to minimise distortion of elements.4 Figure 12.5 is a line drawing representing a crosssectional view of the front seat rail of the settee. It was prepared upon completion of conservation of the settee, several years before the radiograph was taken. The stylised line drawing was made to provide a record of all layers of upholstery fillings and filling covers and where and how they were attached to the frame after conservation. The key includes a brief description of the type of materials. For the investigation the drawing provided a useful document in helping to interpret the radiograph and the degree of information captured. To fully appreciate what has and has not been captured, a summary of the different layers follows. As expected, the radiograph has captured several materials (Figure 12.4b). Some areas are readily distinguishable, for example, the metal elements, including the (now largely shankless) dome-headed decorative nails, which manifest themselves in the radiographic image as rows of white dots. It is interesting to note that the two tiny metal rings replacing the nail shanks are clearly visible where the domed heads are captured in profile; Figure 12.4c). Equally visible are the original hand-forged eighteenth century tacks, captured at various different angles. The lead-based elements in the gesso and paint layers of the carved wood frame are also visible and serve to highlight the carved elements very well. Unfortunately with this particular exposure, neither filling (as shown in Figure 12.5, items 5 and 7) is identifiable as curled hair, although the space it occupies and the profiles of each layer are clearly visible. There is definitely a line denoting the edges of the webbing captured in the view, but no indication that there are two layers of webbing on top of each other (Figure 12.5, items 3 and E) and that they are of different weave and widths. Both wood frames are distinguishable, even though the edges are not sharp (Figure 12.5, items 0 and 2). Surprisingly, the wood grain on the replaced

Evaluating X-radiography as a tool for examining upholstered furniture 181

(a)

(c)

(b)

Figure 12.4 Right front corner of the Audley End settee, (a) three-quarter view showing wool top cover and close nailing, after conservation, (b) radiograph taken from the same viewpoint, (c) detail of radiograph showing converted dome headed nails. (Photograph reproduced by kind permission of English Heritage, Audley End House. Radiographs reproduced by kind permission of English Heritage, Audley End House and Mobile Radiographic Services Ltd.)

element (Figure 12.5, item 2) is clearly visible. Although the frame had been damaged in areas by wood boring insects, none of the bore holes were revealed (Podmaniczcy, 1990). Interestingly, the most easily distinguished layers are the two layers of nylon net (shown ‘on edge’) seen in the radiographic image as white lines (Figure 12.4b), and as

items G and H on Figure 12.5. It is unclear as to why these layers are most distinctive. Perhaps it is because the dyed net comprises fine, dense extruded white synthetic fibre compared to woven linen comprising less dense loosely spun undyed flax fibres and loose upholstery filling of moderately compacted curled animal hair.

182

Case studies

certain. Information can therefore be passed by completely unnoticed as it is neither picked up on the radiograph to start with, nor its presence suspected in the first place. Another exposure designed to image these more ephemeral components, rather than the frame, might have reduced this information gap; however, full exploration of exposure times was beyond the scope of this phase of the research project.

Investigation of a portable medical facility for object examination

Figure 12.5 Cross-section of the Audley End settee showing layers of upholstery fillings and filling covers and where and how they are attached to the frame, after conservation. Key: 9 top cover – wool/linen, stitched 8 second filling cover – linen, stitched H net – nylon, stitched 7 second filling – curled hair, bridal ties 6 first filling cover – linen stitched G net – nylon, stitched 5 first filling (roll edge) – curled hair F linen bridal ties E webbing – linen, stapled D lawn – cotton, stitched 4 base cloth – linen, stitched C base cloth – linen, stitched 3 webbing – linen/hemp B net – nylon, stitched 2 subframe (replica) – wood 1 batten – wood, nailed 0 main frame – wood A stitching ground – linen, staples/stitched I net binding – nylon, stitched 10* dome head nail (converted) – brass, stitched (Line drawing by Kathryn Gill.)

Documentation projects involving in-depth examination of upholstered furniture are often carried out on-site as the furniture is too fragile to be moved to a conservation studio (Gill, 1998; Gill, 2001). Unfortunately, the degree of information that can be acquired in such circumstances may be limited by the difficulty of transporting standard conservation tools and facilities, including X-ray machines, into the historic house environment. Even using a mobile X-ray unit parked in the grounds of the houses would require such furniture to be physically removed from the building. The radiography of the Audley End settee was part of a wider research project to evaluate the potential of a portable X-ray unit to examine fragile upholstered furniture inside historic houses. The Xray unit belongs to a company that provides a unique service intended to enable medical radiographers to X-ray patients at home. The unit is therefore designed to allow access into buildings. Their van includes equipment that allows X-ray film to be developed on-site. The investigation project aimed to evaluate the potential for upholstered furniture examination of this service. Issues explored included: ●

●

Overall, the radiographic images captured some clear information on the configuration and range of the concealed materials. The nylon net warrants further study.5 However, this particular research project provides a useful reminder that, even if the upholstery specialist examining the upholstered structure has a good idea of what is expected to be concealed, what is really there is rarely known for

●

access, specifically assessing the practicalities of setting up the equipment in a historic house, literally around the fragile three-dimensional artefact on display health and safety the quality of the image captured by a piece of medical radiographic equipment which is heavily filtered to protect human tissue.

All the original aims of the project were achieved. Setting up the equipment in a historic building full of precious items posed no unsolvable problem, the equipment was extremely versatile and developing the film on-site was very straightforward. The

Evaluating X-radiography as a tool for examining upholstered furniture 183

radiographer, conservator and curator were all very satisfied with the procedures and happy to repeat the process at another historic site. Given the limitations of the medical equipment, the level of information captured on the radiographic film was good, as was the scanning equipment. This research project shows that using this portable radiography service, in conjunction with other examination techniques, has increased the understanding of the configuration and range of the concealed materials. Having said this, the heavy filtration of the medical equipment, necessary for human safety, produces rather grey images of the textile elements, such as stitched edges or the linear formation of loose fillings. The visibility of the detail in these images has benefited from contrast adjustment after the films have been digitised. Unfiltered industrial units, used in combination with industrial film, should produce more detailed high contrast images of such components. Consequently, one of two remaining phases of this project will involve X-raying one of the three objects used for this investigation again at a facility with an unfiltered unit. The same set-up as with the medical equipment will be followed in order to provide a means of comparing the degree and quality of images from both pieces of equipment. The final remaining phase is to undertake a feasibility study on introducing a company with an unfiltered industrial unit to a company with mobile radiography services. A detailed account of this research project to date is in preparation (Gill, in press).

Acknowledgements I would like to thank the Brooklyn Museum of Art; the Metropolitan Museum of Art, New York, USA; Audley End, English Heritage and the Textile Conservation Centre (TCC), University of Southampton, UK for permission to publish this work. Thanks are also due to Mobile Radiographic Services Ltd, Teddington, UK for permission to publish two radiographic images generated from the mobile X-ray unit research project. Support provided by research funding from the AHRC Research Centre for Textile Conservation and Textile Studies is gratefully acknowledged. I would also like to thank Sonia O’Connor, University of Bradford, for sharing her knowledge and experiences of radiography and image scanning. Thanks also go to colleagues at the TCC, particularly Nell Hoare, Amber Rowe and Maria Hayward. Final thanks go to Mike Halliwell, conservation photographer, for processing the images for this publication and to Dinah Eastop for her editorial input.

Notes 1.

Conclusion Radiography can be a useful tool in the examination of upholstered furniture. No matter what technique or combination of techniques used, interpretation of information revealed during investigation of upholstered objects requires considerable knowledge of upholstery materials and techniques and input from other specialist expertise, for example practitioners and historians of gilding, tacks and nail manufacture, furniture making, textiles and passementerie. Information revealed is often best examined and interpreted by different specialists using a variety of documentary sources, including the actual objects and other well-referenced comparable examples. Without such shared collective knowledge and resources, important evidence may be missed or misinterpreted. The interpretation of radiographic images is no exception.

2.

3.

4.

This radiography was undertaken in 1991. The X-ray machine used was a Baltospot from the Balteau Electric Corp., Stamford, CN, Model # BS160, Serial # 865, volt 110, amp 115, frequency 60. Every X-ray shot on this machine was at 5 amps since this feature was not adjustable. To assist in determining what the kV and time exposure should be for taking a suitably exposed X-ray a quick test was undertaken with Dupont NDT Rapid Film. Based on the quick test results, the radiograph illustrated in Figure 12.1c was taken at 45 kV for 60 seconds. A Philips-Norelco industrial MG300 machine; 150 kV tube; exposure 35 kV; 5 mA for 60 seconds without filters; focus to film distance was approximately 980 mm; the film was Kodak Industrex M. The equipment was housed in a lead lined room designed for the purpose. Manoeuvring of the object and the X-ray head was limiting. The settee was delicate, fragile and heavy and took up most of the floor space. The radiography was undertaken in 2004. All images were digitised onto an AGFA GE FS50B industrial X-ray film scanner by Sonia O’Connor, AHRC Research Centre for Textile Conservation & Textile Studies, Research Fellow in Conservation, University of Bradford. The radiography was undertaken in 2004 using Windsor Medical X-ray equipment, Sovereign II Model. The unit is mounted onto a mobile stand and is easily dismantled into five component parts – originally designed to be taken into people’s homes.

184

5.

Case studies Being a piece of medical equipment, it is designed for human tolerance (i.e. living, moving tissue). Consequently, this unit operates at higher kV and short exposure times; the radiographic output ranges from 45 kV 22 mA to 90 kV 10 mA. The distance from the X-ray beam focus to the film was approximately 1300 mm. The control unit was set at 60 kV 20 mA for 30 seconds for the image shown in Figure 12.4b. Agfa Structurix non-screen film was used. Proposal put to Sonia O’Connor by Kate Gill.

Acronyms AIC WAG

American Institute of Conservation of Historic and Artistic Works Wooden Artefacts Group

References Buck, S. L. (1991). A technical and stylistic comparison of twelve Massachusetts State House Chairs. Wooden Artefacts Group. In Preprints AIC Washington, unpaginated, AIC. Gill, K. (1998). A Pilot Documentation Exercise concerning a Suite of Furniture at Knole, Sevenoaks, TCC Reference 2360.1-4. (Unpublished report, Textile Conservation Centre, University of Southampton.)

Gill, K. (2001). Eighteenth-century close-fitting detachable covers preserved at Houghton Hall: a technical study. In Upholstery Conservation: Principles and Practice (K. Gill and D. Eastop, eds), pp. 113–143, ButterworthHeinemann. Gill, K. (forthcoming). An evaluation of a portable X-ray unit for the examination of fragile upholstered furniture inside historic houses. The Forgotten History – Upholstery Conservation Postprints. Carl Malmsten Center of Wood Technology & Design and Birgitta Forum at Linköping University. Gill, K. and Doyal, S. (2001). A brief object record: the Brooklyn Museum of Art easy chair. In Upholstery Conservation: Principles and Practice (K. Gill and D. Eastop, eds), pp. 186–192, Butterworth-Heinemann. Gill, K. and Eastop, D. (1997). Two contrasting minimally interventive upholstery treatments: different roles, different treatments. In Textiles in Trust. Proceedings of the ‘Textiles in Trust’ Symposium held at Blickling Hall, Norfolk, September 1995 (K. Marko, ed.), pp. 67–77, Archetype Publications/The National Trust. Gill, K., Soultanian, J. and Wilmering, A. M. (1990). The conservation of the Seehof Furniture. Metropolitan Museum Journal, 25, 169–173. Kisluk-Grosheide, D.O. (1990). The Garden Room from Schloss Seehof and its furnishings. Metropolitan Museum Journal, 25, 143–160. Podmaniczky, M. S. (1990). Wooden frame conservation techniques. In Upholstery Conservation. Preprints of a Symposium held at Colonial Williamsburg, February 2–4, 1990 (M. A. Williams, ed.), pp. 29–41, American Conservation Consortium.

13 The use of X-radiography in the Textile Conservation Laboratory, Opificio delle Pietre Dure, Florence: methodological, technical and research approaches towards a non-invasive investigative technique Susanna Conti and Alfredo Aldrovandi

Introduction: concepts and issues The ideal for the conservator and, most importantly, for the work of art would be to extrapolate all possible documentary and technical data solely through simple visual observation alone but this is not, for the moment, possible. The next best option is the use of non-invasive techniques, such as radiography. The systematic use of this technique for investigation in textile conservation is a recent development. In most cases, this is because the advantages that this kind of investigation can bring to a textile conservation project are not well known. It may also be due to a lack of investigative curiosity as well as concerns that radiography may cause damage to artefacts. Mantler and Klikovits (2004) have explored this issue in relation to X-ray fluorescence analysis but there is no evidence in the literature relating to damage caused by radiographic imaging at the kilovoltages (kV) and exposure durations generally used (see Chapter 6, pp. 91, 93–4). Besides this, there have also been problems resulting from the traditional classification of textiles as an ‘applied art’. This meant that they were considered as functional pieces and not as ‘fine art’ objects. Over the last few decades, this attitude has changed resulting in increasing research into the conservation of textile artefacts and the dissemination of that knowledge. However, it is important to emphasise that conceptual questions still remain, including establishing the best process for the identification of a work of art and whether a single approach may realistically be applied to all artefacts. Many famous Italian

theorists have discussed these issues, especially that of the concept of methodological unity.1 Brandi (1972) was the first in Italy to consider this. Baldini (1978: 5–7) developed this concept, trying to create a critical framework for the complex and sensitive operation of conservation for all artistic works. Bonsanti made an interesting comment at the end of the 1990s: If today conservation can no longer be characterised as such on the basis of the object to which it is applied – ‘the work of art’ – then it can only be defined precisely by those ‘practical processes’ rejected by Brandi, the differentiation of which lies in the facts, in the order of things. In other words, our intervention would not be classed as an act of conservation if we were to treat, say, Botticelli’s Primavera in a cursory or formulaic manner (assuming, of course, such a thing were actually possible). Whereas our intervention could indeed be classed as an act of conservation if our approach is fuelled by a spirit of innovative curiosity, conscientiously applied, with a sensitivity to the problems at hand, even if the artefact being treated could lay only dubious claim – if indeed any claim at all – to the title of ‘work of art’ . . . if not quite a ‘theory’ as such, then perhaps a useful practical tool to help focus, inform and guide the thousand and one decision-making processes we embark upon in the conservation workplace each and every day.2 (Bonsanti, 1997: 111–112) In the spirit of this quotation, it should be noted that better knowledge and results can be obtained in the field of conservation through sensitive, intelligent cooperation (as this book itself shows). Above all, 185

186

Case studies

improved knowledge is gained through the ‘innovative curiosity’ which must form part of every conservator’s education and practice.

Selecting appropriate analytical approaches In choosing a particular analytical approach, it is necessary to take into consideration the unique nature of the work of art and its integrity. Along with this concept, it is obligatory that the conservation treatment must take into account the complexity of the work of art, including its intangible aspects. However, the focus is generally upon the materiality of the artefact and its physical appearance and condition (Brooks et al., 1996 and Kroller, 1996: 450–415). Therefore, analytical techniques which do not require sampling are to be preferred to those which require even a minimal amount. Furthermore, non-invasive techniques should be considered including radiography in combination with IR reflectography and, in some cases, UV fluorescence, to enable an even better understanding of the object. It is useful to develop a sampling strategy of significant areas to inform the conservation treatment. When a clearer understanding of the situation has been established, appropriate micro-sampling can be carried out in the selected areas. It will eventually be possible to develop a complete analysis of an object’s condition before undertaking any complex treatment. This approach has the direct and positive consequence of gathering further and more precise information which will be available for future research, thus building a database which leads towards a complete understanding of historical context, technical processes and aesthetic purposes (Cordaro, 2000: 33–37). Such data would be useful to future conservators for the preservation for works of art as well as for preventive conservation strategies.

Radiography applied to textiles: technical issues The specific radiography techniques required for textiles have been discussed earlier (see Chapter 3). As noted, low energy or ‘soft’ X-rays need to be used to get the optimum contrast, with careful attention to appropriate exposure times. Exposure evaluation is fundamental in order to obtain the maximum information from radiography. This needs to be assessed

using a very sensitive instrument. The M-Gil electronic dosimeter by Gilardoni is used at the Opificio delle Pietre Dure (OPD). This is equipped with three separate probes and allows a more precise evaluation of the necessary parameters. The results from this dosimeter have completely satisfied expectations. In every case, it is important to carry out preliminary tests on film using different X-ray energies and exposure times. When X-raying a textile artefact, the first aim is to study the original technique used in its construction (physical structure, material characterisation and so on) and its condition (alterations, previous treatments and so on). The construction and condition may reflect not only the manufacturing process, but also the physical impact of events which have occurred through a textile’s history. These may vary substantially, making the object unique. This situation is very different from that of medical and industrial radiography where the structure is usually well known and the diagnostic aim is the understanding of traumatic or pathological situations with respect to a reference standard. The difficulty and complexity of interpreting the radiograph of a textile artefact can therefore be easily imagined, particularly when this may be the only diagnostic information available.

Radiography of large textiles Lengthy experience has shown that a radiograph limited to a few areas may not always answer specific questions. The individual character of an artefact may make it necessary to X-ray a whole textile in order for general observations to be drawn; for instance, to help explain the condition of some parts. The reasons why the conical projection of the X-ray beam will produce an image slightly bigger than the object have been explored previously (see Chapter 2, pp. 17–8 and Chapter 3, pp. 44–5). In the case of two-dimensional textiles, the radiograph will be only slightly bigger than the original. In contrast, the differences can be substantial in the case of three-dimensional textile artefacts, especially when the film is only partially in contact with the object. Several methods can be used when X-raying large textiles. The largest easily available X-ray films measure 430 mm 350 mm. It is possible to X-ray one area at a time with such single films and then mosaic the single films together in order to obtain a complete image. Even though this process has been simplified by the development of digital imaging, it

The use of X-radiography in the Textile Conservation Laboratory 187

still has a disadvantage. As each exposure is made by conical perspective projection, the combination image contains distortions produced by each point of shooting. However, this method works fairly well for two-dimensional textile artefacts, such as flags or embroideries. With three-dimensional textile artefacts, such as dresses, this process may result in unreliable images because of the optical distortion of the parts of the object which are at some distance in front of the film and the X-ray beam. In such cases, there is only one solution: a single exposure made by placing the radiation source at such a distance that the incident beam radiation totally covers the object. This beam is normally placed centrally and perpendicularly to the object. In the past, an entire film of the right size for the image was made by joining numbers of standard sheet films together. Large film in rolls (1.27 metres wide and 50 metres long) has now become available. The film is cut to length as required by the size of the object and placed in a black plastic sealed envelope which is opaque to visible light and has a negligible radio-opacity. This envelope is also used to develop the film. Developer is introduced into the envelope by means of a tube attached to a corner, as is the fixing bath. This operation allows the complete treatment to be carried out anywhere, not only in a dark room, and utilises very small quantities of development chemicals.

When using large film, the choice of time exposures is determined in two stages. First, the time required by an exposure is determined by placing the object one metre from the beam focus (focus to film distance – FFD). Then the exposure required for the real FFD is calculated by means of the inverse square law (see Chapter 2, Note 1, p. 16). As long exposure times are often required (sometimes of several hours), it is necessary to have a radiographic unit equipped with a cooling system to be capable of continuous operation (Aldrovandi, 1999; Aldrovandi and Ciappi, 1995).

The use of radiography at the Opificio delle Pietre Dure Radiography has been used in the textile conservation department of the Opificio delle Pietre Dure since the 1980s. It was used in several studies on textile manufacturing techniques. Research has been undertaken into the interpretation of these images in order to understand both the radiographic signature of the materials involved (various fibre typologies, their deterioration and deformation) and interactions between these fibres and other materials (wood, ivory, metals, resins, etc.). This interpretative work is further complicated by the fact that such objects are often so severely damaged that they can scarcely be handled.

Table 13.1 Summary of case study objects Object

Date

Dimensions

A Chasuble

14th–18th and 19th century

1.15 m 1.64 m

B Chinese screen

19th century

C Fragment, velvet

Second quarter of 15th century; from a 19th century collection Second quarter of 16th century

D Mitre

E Dressed statue of an unidentified saint F Wax sculpture ‘The Death of the Virgin’

18th century head, arms and feet on a 15th century structure Early 19th century

Materials

Silk, linen, silk embroidery threads, wrapped and membrane gold and silver metal threads 0.76 m 1.7 m (0.76 m Wood, paper, silk, metal, adhesive, 0.57 m per panel) pigments 560 mm 250 mm Silk, blue cardboard 295 mm 285 mm Full height including lappets 720 mm Height without base 1.27 m 0. 62 m Height with base 1.32 m 330 mm 190 mm

Cardboard, linen, silk, membrane, metal threads, gilded copper, precious stones Wood, iron, terracotta, silk, pigments, spun gold, cotton, linen, jute(?) Wax, metal, silk, pigments, purl and spun gold and copper metal threads, glass beads

188

Case studies

Table 13.2 X-ray data Object

OPD X-ray archives number

X-ray film type

Focus to film distance (m)

Voltage (kV)

Anodic current (mA)

Exposure time (minutes)

A Chasuble B Chinese Screen C Velvet fragment D Mitre E Dressed statue of an unidentified saint F Wax sculpture ‘The Death of the Virgin’

RX 336 RX 272 RX 417 RX 430 RX 433

XDA-plus XDA-plus Agfa Curix D4 XDA-plus

2.30 4.00 1.00 1.20 3.50

22 22 15 22 22

5 5 5 5 5

7 70 1 6 40

RX 438

D4

1.20

20

5

4

Table 13.1 summarises these studies. Unless specifically noted, the radiography was carried out using unencapsulated film to avoid interference. As radiographic film is very sensitive to visible light, the exposures were carried out in an X-ray dark room using an Art-Gil machine manufactured by Gilardoni (Mandello Lario, Lecco, Italy). Details of the radiography procedures used are listed in Table 13.2.

Pilot study of the use of radiography in textile conservation: case study of a chasuble The first study was carried out on a chasuble (communion vestment); see Table 13.1–A (Bonito Fanelli, 1994 and Devoti, 1989). The objective was to minimise the amount of handling while exploring hidden elements. This study resulted in a great deal of new and very useful information which was of considerable interest to researchers in the field. It is still considered to be a key study from the OPD textile conservation laboratory. The Chasuble The chasuble’s exterior face is made with two different types of textile: a white silk thirteenth century lampas brocaded with silver and gilded membrane threads and a red silk taffeta decorated with opus senese3 embroidery (Figure 13.1a). Most of the lampas was in very poor condition because of treatments used in silk processing, the long period of use and the high degree of photochemical degradation caused by an inappropriate exhibition environment.

In addition, the chasuble had many old repairs and darns scattered all over the surface. The chasuble linen lining is still in good condition and very flexible. Despite the condition of the silk, this stable foundation made it possible to examine the object. Visual observation of the deteriorated surface and its technical construction, while absolutely essential to understand the chasuble’s condition, was not enough. It was necessary to safeguard the object from damage during all the movements necessary throughout the radiography. In fact, turning the chasuble was impossible because of the risk of losing detaching fragments. Besides this, there were interesting questions about unknown factors which guided the approach to the radiography. What was concealed by the central part of the silk-embroidered Eucharistic symbols? These included a rayed Host (communion wafer), executed in an eighteenth century thread and so from a different period than the late fifteenth century opus senese embroidery. Did traces of the missing braid exist? What was the condition of the chasuble’s construction stitching? A useful discussion with the head of the X-ray department at OPD resulted in the taking of two largescale radiographs, one of the front and one of the back of the chasuble. Preparing the chasuble for radiography One of the most significant problems arose when preparing the chasuble for the radiography exposure. Moving the chasuble was very difficult because of its size and poor condition. To enable it to be moved safely, a light-weight, fire-resistant and highdensity foam polystyrene panel was custom-made,

The use of X-radiography in the Textile Conservation Laboratory 189

(a)

(c)

covered with a silk taffeta fabric and inserted inside the chasuble. A silk covering was chosen to avoid friction with the chasuble’s linen lining. A protective silk overlay was placed over the front and back of the chasuble and stitched through to the lining where this was exposed. With the chasuble lying horizontally, an X-ray film was interposed between the custom-made panel and the linen lining; see Table 13.2–A for procedural details. Two exposures were taken, the front first and then the back (Figures 13.1b and c).

(b)

Figure 13.1 White silk lampas chasuble with opus senese embroidery before conservation, (a) photograph, front view, (b) radiograph of front section, (c) radiograph of back section. The sharply defined edges at the top of the radiographs show the limit of the film rather than the edge of the chasuble. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

Results of radiography Regular comparisons between the object and its radiographs were essential to gain full understanding of the data, including the development of explanatory diagrams of the construction technique and condition. The radiographs were constantly consulted during the conservation treatments which involved the removal of previous restorations, surface cleaning and stitched consolidation. Moreover, the radiography enabled reduction in the handling of the chasuble.

190

Case studies

(a)

(b)

Figure 13.2 Radiographs showing details of the front section of the chasuble, (a) lampas showing construction, repairs and condition, (b) metal thread embroidery. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

It was possible to establish some data more accurately, in particular information about the chasuble’s physical and mechanical degradation including the linen lining which was in good condition. The weaving technique, thread diameter and thread count, stitching, wear, tears and deformations, old insertions, selvedges and so on could be identified (Figure 13.2a). In contrast, it has not yet been possible to characterise the silk fully in all areas using the exposures listed in Table 13.2–A on account of the presence of the other textile layers. The different layers of materials can be most clearly identified at the edges. More specific research is needed into ways of differentiating the silk from the other textile components. Some embroidery, presumably original, became evident. Figure 13.2b shows the location of the metal thread embroideries where the letter ‘a’ is now visible below the silk split-stitch embroidery. The letter was executed with spun gold thread and is located beneath the embroidered Eucharistic symbol. This was worked over the earlier embroidery when the chasuble was reused, a very frequent custom in ecclesiastic circles. As far as the stitching of the lining and the original lampas is concerned, all the elements of the chasuble’s construction and other previous repairs could be seen. The selvedges and the decorative pattern of the lampas are visible thanks to areas brocaded with a silver membrane thread. Evidence about the former dimensions of the lampas and lining fabric was also obtained. The radiography also reveals deposits of silk fragments and silver membrane threads along the edge of the chasuble inside the lining. Unfortunately, it was not possible to obtain any further information

about the missing braid. A different X-ray exposure might enable some evidence of this to be identified. In summary, radiography allowed the interventive treatment to be minimised and informed preventive conservation by confirming the need to display the chasuble horizontally

Case study: a Chinese screen A diagnostic radiographic study was carried out on the Chinese screen, probably executed using Eastern techniques under Western influence (Table 13.1–B; Carlano, 1979). Following the model of the chasuble, it was hoped that radiographing the screen would provide data on its condition, so aiding the development of a minimum conservation strategy as well as furthering understanding of its construction. These panels are mounted in an extraordinarily beautiful bronze frame decorated with the House of Savoy coat of arms, presumably dating from the mid-nineteenth century (Figure 13.3a). The frame was removed to be treated separately. Each of the three panels which form the screen is made of four or five small wooden boards, jointed together and lined with a red silk satin fixed along the outer edge by large-headed nails (Figure 13.4). The front of the panels is covered in another red silk satin, glued with starch paste to a Japanese paper support. Each panel is decorated with figures worked in modelled cardboard, padded and lined by silk textiles, such as lampas, damask and ermesini.4 These figures are also glued to the red satin with starch paste and fixed, in areas with the greatest tension, with

The use of X-radiography in the Textile Conservation Laboratory 191

(a)

(b)

Figure 13.3 Chinese embroidery panels mounted in a European bronze screen, (a) photograph, front view, after conservation, (b) radiograph, before conservation. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

192

Case studies

Figure 13.4 Diagram showing construction of the Chinese embroidered panels. Key: A Lined wooden support: wooden board with bamboo D Construction of decorative elements: cardboard, inserts, nails, printed paper, textile padding, textile, nail B Paper and textile ground (Line drawing by Susanna Conti.) C Paper and textile

small, round-headed nails. The textiles are arranged with masterly skill to give volume and physical realism. These figures are decorated with starch-pasted paper, painted inserts and sequins and are clearly the most important element of the design (Conti and Vaccari, 1999). Preparation of the panels for radiography A single radiograph was taken of the three panels to provide evidence of the screen’s condition as structural damage had occurred to the panel on the left (Figure 13.3a). This image would also enable a more accurate comparison of the materials and problems present in each panel. The most important aim was to carry out the conservation operations while avoiding total disassembly of the object. Only the red silk satin lining on the back of the broken panel was removed to allow access to the wooden pine panel during radiography. The panels were placed horizontally one beside the other on an inert semi-rigid polystyrene support. The X-ray film was inserted beneath the panels so that it was in direct contact with the inert support, taking great care of the broken panel; see Table 13.2–B for procedural details.

Results of radiography These high quality, large-scale radiographic images provided a great deal of information about the construction, condition and degradation of the materials of this screen (Figure 13.3b). The two cracks in the left wooden panel can be seen very clearly in the radiograph (Figure 13.5). The crack to the right of the join formed a longitudinal split which stopped only a few centimetres from the upper edge. This put the panel at risk of breaking in two, with catastrophic consequences for the object’s integrity. Further information was provided on many other aspects of the screen: ●

The unexpected discovery of the support structure which was formed from four or five coniferous wooden boards joined edge to edge by bamboo dowels inserted into drilled holes into the thickness of the boards (Figure 13.5). The dowels, which are characteristic of Eastern screens, confirmed the hypothesis of oriental manufacture. The lack of glue and the use of the dowels showed that the screen had been made so that the panels were free to move, important as the whole artefact

The use of X-radiography in the Textile Conservation Laboratory 193

●

white features in the radiographs (Figures 13.5 and 13.6a). Pigments and gold powder, probably blended with Arabic gum and used mainly in the painting of the figure in the right panel, could be detected. The gold floral motifs used on the blue and white printed paper on the back of the panels became visible (Figure 13.6b). This paper is the original lining, subsequently covered with the red silk satin. Knowing that the silk satin lining was not the original lining enabled us to remove it from the broken panel.

In summary, radiography provided an excellent overview of the screen’s materials and construction. This helped to refine the treatment strategy so that it focused on specific areas and damage types.

Case study: a fifteenth century velvet fragment from a nineteenth century collection

Figure 13.5 Radiograph showing detail of left wooden panel. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

●

●

●

●

was made with very hygroscopic materials. It also became evident that the object was not very stable because the panels were not very thick. The radiograph confirmed that these nails along the edges of the panels were not original but were added later to repair earlier damage. The results of this can be seen in the deformations of the red silk satin ground, particularly in the panel on the right side in the photograph (Figure 13.3a). The presence of glue in very limited quantities on all the figures, as seen on the figure of the girl with the lute, had not been detected before. A previous repair, carried out using an excellent technique of glue and little strips of paper to renew adhesion to the panel, was identified through the radiography (Figure 13.6a). The unexpectedly high radio-opacity of the textiles glued onto paper and padded enabled a better characterisation of this technique (Figure 13.6a). The presence of many knots caused weak points in the wooden panels. They show as blurred, circular

Thanks to recently increased awareness of the importance of our textile heritage, private collectors and established textile industries have given museums a considerable amount of material, including machinery, pattern books and costume. The acquisition of textile fragments from the antique trade began in the second half of the nineteenth century and continued into the early years of the twentieth century, resulting in the reappraisal of ecclesiastical vestments and archaeological finds. This led to new studies and historical documentation, important for the developing museum sector, private collections and fashion designers. Such fragments were often glued or, in a few rare cases, stitched onto various kinds of paper or cardboard, almost never suitable for conservation purposes (Nagasawa, 2000). The subject of this conservation study forms part of a larger collection belonging to the Textile Museum in Prato (Table 13.1–C). A survey was carried out on how these fragments could be detached from their cardboard supports without damage while retaining documentary data in order to meet curatorial inventory requirements and to store them systematically. Moreover, precious space could be saved in museum storage facilities. The fragment studied was a green pile on pile silk velvet with an irregular edge. It had been glued onto a light blue cardboard mount with an acidic pH (Figure 13.7a). Radiographic analysis was proposed to locate the exact position of the adhesive

194

Case studies

(a)

(b)

Figure 13.6 Radiographs showing, (a) previous repair using strips of glued paper, central panel before conservation, (b) gold on the printed paper fragments, right panel. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

used to fix the fragment onto the cardboard. In this particular case, it was not expected that any particular contrasts would be revealed as the fragment was woven exclusively in silk. The weaving technique is actually a very important factor as it gives textiles characteristics which influence their behaviour over time. In fact, the density and different thicknesses of the velvet in contrast to the underlying cardboard support provided a great deal of information. Preparation of the textile fragment for radiography As little variation in contrast was expected, a copper thread was placed around the edges of the textile to map its location on the cardboard. The fragment, still glued to its support, was placed horizontally on the radiographic film; see Table 13.2–C for procedural details.

Results of radiography Old stitching, some weaving faults, the decorative pattern and vertical chain lines relating to the manufacturing of the cardboard support could be detected. Thanks to its relatively high radio-opacity, areas with adhesive were shown as white, almost circular, rings; these are indicated by arrows on Figure 13.7b. The removal of the fragment from its cardboard support was carried out mechanically with very good aesthetic results. The use of radiography in this specific case was undoubtedly useful but the problem of a large-scale operation and the cost/benefits balance still remains when considering the whole collection.

Case study: a mitre This mitre is the traditional shape for bishop’s headwear of the sixteenth century (Table 13.1–D).

The use of X-radiography in the Textile Conservation Laboratory 195

(a)

this would enable treatment to be carried out without having to deconstruct the mitre. It has a red silk cut velvet ground which is embroidered and decorated with stones, miniatures set in mounts and braids along the edges (Figures 13.8a and b). The mounts, decorated with stones, contain painted images on parchment which are further ornamented with baroque pearls and covered with quartz crystals. Other stones, set on gilded copper supports, decorate the velvet ground. The two lappets, narrow strips which hang from the back of a mitre, are also of red silk cut velvet and ornamented with applied embroideries and two metallic thread braids. A fragment of a fringe remains on one. Deformations and damage were evident in areas that had undergone greater mechanical stress. There were also wax stains and superficial fatty stains, mainly located on the embroidery, with traces of inert microbiological damage. The important ethical issue of whether or not to disassemble certain textile artefacts has been overcome in the OPD by making every effort to undertake all possible research in order to obtain the maximum amount of information. This enables the conservation principle of minimum intervention to be respected (Ciatti, 2005). The main reason why radiography was required here was the need to know about the pressed cardboard structure of the mitre and how some of the metals present, besides the ones decorating the mitre, might influence its future condition. This would influence the decision as to whether substantial changes might be required to the structure of the mitre. There was also great curiosity about all those materials used in the construction of the mitre which were still to be fully investigated. Preparation of the mitre for radiography

(b)

Figure 13.7 Velvet fragment on cardboard mount, (a) photograph with the area radiographed indicated, (b) radiograph detail; the arrows show spots of glue. The intersecting bright lines are the copper wires used to outline the textile. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

It is open on the top but stitched on the right and left sides (Demori-Stanicic´, 2001). It was radiographed in order to understand the construction of the applied elements and the deformation in the hope

The mitre was placed on a horizontal inert panel. Two radiographs were taken of the front and back of the main body (central part) of the mitre using a 300 mm 400 mm film; see Table 13.2–D for procedural details. Care was taken in inserting the film into – and through – the mitre. To avoid damaging the stitching, the edges of the film were kept away from the right and left sides. As both sides of the mitre were much flattened, inserting the film up to these edges might have caused damage to the fibres of the fabric in these areas. One side was exposed first and then the mitre was turned over to radiograph the second side. This method did not result in an entire image as the sides of the mitre were beyond the sides of the film. It was later decided to

196

Case studies

(a)

(b)

Figure 13.8 Mitre detail before conservation, (a) front view before conservation, (b) back view after conservation. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

radiograph the sides with the film placed below the mitre to complete the information. This produced an image with the detail of both the front and back surfaces of the mitre superimposed on each other; see insert on the right of Figure 13.9a. Results of radiography The radiographic image showed all the spun gold threads used in the embroidery, including information on their density and various losses (Figure 13.9a). It was interesting to establish that the silk thread used for the embroidery of the faces could not be seen while the wax stains and pigments were visible. Two metal fittings (a small fastener and a pin) were visible. These were clearly of a later period and were easy to remove. Observation of the radiographic images of the details of the sides of the mitre indicated that a slight underexposure enabled better reading of the data (Figure 13.9a). Traces of the thread used in the stitching of the linen lining and its folds became evident under the velvet (Figure 13.9b). The layout of the areas relating to the jewellery work became evident and confirmed that these were in good condition. These areas have been analysed with IR and UV multispectral analysis which has identified some of the pink, green, blue and yellow ‘stones’ set in the brass oval-shaped gilded mounts as

glass. It was possible to verify that the general condition of the mitre was rather good. Its structure was not a cause for concern. No data connected to the cause of the visible deformations was obtained from the radiographic images. This provided further reassurance that there was no need to disassemble the mitre. It was necessary to create a suitable support structure for display so that the pressure points at the sides of the mitre did not stress the stitching. A shortterm period of rotating exhibition has been specified.

Case study: dressed statue This statue (Figure 13.10) represents an unidentified female saint. No documentation concerning its provenance or its previous locations exists with the exception of a date on a small cardboard label reading ‘Eighteenth century’ (Table 13.1–E, Figure 13.10a). The study of this statue was fascinating, not only because of the OPD’s usual study of the documentary data and construction techniques, but also because of the museum curator’s request. All the garments with which the statue was dressed were to be completely removed, right down to the underlayers, to meet teaching and training requirements. This aspect was very important for the teaching programme but,

The use of X-radiography in the Textile Conservation Laboratory 197

(a)

(b)

Figure 13.9 Mitre, radiographs, (a) detail, front and, on the right side, detail of the edge with both front and back superimposed, (b) detail of back. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

before it was done, it was critical to know the real state of the statue including its condition and the number and nature of garments with which she was dressed. This kind of statue is an assembly of parts made of different materials from various workshops. The figure stands on a wooden base. It was established that the statue had a painted terracotta head, arms and feet and, probably, wooden legs. The arms have articulated joints and there is a protruding nail in the left hand. The figure wears sandals. The presence of a halo, now lost, is suggested by a screw on top of the head and the loss of material in the area above the nape of the neck. The statue’s outer dress is made of two different types of textile: a c. 1730 Revel-type lampas in front and a seventeenth century damask at the back. Two cotton undersleeves, decorated with lace, emerge from the three-quarter length sleeves. The neck opening is covered with a lace collar worked in a vine-shoot motif. The cut of the dress suggests fashions of the 1840s–1850s. A remarkable quantity of undergarments, of different types and in varying condition, appeared under the dress’s hemline. General examination suggested that the statue was in a fairly good condition. It was not possible to establish the real condition of the joints in the arms which were only partially moveable. The front of the outer lampas dress did not have any losses and only partial photochemical degradation. However, it was clearly being placed under tension by the garments underneath. The first decision, taken together with the curator, was to remove the upper dress in order to evaluate the next steps as this garment did

not, at that point, present any particular conservation problems. The second dress, a white starched heavily creased cotton undergarment, was at least two sizes too big for the statue (Figure 13.10b). It was removed from the statue. The style of this undergarment was very interesting; it appeared to be an example of late nineteenth century underwear. At this stage, another more modest, chaste white cotton undergarment in a size suitable to the statue became evident. The proposal to carry out radiography was then formulated. This aimed to gain an understanding of the issues which had to be considered in order to justify the removal of the other undergarments. Preparation of the statue for radiography It was important to decide which aspect of the statue would be best for radiography. As it was a threedimensional object, several views were possible. In this case, the front was chosen, as it did not show specific damage and only the total image would characterise all aspects. A large, especially prepared film was placed behind the statue with the X-ray beam placed centrally and perpendicularly in front; see Table 13.2–E for details of radiography procedures. Results of radiography The radiography helped to determine the previous history of the statue and to develop the methodology of the conservation treatment. 5 Three more undergarments, placed one on top of another, were evident on the radiograph (Figure 13.10c). Some details

198

Case studies

(a)

(b)

(c)

Figure 13.10 Dressed figure of a saint, (a) front view, fully dressed before conservation, (b) front view showing the third layer of clothing, (c) radiograph. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

The use of X-radiography in the Textile Conservation Laboratory 199

radiography, the curator was consulted and it was agreed that the statue only needed to be partially undressed.

Case study: wax sculpture Dormitio Virgini (‘The Death of the Virgin’)

Figure 13.11 Radiograph detail showing the structure of the proper left arm. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

of the first of these were visible. Reinforcements added to the skirts of the second undergarment to give volume to the dress became visible. The third undergarment was characterised by a coarser textile together with some thickness at the hips. This suggested that non-woven (jute?) fibres had been placed there to create volume. The radiograph of the statue’s support structure was very interesting (Figure 13.11). A much older construction than was expected was revealed. According to wooden sculpture experts, this could probably be dated to the fifteenth century. Using radiography plates in the OPD archives, a comparison was made to a sculpture of the Annunciazione (Annunciation) executed by Mariano Agnolo Romanelli, a fifteenth century Tuscan sculptor. Again, this type of statue was often constructed by assembling elements from various artisans’ workshops, generally located in southern Italy. As a result of the information gained through the

This case study examines a wax sculpture showing the Death of the Virgin. This probably came from a convent and was produced in Sicily, where this style was very popular (Table 13.1–E, Figure 13.12a). This subject of the death of the Virgin Mary was common from the thirteenth century onwards. Radiography was primarily undertaken to understand the condition of the underlying structure, particularly because of the juxtaposition of the wax and textile components. This information would help to determine the most appropriate treatment and the order in which this needed to be carried out. This sculpture represents the Virgin wearing a white dress, and, following traditional iconography, lying on a bed, with her eyes closed and her hands stretched out along her sides (Palmisano, 2004). The wax limbs and face are pressure inserted into the internal structure of the sculpture. The dress has a silk satin ground fabric embroidered in a design of flowers and sprays, symmetrical about a central axis using a golden metal purl thread. The pistils of the flowers are in pearl-coloured and red glass beads. This embroidery is only worked on the front of the dress and sleeves. The statue’s legs are clad in a pair of drawers in a white, plain weave cellulosic textile. The sculpture arrived in the OPD laboratory lying on a modern embroidered cushion. It was completely covered with dust. The limbs were partially mutilated and part of the gathered flounce showed wear from use. Photodegradation and dust had caused damage to the fibres. Mishandling of the statue had worsened its condition. It was immediately clear that an interdisciplinary approach, as is usual at the OPD, was needed. Radiography was required to determine the statue’s structure more clearly and to clarify the different problems in order to develop the course of the treatments (Conti et al., 2004). Preparation of the wax sculpture for radiography The very delicate sculpture was placed horizontally on an inert panel and radiographic film, without an envelope, was interposed between the two. The radiation exposure was carried out perpendicularly

200

Case studies

(a)

(b)

Figure 13.12 Wax sculpture of the Dormitio Virginis, (a) photograph before conservation, (b) radiograph. (© Ministero per i Beni e le Attivita’ Culturali. Opificio delle Pietre Dure e Laboratori di Restauro, Firenze.)

and centrally; see Table 13.2–F for details of the radiography procedures. Results of radiography The radiograph revealed that the main structure was made with twisted iron wire armature, reminiscent of a human skeleton (Figure 13.12b). The wax elements were inserted into this armature. Unexpected tow fibres were evident in the abdominal area, giving more volume here. Breaks in the purl thread along the hem of the gown as well as the disrupted position of the feet could clearly be seen. The condition thus revealed the necessity to carry out the conservation treatment of the anatomical elements first.

Carrying out conservation treatments on the textiles first would have further weakened the underlying structure, even if executed with the utmost care. It was necessary, on the other hand, to protect those textiles in poor condition from the handling necessary to carry out consolidation to the wax elements. Fine silk overlays were placed over the weakened areas before delivering the object to the wax conservators. After conservation of the wax parts, the overlays were removed one by one. The textiles were then treated one at a time using appropriate techniques. Observation derived from the radiographs suggested that it was necessary to make two small pillows to stabilise and support the feet which were otherwise going to bend over on their sides.

The use of X-radiography in the Textile Conservation Laboratory 201

These were placed inside the skirt, one on top and one under the legs.

2. 3.

Discussion The examples presented in this chapter are only a few of those treated in the OPD textile conservation laboratory. However, they represent a significant group for which specific X-ray diagnostic methodologies have been formulated to satisfy the conservator’s requirements. They are still very limited, considering the variety of textile techniques. Although it cannot be a substitute for the conservator’s analytical and critical examination, which comes from experience acquired over time, radiography can become one of the main methodologies to provide information for textile conservation decision making. Such information is, above all, a foundation for developing an understanding of the condition of an object. X-ray examination can enable greater understanding of construction techniques, which are a fundamental aspect of object consolidation, and thus improve treatment so enabling artefacts to be returned to museum display. Radiographs can also lead to more discriminating and reliable research into identifying and differentiating materials. All this evidence can be enhanced by digitising radiographic images. Non-invasive analysis will, it is hoped, be employed as standard practice during the study of a textile work of art, a unique point where the expertise of the conservator, radiologist and curator can interact. Growing research will also hopefully enable the development of the qualitative and quantitative data, resulting from the use of more sophisticated equipment such as computerised axial tomography (CT) coupled with increasing expertise in interpretation of evidence from the widest possible range of examples.

Acknowledgements The authors would like to express their thanks to the following: G. Bonsanti, C. Acidini Luchinat, M. G. Vaccari, M. Ciatti, C. Frosinini, O. Ciappi, M. Brancatelli, L. Cerretini, F. Cinotti, A. Keller and P. Cesari.

Notes 1.

There is a large bibliography exploring issues relating to the definition of a ‘work of art’; for example, see

4.

5.

Brandi (1972: 3–8), Conti (1992) and Bonsanti (1997). English translation by Gabriella Barbieri. Opus senese embroidery is a particular type of embroidery made in Siena at this period. For a definition, see the exhibition catalogue by Bonito Fanelli (1994: 53) Ermesini is a luxurious, very fine, light-weight silk usually used for women’s dresses. Its name comes from the town of Ormuz, Persia, where it was originally made. The conservation treatment carried out on the first dress consisted of humidifying the dangerously brittle fibres and in realigning warps and wefts where possible. The undergarment forming the second layer had been added, possibly for devotional reasons, into an unsuitable context. In fact, owing to its larger dimensions, it had been pleated in order to be put on the statue. The methodological choice made by the curator led to the removal of this undergarment. After wet cleaning, it recovered its original shape. Replacing it would have required compressing the garment and recreating the original pleating. This undergarment was, however, inseparable from the statue as it belonged to it historically and so will always be related to it.

References Aldrovandi, A. (1999). Indagini radiografiche nelle opere d’arte: principi di base e nuovi sviluppi. In OPD Restauro. Rivista dell’Opificio delle Pietre Dure e Laboratori di Restauro di Firenze, 11, pp. 146–157, Edizione Centro Di. Aldrovandi, A. and Ciappi, O. (1995). La radiografia di grande formato: problemi e soluzioni tecniche. In OPD Restauro. Rivista dell’Opificio delle Pietre Dure e Laboratori di Restauro di Firenze, 7, pp. 163–168, Edizione Centro Di. Baldini, U. (1978). Teoria del Restauro e Unità di Metodologia. Edizione Nardini. Bonito Fanelli, R. (1994). I ‘doni’ dall’Oriente e da ‘Champagne, Provinces et Ultramonti’: commerci, chiesa e moda. In Drappi, Velluti, Taffetas et altre Cose. Antichi Tessuti a Siena e nel suo Territorio (M. Ciatti, ed.), pp. 52–53, Edizione Nuova Immagine. Bonsanti, G. (1997). Riparare l’arte. In OPD Restauro. Rivista dell’Opificio delle Pietre Dure di Firenze, 9, pp. 109–112, Edizione Centro Di. Brandi, C. (1972; first edition 1963). Teoria del Restauro La Carta del Restauro del 1972. Edizione Giulio Einaudi. (An English edition is now available: Brandi’s Theory of Restoration (2005). Edizione Nardini.) Brooks, M. M., O’Connor, S. and McDonnell, J. G. (1996). The application of low energy X-radiography in the examination and investigation of degraded historic silk textiles: a preliminary report on work in progress. In

202

Case studies

Preprints, 11th ICOM Triennial Meeting, Edinburgh. ICOM Conservation Committee (J. Bridgland, ed.), pp. 670–679, James & James (Science Publishers). Carlano, A. (1979). Vicende della guardaroba di Palazzo Pitti. In Curiosità di una Reggia. Catalogo della Mostra. Edizione Centro Di, 250. Ciatti, M. (2005). Il minimo intervento nel Laboratorio della Fortezza fra utopia e applicazioni. In Atti del II Convegno Cesmar 7, Tema ‘Minimo Intervento Conservativo’ nel Restauro dei Dipinti, pp. 53–60, Il Prato. Conti, A. (1992). Restauro. Un enciclopedia d’orientamento. EDO, 32, p. 10, Edizione Jaca Book. Conti, S., Degli Innocenti, D., Laurini, S. and Rossignoli, G. (2004). Intervento conservativo su abito di statua devozionale del Museo del Tessuto di Prato: un assemblaggio di varie epoche. In Lo Stato dell’Arte 2 Conservazione e Restauro, Confronto di Esperienze. II Congresso Nazionale IGIIC, Genova, Palazzo Reale 27–29 Settembre 2004, pp. 340–349, II Congresso Nazionale IGIIC. Conti, S., Lorenzini, M., Bellina, R., Stragapede, M., Zingarelli, M. and Zonta, E. (2004). Fiori, cere, carta, volant e canutiglie. In Le Raccolte di Arte Devota Popolare dell’Abbazia di S. Spirito a Caltanissetta, pp. 37–55, Edizione Nuova Graphicadue. Conti, S. and Vaccari, M. G. (1999). Paravento Cinese. In OPD Restauro. Rivista dell’Opificio delle Pietre Dure e Laboratori di Restauro di Firenze, 11, pp. 326–332, Edizione Centro Di.

Cordaro, M. (2000). Restauro e tutela scritti scelti (1969–1999). Annali dell’Associazione Rancio Bianchi Bandinelli fondata da Giulio Carlo Argan, 8, pp. 33–37, Fotocomposizione e stampa CSF. Demori-Stanicˇic´, Z. (2001). Tesori della Croazia Restaurati da Venetian Heritage Inc. Catalogo della mostra, pp. 150–152, Edizioni Multigraf. Devoti, D. (1989). Un’ arte decorativa e industriale: ‘ . . . Centum XII Pannos Lucanos . . . de Serico cum Auro . . .’. In La Seta. Tesori di un’antica Arte Lucchese. Produzione Tessile a Lucca dal XIII al XVIII Secolo, p. 16, Lucca Museo Nazionale di Palazzo Mansi. Koller, M. (1996). Polychromed sculpture and textile arts. In Preprints, 11th ICOM Triennial Meeting, Edinburgh. ICOM Conservation Committee (J. Bridgland, ed.), pp. 410–415, James & James (Science Publishers). Mantler, M. and Klikovits, J. (2004). Analysis of art objects and other delicate samples: is XRF really necessary? Powder Diffraction, 19(1), 16–19. Nagasawa, T. A. (2000). Distacco di manufatti tessili antichi da supporti cartacei e/o eterogenei da adesivi. (Tesi di diploma – Settore Tessili. Opificio delle Pietre Dure e Laboratori di Restauro Corso 1998–2000. Relatori: S. Conti, M. Ciatti, I. Tosini, D. Degl’innocenti). Palmisano, M. E. (2004). Le raccolte di S. Spirito. Un museo per l’arte devota popolare. In Le Raccolte di Arte Devota Popolare dell’Abbazia di S. Spirito a Caltanissetta, pp. 17–29, Edizione Nuova Graphicadue.

14 The role of X-radiography in the documentation and investigation of an eighteenth century multilayered stomacher Gabriella Barbieri

Introduction The research described here was undertaken as part of a much broader investigative study entitled Memoirs of an Eighteenth Century Stomacher: A Strategy for Documenting the Multiple Object Biographies of a Once-Concealed Garment, and was submitted in partial fulfilment of the requirements for the MA Textile Conservation, Textile Conservation Centre (TCC), University of Southampton (Barbieri, 2003). The project was based on the detailed study and documentation of a multi-layered eighteenth century stomacher (Figures 14.1a and b) which was found concealed in the timbers of a fifteenth century thatched cottage in Nether Wallop, Hampshire.1

The Nether Wallop cache The Nether Wallop cache was discovered in 1978 while renovation works were being carried out. The objects were subsequently stored in the house until 2001, when they were taken to Winchester Museum by the owners for identification. The cache consists of three items: a boned stomacher (1700– 1770), a tan velvet waistcoat (1790–1840) and six paper patterns. Three of these are cut out of newspaper, one of which bears the inscription The London Evening Post, from Thursday May 14, to Saturday May 16, 1752. The objects were found in the wall of a first-floor bedroom, wedged into a natural knothole located in one of the timbers forming part of the original fifteenth century cruck frame of the building. The knothole opens into the wall cavity above the ground-floor inglenook fireplace and was

unsealed, being accessible from either the bedroom or the roof space. Research by the owners into the history of the building showed that the farmhouse was extended during both the seventeenth and eighteenth centuries, and it is likely that the cache was concealed while the latter alterations were taking place.

The practice of concealment: a contextual framework The term concealed objects is used to refer to artefacts which are found hidden, buried or built into the fabric or foundations of a building – e.g. in the walls, chimney or rafters, around entry points such as doors and windows, and under the hearth, threshold or floorboards. Such objects are frequently described as being deliberately concealed to indicate that their placement within the concealment site is believed to result not from accidental loss or rubbish disposal, but from a deliberate act on the part of the householder or builder. The practice of such concealment is thought to form part of a long-standing folk tradition whereby an appropriate object or group of objects – frequently, though not exclusively, garments or shoes – were specifically selected and deposited in strategic locations within the building as part of a superstitious ritual to protect the home and its residents against witches, evil spirits and general misfortune (Brooks, 2000; Easton, 1995; Merrifield, 1987). These objects tended to be placed at ‘weak’ points in the structure of the house, e.g. doors and windows, where the physical – and therefore spiritual – integrity 203

204

Case studies

(a)

(b)

Figure 14.1 Nether Wallop stomacher, (a) obverse face, (b) reverse face. (Photograph by Mike Halliwell; © Textile Conservation Centre.)

of the building was seen as being breached, and where evil spirits were believed to gain entry. The chimney or fireplace, which unlike windows and doors could not be closed off, seems to have been considered the most vulnerable point, and it is here that the greatest numbers of caches are found (Swann, 1996). Such artefacts are believed to have acted as substitutes for the occupants, a sort of decoy which would confuse the witches as they entered the building and divert them away from their intended victim(s). Hence the use of personal items, such as clothing or shoes, which would retain the imprint or ‘essence’ of the wearer.

The stomacher Randle Holme, in his contemporaneous definitions of ‘Terms used by taylors’ states that ‘The STOMACHER is that peece as lieth under the lacing or binding on of the Body of the Gown. . .’ (1688; cited in Arnold, 1977: 3). Such stomachers were decorative triangular panels designed to fill the gap between the edges of an open-fronted gown

(Baumgarten and Watson, 2000; Hart and North, 1998). They were worn from the late seventeenth to the late eighteenth centuries and were fastened either side of the bodice by means of tapes, laces or pins which would have been concealed from view by the dress robings (decorative revers either side of the bodice). By 1775, stomachers were becoming redundant as open-fronted gowns began to go out of fashion, replaced with bodices which closed edge-to-edge at the centre front. The Nether Wallop stomacher is a mixed-media artefact composed of five principal component layers which appear to have formed the basic structure of the original garment (Figures 14.2a and b). These are, working from the obverse through to the reverse face: ●

● ●

● ●

a ribbed yellow silk, now fragmentary and heavily abraded cream card for stiffening baleen (whalebone) boning strips to impart shape and structure a coarse, dark brown linen interlining a block-printed linen lining.

The role of X-radiography in the documentation and investigation 205 (a) Exploded view

(a) Exploded view REVERSE FACE

REVERSE FACE

Layers stitched together to form boning channels

Linen patch encasing left side edge

Printed lining

Printed lining

Parallel lines of backstitches

Linen interlining

Linen interlining

Card

Baleen strips Card

Ribbed silk Orange/cream wool twill

OBVERSE FACE

OBVERSE FACE

Ribbed silk (b) Cross-section OBVERSE FACE

(b) Cross-section

Ribbed silk OBVERSE FACE

Ribbed silk

Channel stitch lines

Card Card

Linen interlining REVERSE FACE

Figure 14.2 Original structure of stomacher showing sequence of component layers, (a) exploded view, (b) cross-section. (Drawing by Gabriella Barbieri.)

In addition to the five principal layers, there are also remnants of two further ‘patch’ layers which appear to constitute later additions, repairs or alterations to the original garment (Figures 14.3a and b). These layers consist of two fabrics: ●

●

Linen patch

Baleen strips

Baleen strips Printed lining

Orange/cream wool twill

an orange and cream wool twill, of which only a few, heavily soiled, fragments survive an open-weave, undyed linen.

The stomacher is in extremely poor condition, being heavily soiled and structurally weak with substantial areas of loss. In addition to deterioration resulting from concealment, the stomacher has undergone heavy wear and extensive alteration and/or repair during its functional life as a garment. There is also evidence to suggest possible deliberate damage prior to placement in the cache site. In fact, it would appear that the artefact had at least three different ‘object biographies’ (Kopytoff, 1986) during its lifetime. Starting life as a finely constructed undergarment (part of a pair of stays, a boned undergarment that was a precursor of the corset), it

Linen interlining REVERSE FACE

Printed lining

Figure 14.3 Altered structure of stomacher showing location of patch layers relative to main component layers, (a) exploded view, (b) cross-section. (Drawing by Gabriella Barbieri.)

was subsequently remodelled and recycled for use as a stomacher (probably as an item of working dress), before finally assuming a spiritual role as an object of ritual and magic.

Rationale for research Concealed garments embody some of the most complex ethical issues facing professionals involved in the care and preservation of textile and costume collections. They constitute prime examples of how objects in general – and garments in particular – can acquire changing social roles and functions which frequently have far-reaching implications on the subsequent interpretation, treatment and display of these artefacts (Eastop, 1998, 2000). The prime objective of any conservation strategy should be the preservation of an artefact’s true nature (UKIC, 1990). Yet in order to preserve that true nature effectively, it must first be clearly identified. Defining the true nature of concealed garments is, however, often a far from straightforward matter.

206

Case studies

First, such artefacts are severely under-researched as a field of study in their own right. Second, precisely because they have fulfilled a number of often widely disparate roles in their long histories, these garments do not easily fit into one particular object category within a museum collection. Should they be accessioned as part of the costume collection? As objects of social history? As archaeological artefacts? Third, their frequently exceedingly poor physical condition means that they can present seemingly insurmountable problems in terms of study, documentation, interpretation, access and display.

the stomacher. Both these findings would in turn inform conservation treatment decisions as well as, potentially, the institutional context eventually defined for the artefact. However, since the boning strips were encased between the card and linen interlining layers, they could not be examined directly without causing further damage to the artefact and without disturbing the pattern of particulate soiling. Radiography was therefore used to reveal the garment’s boned structure without compromising its physical and historical integrity.

Methodology General aims of project The primary aim of this project was, therefore, to attempt to identify and define the true nature of the concealed stomacher by exploring the multiple roles it had fulfilled, and might still fulfil, during its lifetime as a fashionable, functional, ritual, archaeological and study textile. A further aim was to attempt to reveal how it may have made the transition across boundaries, from one social group to another, from the physical and functional plane to a spiritual one. These objectives were achieved by providing a detailed physical profile of the stomacher as it survives today, through direct study of its appearance, structure, condition and soiling. Research was also undertaken into the practice of concealment and the way in which concealed artefacts have been studied, interpreted and conserved to date. This helped both to assess the significance of the stomacher as a deliberately concealed object and to provide a contextual framework for the project as a whole. Additional research into costume history further helped to place the stomacher within its correct functional, cultural and historical context, and involved study of the relevant literature, consultation with dress and costume historians, and examination of the materials, cut and construction of similar boned garments of contemporary date.

Specific objectives of X-ray examination The baleen components of the stomacher form the innermost layer of the garment and thus the very basis of the object’s structure. As such, their nature and configuration could offer potentially vital clues as to the garment’s original and subsequent roles. It would also be crucial to assess the condition of the baleen layer in order to gauge the general stability of

The low-energy (15 kV) radiographs were taken using a Hewlett Packard Faxitron and Agfa Structurix D4 radiographic film. Three different exposure settings were used (2, 4 and 8 minutes) in order to allow for the different radio-opacities of the stomacher’s various component layers – i.e. baleen, card and several different types of textile, as well as organic and inorganic soiling from building debris. The garment was placed directly onto the film, as any intervening material (e.g. film wrappers or tissue paper) would degrade the image. This was particularly important in the presence of so many textile layers, which would already be difficult to differentiate from one other. Thin aluminium foil was inserted between the X-ray tube and the object so as to filter out the lowest energy particles of the beam and so reduce film fogging from scattered radiation. Specific details regarding the processing and digitisation of Xray images have already been described elsewhere in the present work (see Chapter 4) and so will not be discussed here. Although the stomacher fitted into the X-ray chamber, it was too large for a single X-ray film. Two films were therefore taken for each exposure setting: an upper one across the chest and shoulder area and a lower one of the waist and base, using the central busk fragment as a common reference point. There was a slight disadvantage to this method as it necessitated additional handling of the object and, each time the latter was moved, the particulate soiling was inevitably disturbed to some degree.

Interpretation of the radiographic images Once developed, the X-ray films were examined on a light box using a 10 magnification loupe.

The role of X-radiography in the documentation and investigation 207

Figure 14.4 Composite radiographic image of stomacher showing boned structure; exposure 15 kV for 4 mins. (Radiography and digital manipulation by Sonia O’Connor, University of Bradford.)

In order to examine the overall garment structure more easily, the upper and lower films were also overlapped and ‘knitted together’ digitally to form a single, composite image2 (Figure 14.4). The individual boning strips are clearly visible, as is the attached busk fragment which shows up as a lighter, butterfly-shaped piece in the centre of the garment. The thick accretion of particulate soiling on the obverse face, thought to be residues of building material (probably mortar or plaster), shows up as an opaque white mass top centre with smaller deposits visible down the right side proper. The full extent of damage to the baleen layer is also evident. A transparent polyester overlay was traced directly from the 1:1 scale X-ray image to give a very accurate life-size configuration of the stomacher’s boned structure (Figure 14.5). The overlay was then superimposed onto the object itself so as to examine the

Figure 14.5 Scale diagram of baleen layer. (Drawing by Gabriella Barbieri.)

precise relationship of the baleen layer relative to the other textile layers, as well as to the patterns of soiling and damage. A great advantage of using the overlay as opposed to the radiographic films themselves was that the former could be handled as necessary, without the risk of scratching the films and so degrading the original image. A further advantage was that, being portable, the overlay could be more safely and conveniently taken to museums or study centres and superimposed directly onto other boned garments of the period for comparative purposes.

Materials and construction By studying the radiographic images in conjunction with direct visual examination of the garment, it was possible to deduce that each panel of the stomacher would originally have been fully boned, containing about thirty baleen strips of varying length. These

208

Case studies

(a)

(b)

Figure 14.6 Details of radiographic image, (a) peeled-back edge of baleen strip, (b) moth damage to the baleen. (Radiography by Sonia O’Connor, University of Bradford.)

would have moulded the wearer’s torso to give the smooth tapering line of an inverted cone – the fashionable shape of the eighteenth century (Buck, 1987). A fully boned construction of this type begins to suggest that the garment started its functional life as a pair of stays, not a stomacher – a theory which will be explored further below. A further constructional feature highlighted by the radiographs, and one which may be measured very accurately, is the steeply diagonal orientation of the baleen strips. This points towards an early to mid-eighteenth century date for the original garment. Late seventeenth century boning was almost vertical, giving a long, tubular shape, while the fashionable pigeon-chested look of the latter part of the eighteenth century required fan-shaped boning set at a shallower angle to the waistline (Hart and North, 1998). The boning is of good quality, with thin, slender strips measuring just 4 mm in width and less than 1 mm thick, inserted into regularly spaced and flawlessly stitched bone casings. The radiographic image clearly shows how the ends of the baleen strips were cut obliquely, with considerable precision, to ensure that they aligned perfectly along the centre front seam – strong evidence to support the theory that the object’s boned structure was professionally crafted and formed part of the construction of the original garment. Yet a further detail of the stomacher’s construction that is visible only from the radiographs is a thin strip of baleen which has split away from the

main boning strip and has been bent back on itself (Figure 14.6a). Such delamination could only have occurred as the ‘bones’ were pushed into the tightly stitched casings and is indisputable material evidence that the channels were stitched first and the bones inserted subsequently, rather than vice versa.

Patterns of use The object’s boned structure is the single feature which most distinctly characterises the original garment. The structure revealed by the radiographs offers compelling evidence to suggest that the garment originally formed part of a pair of stays, either the triangular front panel of back-lacing stays or the separate stomacher panel of front-lacing stays. Furthermore, the quality of materials and standard of craftsmanship present suggest that these were expensive, professionally made, fashionable stays and were therefore likely to have been bought and worn by a woman of considerable financial and social standing, possibly a member of the country gentry. Although fully boned stomachers are known (Bradfield, 1997), the numerous museum examples examined as part of the costume research undertaken for this project revealed that boned ones were the exception rather than the rule, with the vast majority of stomachers not being boned at all. Those that are boned tend to have only a dozen or so baleen strips down the centre of the garment.

The role of X-radiography in the documentation and investigation 209

Here again, the radiographs offer a valuable insight into establishing the object’s sartorial history: they show that approximately five baleen strips are missing from the right side proper of the garment. It also seems as though some of the original strips were removed from the left side proper but were then replaced with slightly thicker and wider strips. A narrower central panel of boning of this type matches the configuration found in other boned stomachers, suggesting that the original stays were at some stage remodelled as a stomacher – possibly by a member of the working class, given the materials used and the crude nature of the alterations. The subsequent reintroduction of two thicker strips down the left side edge may have been because women’s clothing traditionally fastened on the left. This configuration would have placed that side under repeated physical stress from fastening and unfastening, necessitating reinforcement or periodic repair – hence the thicker strips on that side.

Patterns of degradation The radiographic images also proved invaluable in revealing the nature and extent of damage to the stomacher’s boned structure, showing a loss of baleen across the entire upper chest area and from the midriff down into much of the base. It would appear that, in addition to heavy wear and soiling, insect attack may well have played a significant part in the stomacher’s overall physical degradation, since many of the boning strips exhibit gnarled and misshapen ends and small, semi-circular indentations along their edges, as though gnawed (Figure 14.6b). The type of material attacked (keratin), together with the insect debris found littered across both faces of the stomacher (silk cocoons and gritty frass), suggest that the insects responsible were probably clothes moth larvae (either Tineola bisselliella or Tinea pellionella (Pinniger and Winsor, 1998)). The sheer scale of insect damage to the baleen layer is simply not visible through a solely visual inspection of the object. In fact, by superimposing the transparent overlay over the garment, it is clear that the feeding sites extended well into the boning channels and underneath the other layers, indicating that the larvae were eating their way along the casings to reach the unexposed baleen. The radiographic images also reveal many more pupae cases than are visible on the surface layers (Figure 14.6b), the frass produced by the larvae showing up very

distinctly as tiny bright white specks, outlining the shape of the cocoons. The fact that these specks are opaque to the Xrays suggests that they contain appreciable amounts of particulate soiling (see the opaque white mass, top centre) which would, in turn, imply that the bulk of the insect damage took place after concealment and not during use or storage of the garment. If insect attack is indeed responsible for the damage to the baleen, it may assumed that, even with the heavy wear and crude alterations that had already taken place, the stomacher’s physical appearance and condition pre-concealment were drastically different from those post-concealment.

Conclusion It is clear from the above findings that the application of radiography techniques in this investigative study played a crucial role in gaining a clearer understanding of the true natures of the concealed stomacher. By helping to establish a detailed and accurate biographical profile of the various functional and cultural roles that the garment has fulfilled over time, the technique has enabled the conservator to make more sensitive and better informed decisions as to what may or may not be appropriate conservation treatments. Since many of the details revealed in the radiographic images, would not have been gleaned – or at least not in such detail – from a purely visual inspection of the object, it is not unreasonable to assume that quite different – and possibly erroneous – conclusions may otherwise have been reached regarding the garment’s object biographies. The radiographic images also highlight the single aspect of the stomacher’s condition which is possibly of greatest concern – its physical integrity. Given the fundamental importance of the boning to the garment’s structure, it is clear that the degree of loss suffered by the baleen layer has seriously undermined its overall structural stability – knowledge that will be of crucial importance when devising future conservation, handling, access and display strategies. As a result of the combination of documentary, investigative and analytical techniques employed in this study, it was decided that no remedial conservation could ethically be undertaken on the stomacher at that time. All object biographies were deemed to be of equal importance as each formed an integral part of the garment’s history. The primary focus of the preservation strategy was, therefore, to stabilise the object and prevent further deterioration until

210

Case studies

such time as its future institutional context may be more clearly defined. As technology improves and interpretation of the results obtained using these techniques becomes more sophisticated, there can be little doubt that radiography will become an ever more useful investigative tool in the study, not just of concealed garments, but of other complex, multi-layered textiles. For instance, in the future it may be possible to ‘deconstruct’ the various component layers of the stomacher by digitally identifying, differentiating and separating out the various layers captured in the image according to their respective densities, as is currently possible with three-dimensional computed tomography (CT) in medical imaging. Such a technique could play a significant role in constructing a surrogate or replica stomacher, which could be used to great advantage in increasing public access to these objects and enhancing their interpretation. Typically, when documentation is carried out at the outset of a conservation project, the object to be treated already has an institutional context and its role within that context has already been decided, the treatment implemented being dictated by that role. This is rarely the case with concealed garments which, as has been explained, tend to defy straightforward classification. The rationale underlying this project was therefore to see whether thorough, detailed documentation could help to define the true nature of concealed garments, identify their potential long-term roles within the wider heritage context, and thus play a much more active, even decisive, role in the general curatorial and conservation decision-making processes. Radiography proved an invaluable investigative and analytical tool in that documentation strategy.

Acknowledgements The documentation of the Nether Wallop stomacher was undertaken at the Textile Conservation Centre (TCC), University of Southampton, in 2002 under the supervision of Dinah Eastop, Senior Lecturer, and Paul Wyeth, Conservation Science Lecturer. The radiographs were taken, processed and digitised by Sonia O’Connor, University of Bradford. The stomacher forms part of a privately owned cache and the author would like to thank the owner for granting permission to publish the case history and for her generosity and enthusiasm in allowing her objects to be studied.

Notes 1.

2.

The stomacher forms part of a wider research project on Deliberately Concealed Garments led by Dinah Eastop and funded by the Arts & Humanities Research Council; see: www.concealedgarments.org and Eastop & Dew (2003). Scanning was undertaken by Sonia O’Connor, AHRC Research Centre for Textile Conservation & Textile Studies, Research Fellow in Conservation, University of Bradford, using an Agfa FS 50B industrial X-ray scanner.

Acronyms NATCC North American Textile Conservation Conference UKIC United Kingdom Institute for Conservation of Historic and Artistic Works V&A Victoria & Albert Museum

References Arnold, J. (1977 [1964]). Patterns of fashion 1: Englishwomen’s dresses and their construction c. 1660–1860 (revised edition), p. 3. Macmillan. Barbieri, G. (2003). Memoirs of an 18th century stomacher. A strategy for documenting the multiple object biographies of a once concealed garment. (Unpublished MA Dissertation. Textile Conservation Centre, University of Southampton.) Baumgarten, L. and Watson, J. (2000 [1999]). Costume Close-up: Clothing Construction and Pattern 1750–1790 (second printing). The Colonial Williamsburg Foundation/Quite Specific Media Group. Bradfield, N. (1997 [1968]). Costume in Detail:Women’s Dress 1730–1930 (third edition). Eric Dobby Publishing. Brooks, E. (2000). Watch your step. The National Trust Magazine, 91, 66–68. Buck, A. (1987 [1954]). Women’s Costume: The Eighteenth Century. The Gallery of English Costume, Picture Book Number 2 (reprint). Manchester City Art Gallery. Easton, T. (1995). Spiritual middens. In Encyclopaedia of Vernacular Architecture of the World (P. Oliver, ed.), 1, p. 568. Cambridge University Press. Eastop, D. (1998). Decision-making in conservation: determining the role of artefacts. In International Perspectives on Textile Conservation. Papers from the ICOM-CC Textiles Working Group Meetings, Amsterdam 13–14 October 1994 and Budapest 11–15 September 1995 (A. Tímár-Balázsy and D. Eastop, eds), pp. 43–46, Archetype. Eastop, D. (2000). Textiles as multiple and competing histories. In Textiles Revealed: Object Lessons in Historic

The role of X-radiography in the documentation and investigation 211 Textile and Costume Research (M. M. Brooks, ed.), pp. 17–28, Archetype. Eastop, D. and Dew, C. (2003). Secret agents: deliberately concealed garments as symbolic textiles. In NATCC Biannual Conference 2003: The Conservation of Flags and other Symbolic Textiles (J. Vuori, ed.), pp. 5–15, NATCC. Hart, A. and North, S. (1998). Historical Fashion in Detail: The Seventeenth and Eighteenth Centuries. V & A Publications. Kopytoff, I. (1986). The cultural biography of things: commoditization as process. In The Social Life of Things.

Commodities in Cultural Perspective (A. Appadurai, ed.), pp. 64–68, Cambridge University Press. Merrifield, R. (1987). The Archaeology of Ritual and Magic. Batsford. Pinniger, D. and Winsor, P. (1998.) Integrated Pest Management. Museums and Galleries Commission. Swann, J. (1996). Shoes concealed in buildings. Costume: The Journal of the Costume Society, 30, 56–69. UKIC (1990). Guidance for practice. In Members’ Handbook. UKIC.

15 Hidden Structures: the use of X-radiography in the Fashion Gallery at Snibston Discovery Park, Leicestershire Clare Bowyer

Introduction Snibston Discovery Park opened in 1992, on the site of the last working mine in Leicestershire which closed in 1986. It was established to tell the story of technology through working lives using Leicestershire museum collections combined with interactive exhibits. The emphasis is on educational fun for the family. In 1997 local government changes meant that Snibston remained under the ownership of Leicestershire County Council but many of the objects displayed on site were now owned by Leicester City. This was particularly so in the Textile Gallery and so plans were begun to redesign the display. The new £800 000 Fashion Gallery, jointly funded by the Heritage Lottery Fund, NEXT plc and Leicestershire County Council, opened on 28 May 2005.

The Fashion Gallery, Snibston Discovery Park Its remit was to use Leicestershire County Council’s costume collection in a thematic display to show visitors why we wear clothes, how clothes have transformed the shape of our bodies, the types of fibres used and how clothes are made and sold. The outfits are not shown in a chronological order; instead they are juxtaposed according to the use, shape, type of fabric or construction of the garment. The Fashion Gallery also has a temporary exhibition space. The interpretation of the collection was carefully planned. It is layered with different levels of text ranging from panels, case banners and individual object labels. 212

There is also computerised information relating to the fashion industry and, finally, a hands-on series of interactive exhibits. An important part of the interpretation is found in the area called the Fashion Theatre, which houses a variety of interactive exhibits. Some are based on simple creative ideas making simple woven structures and art and design tables while others explore construction of different fabrics or how a sewing machine works. The radiography interactive exhibit Hidden Structures, which is the subject of this case study, is part of the Fashion Theatre.

Hidden Structures Hidden Structures aims to show visitors the structure of relatively familiar clothes which is not normally visible. Unless they are lucky enough to handle historic costume, the average visitor will have little understanding of how clothing is constructed and just how much is hidden underneath the surface. The finished interactive shows that some items have fewer components than might be expected. The aim was to enable visitors to explore hidden structures themselves, to compare a variety of objects, and sometimes to surprise themselves with unexpected answers. The final solution had to be a low maintenance, low-tech display which would stand up to robust use by large numbers of predominately young visitors, often without supervision. Different approaches were considered before the final decision to use radiography was made. One option was to buy garments which could then have been partly dismantled for the purpose of display. This would have shown the complexities of

The use of X-radiography in the fashion gallery 213

construction but it would have remained a static display without any interaction. The Fashion Gallery does display a silk panel from a late eighteenth century silk waistcoat shown alongside a complete version. This demonstrates how the garment was embroidered, and how carefully the garment was made to ensure that the expensive silk was not wasted. It is useful for showing the layout of the individual pieces but does not show the structure of the assembled garment. Another option was to have a handling collection in the Gallery for all visitors to touch. This was rejected for several reasons: the collection did not have any suitable objects which could have been left in the Gallery; they would have needed constant attention by museum staff to ensure that the items were put back tidily (or even to ensure that they remained in the Gallery). The best way to understand each object would have been to have a member of staff explaining what to look and feel for. These options were all unsustainable and not necessarily suitable for the Snibston visiting experience. It was therefore decided that radiography would be the best way of showing internal construction in a quick, clear and familiar way for a general audience. As Leicestershire County Council’s Museum Service has no radiography equipment, the University Hospitals of Leicester National Health Service (NHS) Trust was approached. Once it had agreed to help, the design of the interactive display was begun. The final design has an ankle-height round base on which are mounted four clear acrylic boxes and four tall light boxes. These light boxes appear black until one waves a hand over a sensor and they light up to show the X-rays (Figure 15.1). Originally, the radiographic boxes were not going to be black and the radiographic plates were going to be visible all the time, but it is much more exciting to light up a black box and see the image suddenly. This has the incidental benefit of prolonging the useful life of the radiographic images. The images themselves are sandwiched between two layers of acrylic and then lit from behind in the metal light box. The design is accessible for wheelchair users, children and adults. Although plugged into floor sockets, the whole base is on wheels and can be moved around relatively easily. The clear acrylic cubes, measuring approximately 400 mm 400 mm 400 mm, were designed to protect the items of costume. This gave the maximum size of object for display and therefore limited the choice of garments that had an interesting construction. Ideally, the objects would have been selected and a system designed around them.

Figure 15.1 Hidden Structures interactive display in the Fashion Gallery, Snibston Discovery Park. (Photograph by Grace Deeks; © Leicestershire County Council.)

However, the size of the Fashion Theatre as a whole, the range of activities planned for the space and budget constraints dictated the size of the display cases. The symmetrical layout of the interactive display resulted in costs being lower than would have been the case had all the boxes been different shapes, tailor-made to each different object. From a design perspective, the symmetry also looks better than random shapes and sizes.

Selecting objects for radiographic display images Although large items of costume were rejected as they would never have fitted in the cases, Leicestershire County Council does not collect children’s dress, which might have supplied some items of the required size so this was not an option for the Gallery. Instead corsets and bodices, handbags, footwear, an eighteenth century stomacher, a crash helmet, and even a World War I lifesaving waistcoat were considered as these all had hidden construction elements. The late nineteenth century bodices had boning inside them which was not visible externally. As there was no guarantee which object from this selection would provide a suitable radiographic image,

214

Case studies

results which were interesting for a curator, they were deemed too difficult to be read by general visitors. The lines created by the whalebone stiffening were too shadowy and the three dimensionality of the corset was lost. The stomachers were also rejected because they would have needed an explanation of their use and space in the design was too limited for such detailed labels. It is still useful to have the images as a record for each item but it was necessary to weigh up the knowledge that most visitors would bring with them to the object and assess how they would respond to each item on display. Once the objects had been X-rayed the hospital was able to give us a disk for our designer, Ian Jones, to work on. He was able to tidy the images, extend the canvas and clone in extra background to fit the aperture of the interactive. The digitals were then reverse mounted onto acrylic. It was comforting to know that the hospital retained copies in its database should any problems arise and could have e-mailed us repeat copies if needed.

Figure 15.2 Radiograph of an 1890s corset containing pre-formed steels. (Radiography by University Hospitals of Leicester National Health Service Trust; © Leicestershire County Council.)

far more items were taken to the hospital for radiography than were necessary for the final exhibition. The staff were very helpful and interested as they had never undertaken such a request before. They managed the X-ray process from this point on. The X-rays were generated by a Philips Optimus machine. The images were captured digitally. They were transferred to an Agfa IMPAX PACS system for storage. All the radiographs were produced to a set standard for medical images. Existing supports, which were used for patients, were adapted to prop up the items during this process. Providing supports for the boots and shoes was relatively easy as it was for flat items such as the stomacher. Only the corsets proved difficult as they were initially supported by a foam core form. This made the image confusing to interpret as the foam core appeared in the radiograph so clear acrylic supports were substituted for this. These had no impact on the final image and maintained the corset shape perfectly. Metal, not surprisingly, gave one of the better contrasts. While the radiographs of the whalebone corsets and eighteenth century stomachers gave

Chosen objects and radiographs The final items chosen were a corset containing preformed steels, dating from the 1890s (Figure 15.2), a 1910 woman’s ankle boot (Figure 15.3), a man’s modern safety boot (Figure 15.4) and a motorcycle helmet (Figure 15.5). The contrast between the main material and the other components was one of the deciding factors when making the final selection for display. The corset is a striking image, virtually modern art in its own right: the fabric almost disappears leaving only the supports and fastenings showing, contrasting beautifully with the completely black original beside it in the interactive display. The 1910 boot shows the thin metal support down the centre of the insole, which had not been expected in a boot of this date, as well as the large number of nails holding the sole to the uppers. It also shows how the maker has bent the sharp ends over so that they will not stick into the wearer should the insole wear thin. This is in contrast to a 1960s stiletto X-ray, not on display, showing all the nails pointing straight up – and anecdotal evidence makes it clear that wearers did suffer from the nails piercing through. Visitors can compare the construction of this earlier boot with the safety boot. Although they both show stitching marks, the modern boot has no nails. The safety boot also has a layer of Neoprene inside which shows clearly and is described on the label.

The use of X-radiography in the fashion gallery 215

Figure 15.3 Radiograph of a woman’s 1910 ankle boot. (Radiography by University Hospitals of Leicester National Health Service Trust; © Leicestershire County Council.)

Figure 15.4 Radiograph of a man’s modern safety boot. (Radiography by University Hospitals of Leicester National Health Service Trust; © Leicestershire County Council.)

The display also intended to convey the message that safety does not always mean having metal as one of the safety components. The safety boot radiograph, for example, clearly shows the metal toecap hidden under the leather, as would be expected, whereas the motorcycle helmet looks terribly flimsy in comparison. Indeed, the only metal elements in the helmet are the fastenings for the neck strap, and

the padding around the head looks as insubstantial as tissue paper. Thanks to computer technology, the digital images can be manipulated easily and used in different ways; for example, the corset image is available as a postcard, a limited edition mouse mat, and many people have suggested its use as a poster. It would not have been possible to do this as easily prior to computerisation.

216

Case studies

Feedback and evaluation

Figure 15.5 Radiograph of a motorcycle helmet. (Radiography by University Hospitals of Leicester National Health Service Trust; © Leicestershire County Council.)

There is also the option to return to the unused radiographs to provide a change to the existing interactive display. The clear acrylic cases and light boxes are easily opened so that display items can be changed. In addition, the radiographs can easily be used in talks and during supervised visits to the Fashion Gallery or to the Collections Resources Centre which acts as the museum’s store. As the images are stored digitally, replacement copies can easily be made as the display copies become worn or faded.

What has been the response to Hidden Structures? The interactive tends to attract older children (seven years and upwards) as it looks sleek in silver and black in contrast to the surrounding multi-coloured interactives which the younger children favour. To many children it is just a game where you can use your hand to make the sensor work. This is not necessarily a bad thing as the aim was for the children to have fun. The objects displayed had the benefit that most, save for the corset, were modern, or had modern equivalents, with which visitors would be familiar. Even if the children do not compare one image with another they are surprised at what is inside the objects, particularly in the 1910 boot, and do start to think about their own shoes. It was not necessarily expected that children would automatically compare and contrast each image used but, with the help of teachers or parents, the interactive could help them explore further the ideas of visible and invisible construction. Adults react more to the corset and are amazed that women surrounded themselves in so much metal. It also challenges people’s preconceptions that only whalebone was used in corsets. Hidden Structures works from a curator’s point of view in that it displays original garments and reveals their construction without damaging them in any way. While its size was a constraining factor, it fits into the overall scheme of the Gallery. By using relatively simple technology it is robust and sustainable. Finally, it has an element of surprise and gives the visitor an experience they would not normally expect.

16 X-radiography of a knitted silk stocking with metal thread embroidery Sonia O’Connor, Mary M. Brooks and Josie Sheppard

Introduction This rare, although not unique, eighteenth century stocking from York Castle Museum, York Museums Trust (YORCM: 1175-76 362/41), was selected for radiography because of its intriguing combination of a knitted structure with metal thread embroidery and the interesting pattern of wear, holes and darns.

The stocking The stocking is machine-knitted in silk yarn. It is shaped to fit the leg and is the typical length of an eighteenth century stocking, finishing above the knee (Figures 16.1a and b). The main body of the stocking is salmon pink although this varies considerably, giving a rather blotchy effect overall. A decorative pattern or ‘clock’ knitted in a cream coloured yarn is located on the ankle area on both sides. The knitted stitches in these clocks run at right angles to those in the rest of the stocking. A group of small circles is visible at the top of the clocking but only glimpses of the rest of the clocking are visible between the elements of the metal thread embroidery motif. The foot of the stocking has been cut off, presumably because it became damaged as a result of wear. The bottom edge of the clocking has been stabilised with a coarse overstitching in a yellow thread. The welt (top) of the stocking has a number of green and cream stripes separated by bands of the main salmon pink colour and is finished at the top with a row in a red yarn. The stocking would have originally been worn by tying a ribbon or garter around this welt which

might have been folded over and stitched down; no stitched evidence for this remains (Farrell, 1992: 23). The lower section of the stocking has a great deal of metal thread embroidery, probably the reason why it has been preserved. The dominant motif is a large stylised plant form, possibly including a pineapple, in an elaborate vase (Figure 16.2a). This is worked on both sides of the stocking and is placed above the clocking although slightly offcentre. A small male figure, holding what might be a gun over his shoulder, is worked at the centre front of the stocking above a stylised floral motif and a small isolated scroll decorated with leaves (Figure 16.3). Two small squirrel-like animals, each wearing a chain, appear on either side of this figure. Three small diamonds are placed near the figure together with some detached circles. A scrolling line of stems, leaves and buds borders either side of the back seam, finishing in a stylised leaf motif. The majority of the metal thread embroidery is worked in a wrapped thread. The metal wrapping strip probably has a high silver content. Two embroidery techniques have been used. The centre of the plant motif is worked mainly in laid and couched work using a green couching thread in a decorative pattern. Some details have been worked through the fabric, using the metal thread as a passing thread. Looking at the inside of the stocking, an additional unevenly balanced plain weave two-coloured fabric can be seen beneath the plant motifs (Figure 16.2b). This forms a support for the heavy metal thread embroidery. The cream threads appear to have functioned as a grid for locating the metal thread work. The yellow stitches relate to the more raised areas of the motif. Neither of these threads is visible on the exterior of the stocking. 217

218

Case studies

(a)

(b)

Figure 16.1 Photograph, stocking, (a) Side 1, (b) Side 2. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

Condition At first sight, the stocking appears fairly stable. In the lower portion, any disruption of the knitted structure tends to be at the very edge of the metal thread embroidery. This is probably the result of stress due to

the junction of the flexible knitting with the very much stiffer embroidered motifs. Most of the holes are in the top unembroidered area. The location of darned holes compared with undarned holes is interesting. Logic would suggest that the undarned holes were formed later than the darned holes. There are four

X-radiography of a knitted silk stocking with metal thread embroidery 219

(a)

(b)

Figure 16.2 Photograph, detail of the metal thread embroidered plant motif, Side 1, (a) exterior showing decorative laid and couched green stitching; note the paler knitted clocking to the proper right side, (b) interior showing the green couching thread, cream stitches for the grid and yellow stitches relating to the raised areas of the motif. The plain weave support can also be seen. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

pairs of undarned holes spaced around the circumference of the stocking. Each pair is vertically aligned so one of the holes is in the welt and one is in the main body of the stocking. This distribution suggests that the welt may have been folded over and secured with suspenders, perhaps during reuse as fancy dress. A harder plied brown thread has been used as the darning thread. The upper part of the seam has been repaired with the same thread although the stitching is now coming apart in places. The metal thread has been rubbed in places, revealing its inner fibre core. As the stocking is embroidered on both sides, it could have been worn on either leg and it is hard to define an ‘outside leg’ and an ‘inside leg’. These were therefore designated Side 1 and Side 2.

Radiography

Figure 16.3 Photograph, detail from the front of the stocking showing embroidered figure of a man. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

Radiographs of the stocking were taken at different exposures in order to image the knitted fabric and the metal thread work most effectively. Agfa D4 film was used in a Faxitron 43855 single cabinet unit. The main problem when radiographing this stocking is that it now forms a flattened tube. X-raying from one side with the film beneath the flattened stocking would provide a confused image with information superimposed from both sides of the stocking. This problem was overcome when taking radiographs designed to image the knitted areas by inserting the radiographic film inside the stocking. The corners of the film were carefully rounded off to prevent

220

(a)

Case studies

(b)

Figure 16.4 Radiograph of the stocking taken to explore the bands in the knitted textile (15 kV, 2 mins, aluminium filter, c.0.6m FFD) Side 1, (a) welt and top of stocking, (b) lower area with metal thread embroidery. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

X-radiography of a knitted silk stocking with metal thread embroidery 221

snagging. An exposure was taken of one side before the stocking was turned and an exposure taken of the second side with new film (Figure 16.4a). To take the image of the back seam, the stocking was carefully padded out with soft balls of acid-free tissue which both supported the shape of the stocking and held the film in close contact with the seam. The stocking was placed with the seam uppermost because otherwise the tissue paper and the stocking itself would have overlain the image while the tissue would also have attenuated the beam. At these kilovoltages, even a single sheet of tissue paper will produce a detectable image on the film. Radiography of the metal thread embroidery presented a different challenge. Images taken at 60 kV with the film inserted into the stocking produced an image of the metal thread embroidery but it was difficult to discern the fine detail. Normally, when presented with such a situation, i.e. thin structures of a high atomic metal producing an image with big variations in optical density, the ideal radiographic technique would be to use lead screen intensification. This technique combines beam filtration and electron production to produce images with the penetration of a high kV beam (120 kV) coupled with the contrast normally associated with lower kV images (see Chapter 3, pp. 41-3). Unfortunately the film and lead screen intensifiers used for this are normally housed in an aluminium cassette which is too big to insert into the stocking. Instead, an image was taken with the cassette placed below the stocking.1 Despite the superimposition of the motifs, this provided useful information about the metal thread embroidery which could not be obtained through visual examination.

(a)

What the radiography revealed The yarn Although visually the knitting and the yarn seem very consistent, the radiograph shows something different. In the main body of the stocking, darker and lighter bands are visible in the radiograph which seems to relate to variations in the denier of the yarn (Figure 16.4a). In the welt, these bands are more marked and relate directly to the differently coloured stripes. The variations in radio-opacity between the coloured yarns may also reflect differences in denier but could be influenced by different dye and mordant combinations. Structure and construction Radiography enabled greater understanding of both the knitted structure and the construction of the stocking. Figure 16.4b highlights the junction between the main knitted structure and the clocking where the stitch direction changes. The radiograph also shows that the lighter coloured yarn used for the small circles at the top of the clock was knitted in with the main yarn. This technique prevents the formation of loops on the back of the knitting which could snag (Figures 16.5a and b). Joins between two yarns may be visible as a few thicker stitches when inspecting the stocking with the naked eye. This may be quite a subtle difference. However, these are highlighted on the radiograph, appearing as a short line of brighter knitted stitches side by side (Figures 16.5a and b). This can be compared

(b)

Figure 16.5 Detail, showing a join in the yarns and circles at the top of the clocking, (a) photograph, (b) radiograph. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

222

Case studies

(b)

(c)

(a)

Figure 16.6 Back seam of the stocking, (a) radiograph; the pale halo shows the extent of the support fabric below the metal thread embroidery, (b) photograph, detail of darn, (c) radiograph, detail of darn. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

X-radiography of a knitted silk stocking with metal thread embroidery 223

with the similar image of joined yarns on a modern example (Figure 9.18, p. 139). A very few knots joining yarns, hidden on the interior of the stocking, may be seen on the radiographs as irregular, denser and therefore brighter bobbles. The radiograph makes it possible to establish that most joins in this stocking were created by overlapping the end of one yarn with the start of another. Evidence from the radiograph of the seam and remaining stitching confirms that this is a knitted, not a cut, edge (see Chapter 9, pp. 139–140 and Figure 16.6a). The stocking was shaped by increasing/ decreasing stitches at this edge. Towards the top of this radiograph, misalignment of the bands of different radio-opacity in the knitting may be due to repairs in the seam. The radiographic evidence might suggest that the metal thread work overlaps the back seam of the stocking, indicating that the embroidery was carried out after the seam had been closed. However, the apparent overlaps are all in the upper part of the seam which has been damaged and crudely mended with the same plied brown thread seen in the darns. Lower down, the placing of the metal thread along the seam confirms that, as might be expected, the embroidery was carried out before the stocking was sewn up into a three-dimensional form. Metal thread embroidery The radiography highlighted the structure of the metal thread and the embroidery technique. Valuable information about the different thread types (wrapped and coiled ‘purl’) and the embroidery techniques, including the passing of threads from one area to another, became evident. The underlying support textile can be seen as a pale halo around the motifs and the decorative stitching of the back seam, confirming its presence in areas where it is difficult to inspect the inside of the stocking (Figure 16.6a). The full extent of a metal thread embroidered leaf, previously obscured by the darn, can be seen. The plain weave of the support fabric, not visible to the naked eye, can also be glimpsed below the crisscrossing threads of the darn (Figures 16.6b and c). The bright dot below the weave is a knot in the metal thread concealed inside the stocking. Condition The radiograph gives a strong indication of the fragility of the knitted structure. Worn areas, less visible to the naked eye, show up as darker, more open areas on the radiograph where the yarn has been abraded to a fraction of its original diameter, possibly

Figure 16.7 Detail of radiograph showing hole and areas of thinning. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

as a result of wear (Figure 16.7). The radiograph helps characterise the worn areas and the extent to which the yarn and structure may soon break down. Comparable information might be obtained for a knitted structure by photographing it backlit on a light box, or similar, but this is not possible in the case of a tubular artefact. This information can only be obtained by inserting film inside the stocking. Holes and runs Information about the holes and runs is also made available through the radiographs. The brightish edge seen intermittently around the holes is in some cases due to the overlapping of curling yarns of broken stitches. In others, it is the result of the presence of a darning thread (Figure 16.7). The radiograph of one of the holes located on the striped welt has two bright white spots (Figure 16.8). Their presence prompted a re-examination of the holes, enabling the identification of the dark redbrown staining so characteristic of iron corrosion on several of them. This suggests that metal has been in contact with the stocking. There are several possibilities here but one is that metal suspenders might have been used with the welt folded over if the stocking was worn as fancy dress. This is corroborated by the number and distribution of these holes noted earlier; evidence from the radiograph adds extra weight to this interpretation of the damage.

224

Case studies

here the stitches look brighter but foreshortened. In the other type, where the creases run across the image, the knitted stitches, although brighter, do not appear foreshortened. The edges of the creases have a straight edge rather than following the edges of the knitted stitches. These creases are in the woven support fabric below the knitting.

Summary The radiography of this stocking has not only given greater insights into its construction but has provided additional information about its condition and fragility. This enables the development of a better quality risk assessment and improved conservation strategies.

Figure 16.8 Detail of radiograph showing two bright white spots on the hole’s top left edge. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

Darns The plied brown thread is distinct from the main yarn when viewed in the radiographs. This makes the darns more readable and helps to show the tension they are generating in the knitted structure. It is also possible to see how much of the original knitted structure remains under some darns. This may help in the decision about whether a particular darn should be left in situ as part of the evidence of the use and history of the stocking or whether it should be removed because of the risk to the stocking’s long-term stability.

Acknowledgement Thanks are due to Josie Sheppard, Keeper of Textiles and Dress, York Castle Museum, York Museums Trust, for enabling the study of this stocking.

Note 1.

It is possible to purchase films of various sizes vacuum packed in plastic sleeves with integral lead screens. This would have been ideal for inserting into the stocking in order to radiograph the metal thread embroidery but was not in stock at the University of Bradford. The short shelf life of radiographic film makes it necessary to rationalise stocks to one or two frequently used formats.

Creases Two types of creases can be distinguished on the radiographs (Figure 16.6c). One type, running almost vertically near the edge of this image, is in the knitting;

Reference Farrell, J. (1992). Socks and Stockings. Batsford.

17 A chalice veil rediscovered Sonia O’Connor and Mary M. Brooks

Introduction The examination of this chalice veil became a story of surprises. The goal of the radiographic study was to understand the construction of the object and gain insights into its history. The identity and construction of this beautiful and complex embroidered silk textile was re-evaluated during the course of study and radiography played a critical part in enabling enhanced understanding of its biography. The textile was thought to be made up from parts of a Portuguese ecclesiastical vestment, possibly a sixteenth century chasuble remade into a chalice veil. It had been displayed in a private home in a glazed frame for many years. A chalice veil is one of the textiles used during the celebration of the Eucharist and is still used in many Roman Catholic and Anglican churches. It takes the form of a cloth, generally square, which is used in dressing a chalice for a communion service.1 The chalice veil often forms part of a matching set with a burse, a folding textile-covered case which is placed on top of the chalice. Chalice veils and burses may be made using fabrics which correspond to the different colours used for ecclesiastical textiles, including priests’ vestments and altar frontals, at various stages of the liturgical calendar. Saint Charles Borromeo (1538–1584) stated that chalice veils could be made of silk and decorated with silver and gold metal threads unlike other elements of the altar set which had to be made from linen (Mayer-Thurman, 1975: 27). The well-known ecclesiastical embroiderer Beryl Dean noted that the usual measurement for a chalice veil is a square of between 510 mm2 and 610 mm2 (1981: 98).

Description This chalice veil is a square measuring approximately 650 mm by 650 mm, excluding the machine-made

metal thread lace edging. It is made up of three layers of different fabrics, one of which has a small patch (Figures 17.1a and b). Visual examination established that the embroidered design is a composite, consisting of elements cut from an earlier textile. These consist of a central roundel with an embroidered and painted image of the Madonna and Child surrounded by a border of stylised flowers and leaves. Similar flowers and leaves form a deeper border around the edge of the veil (Figure 17.2a). The motifs are worked in multicoloured silk threads using satin stitch and laid and couched work, enhanced with laid and couched metal thread embroidery. The original plain weave silk ground fabric can be seen behind the gold thread work of the roundel and around the closely trimmed embroidered motifs, particularly where the embroidery stitches have become worn or lost (Figure 17.2b). The underdrawing can be seen here. This embroidered fabric is the layer which was initially thought to have been part of a chasuble or other liturgical textile. The embroidered motifs have been remounted onto a support fabric of cream ribbed silk with a moiré finish. This is made up of three pieces – one large oblong and two smaller pieces, a smaller oblong and a square, stitched together on the left side so as to form the requisite overall square shape. These handstitched seams can be seen with the naked eye (Figure 17.2a). The differing directions of the ribbing can easily be seen. In the larger square, these run vertically across the veil whereas the smaller pieces are placed so that the ribs run horizontally (Figure 17.2c). The backing is made up of four pieces of a lightweight plain weave pinkish silk (Figure 17.1b). Again, there is one large oblong which has been made into a square by attaching three smaller pieces, one long oblong and two smaller irregularly shaped oblongs. The main seams follow approximately the same alignment as the seams in the ribbed silk support 225

226

Case studies

(a)

(b)

Figure 17.1 Chalice veil, (a) photograph, front view, (b) photograph, back view. The arrow indicates the same corner of the front and back. (© Sonia O’Connor, University of Bradford.)

fabric. The long oblong has a small patch inserted in a different, but similarly coloured, silk.

Condition The veil is remarkably stable considering the multiple layers and elements. The carefully worked overstitching used to attach the embroidery motifs to the support fabric is generally stable (Figure 17.2b). There are some losses in this embroidery. Some of the stitching and pigment in the central roundel has suffered loss due to wear which might have included being fingered or kissed as the subject of devotion. Each layer carries in it indications of its separate history. The marked crease in the ribbed support fabric does not relate to creasing in the backing fabric so it is probably an echo of its former use. The pinkish backing has a pattern of differential colour changes, generally being darker in the embroidered areas. The repair patch in this backing layer was carried out when the veil was in its final form as the stitching goes through all the layers and may be seen on the front.

Evidence from radiography All the fabrics, seams, silk and metal thread embroidery are recorded by the radiography (Figure 17.3a). The image of the ribbed plain weave of the support fabric and the plain weave of the backing fabric

interfere visually; an almost moiré effect is created at places. This is characteristic of radiographs of two similar weaves superimposed on each other although it is the image of the ribbed weave which predominates. The seams in the original embroidered fabric become visible underneath the embroidery. These were otherwise hidden by the embroidery and very difficult to spot at the closely trimmed edges of the motifs although, on careful inspection, they can be seen running across the base of petals of the embroidered roses where the cream thread has been lost. Two of these seams can be seen in Figure 17.3a. One runs horizontally through a tulip in the top right while another is intermittently visible from the top centre of the image down through the rose in the lower centre. An enlargement of the vertical seam is shown in close-up running through another flower in Figure 17.3b. Following a full radiographic survey of the veil it was possible to map the position of all these seams in the embroidered elements. The placing of these seams led to the deduction that changed the thinking about the nature of the original function of the embroidered fabric. The location of the seams makes it clear that the embroidered motifs were always intended to form a square and the original function of the piece was as a chalice veil (Figure 17.4). Examining the shape and relationship of the motifs supports this revision of the veil’s history: it is a remounted chalice veil rather than a chalice veil made from fragments cut from a chasuble or other liturgical textiles. The materials and techniques of the

A chalice veil rediscovered 227

(a)

(c)

central motif are entirely consistent with it being part of the veil but, as no seams run through it, this cannot be substantiated. The seams in the support and backing fabrics stand out on these radiographs (Figure 17.3a) providing more evidence about the veil’s history. Their near alignment is evident; in this radiograph, the seam nearer the edge of the veil is the seam in the support fabric. The turnings of these seams are pressed together towards the edge of the veil. Close

(b)

Figure 17.2 Chalice veil front view, (a) lower left-hand corner with vertical seam, (b) embroidered motif showing cut edge of the original ground fabric and the overcast stitches attaching it to the ribbed support fabric, (c) detail of vertical seam showing the different alignments of the ribbed silk support fabric. (© Sonia O’Connor, University of Bradford.)

examination of these showed the turnings to have one layer with a frayed edge and one with a hard bright edge. Such hard bright edges indicate the presence of a selvedge rather than a cut edge. Plotting the position of the selvedges in all the seams made it possible to relate these pieces to the shape of the fabric from which they were cut and establishes the loom width of the support and backing fabrics to be approximately 500 mm or about 20 inches with the turnings. The smaller pieces are narrower

228

Case studies

(a)

(b)

sections cut across the width of the fabric and rotated at right angles before being stitched to the main piece. This reconstruction also links the creases visible in two places in the support fabric. The deep

Figure 17.3 Radiographs (15 kV, 2 mins, c.0.6 m FFD), (a) seams in all three fabrics, (b) detail of the seam in original fabric revealed under the embroidery. (© Sonia O’Connor, University of Bradford.)

crease across the lower section of the long oblong is part of the same crease seen in the larger oblong section. In combination with the evidence of the loom width, it is possible to deduce that the support

A chalice veil rediscovered 229

fabric was folded in half across its width before it was used in the chalice veil. The width of the original silk ground was narrower. It is approximately 445 mm (about 171⁄2 ) wide with turnings of approximately 5 mm (1⁄4 ) giving a total width of about 450 mm (about 18) and was made up to the full width required for the square veil by attaching strips to both selvedge edges. The seam turnings here are not folded to one side but are opened out flat.

Figure 17.4 The front of the chalice veil; the black lines indicate the position of the hidden seams of the original fabric revealed by radiography. (© Sonia O’Connor, University of Bradford.)

(a)

Radiography also provides information about the contrasting quality of the hand stitching of the seams. Both are sewn with a backstitch but the seam in the support fabric is sewn much more tightly whereas the stitching of the seam in the backing fabric is looser and the stitches more widely spaced. The overcast stitching securing the cut edges of the original ground of the embroidered motifs appears quite careful and controlled in the photograph (Figure 17.2b) but seems much less neat and controlled in the radiograph (Figure 17.3b). Here it is most evident as it passes in long, often diagonal lines, from one motif to another. Knots in the embroidery stitches can be seen at intervals as can the fibre core of the metal thread where the metal is missing (Figure 17.3a). The inserted patch is carefully stitched into place, even if the stitches do go through all the layers so they are visible on the front, and the weaves of the two different fabrics are not aligned (Figure 17.5a). The full extent of the oblong used as a patch and its diagonal alignment can be seen in the radiograph (Figure 17.5b). Also revealed in this radiograph (Figure 17.5b) was a totally unexpected element – a small fragment of plant material, caught between the layers, and presumably relating to the backing of the veil (Figure 17.6). Identifying a small fragment of a botanical specimen from its radiographic image when it cannot be accessed in any other way or viewed from other angles is very difficult. Had this been a uniquely identifiable fragment of a plant with distinctive native origins, this fragment might conceivably have added to the understanding of the

(b)

Figure 17.5 Detail of the back of the veil, (a) photograph showing the oval hole, (b) radiograph revealing the oblong shape of the patch and plant fragment, lower left. (© Sonia O’Connor, University of Bradford.)

230

Case studies

Conclusion Radiography of the chalice veil enabled greater understanding of the biography of this complex textile. More information became available about fabric widths, seam types and stitching quality and techniques. Most significantly, it became clear that this piece had always been intended to be a chalice veil.

Acknowledgement The authors would like to thank Mr James Sprigg, the owner of the textile, for generously agreeing to its study and radiography.

Notes 1. Figure 17.6 Radiograph of plant stem fragment trapped between the support and backing fabrics at the edge of the veil. (© Sonia O’Connor, University of Bradford.) 2.

Other textiles used in dressing the chalice may include a purificator, a smaller square of white linen used to wipe the cup after the celebration of communion, and a folded cloth known as a pall. The authors would like to thank Dr Allan Hall, Department of Archaeology, University of York, for this identification.

References biography of the veil. Sadly, this has been tentatively identified as part of a dock plant (Rumex sp.).2 Rumex species are common throughout Eurasia, so this identification does not help to provenance any event in the story of this textile.

Dean, B. (1981). Embroidery in Religion and Ceremonial. Batsford. Mayer-Thurman, C. C. (1975). Raiment for the Lord’s Service. A Thousand Years of Western Vestments. The Art Institute of Chicago.

18 The use of X-radiography in the analysis and conservation documentation of a set of seventeenth century hanging wall pockets Mary M. Brooks and Sonia O’Connor

Introduction

Materials and construction

This case study is an example of the early use of radiography as part of the conservation documentation and treatment of textiles. Following flood damage, the seventeenth century hanging wall pockets had become covered in fungi and radiography was undertaken in order to understand the nature of the concealed embroidery and plan the conservation intervention. The pockets were treated by Lynn Grant at the Textile Conservation Centre (TCC) in 1983 as a Final Year Student Treatment Project for the post-graduate Diploma in Textile Conservation (TCC 0499; Grant, 1983).

The four card-stiffened crimson silk pockets are stitched to a foundation of crimson silk, also on a card stiffening. The whole has an edging of metal thread lace. The pocket fronts and the panels between them are decorated with silver-gilt raised work metal thread embroidery and further ornamented with sequins. The figures are heavily padded, giving a pronounced sculptural effect; see detail of Faith (Figure 18.2).

The hanging wall pockets Unlike the hanging pockets worn by women, these were designed as decorative and functional objects to hang upon a wall. They are constructed with a number of small pockets to hold personal items such as letters or combs. These pockets belong to St John’s House Museum, Warwick. They probably originate from southern Italy and may have been made in a professional workshop. Each pocket, and the flat panels between them, is embroidered with the figure of a woman enclosed in a decorative frame (Figures 18.1a and b). Their attributes suggest that these figures may represent an unusual variation on the theological virtues of Faith, Hope and Charity; see Table 18.1.

Condition before treatment The pockets were structurally weak having been damaged by folding. There were areas of loss due to water damage. The fungal covering (Aspergillus flavus, A. repens, Penicillium extansum and P. frequetans) formed a fibrous mat which obscured much of the surface of the piece and made it difficult to read the design (Figure 18.1a). Fungal threads even extended into the tunnel holes produced by insect infestations, probably carpet beetles (Anthrenus verbasci or Anthrenus scrophulariae) and clothes moth larvea (Tineola bisselliella or Tinea pellionella).

Radiography Understanding the structure and condition of the piece beneath the obscuring deposits was essential in order to develop a conservation treatment strategy. Examination under ultraviolet light and infrared 231

232

Case studies

(a)

(b)

Figure 18.1 Hanging pockets, (a) photograph, front view before conservation showing the fungal covering, (b) photograph, front view after conservation. (Photograph © Textile Conservation Centre; reproduced by permission of St John’s Museum, Warwick.)

The use of X-radiography in the analysis and conservation documentation 233 Table 18.1 Iconography of the figures on the hanging pockets Top panel Pocket 1 Panel 1 Pocket 2

Charity Prudence Prudence fallen Chastity

Panel 2 Pocket 3 Panel 3

Chastity fallen Faith Faith fallen

Pocket 4

Hope

May be giving gifts to the putti accompanying her Clothed figure, holding a mirror and has a snake around her wrist Nude figure with hair cut off. She is no longer holding her mirror which lies beside her Clothed figure holding a palm leaf, the attribute of the Virgin Martyrs and accompanied by a small animal. This is possibly an ermine which is also an attribute of Chastity Nude figure with hair cut off Clothed figure holding a book and cross Reclining nude figure with hair cut off. She is no longer holding her book and the cross lies beside her Clothed figure leaning on an anchor and accompanied by a dove with an olive branch

Figure 18.2 Detail of radiograph showing the Faith pocket. (Radiograph © Textile Conservation Centre; digitisation and image processing by Sonia O’Connor, University of Bradford.)

light did not enable this. The fungi did not fluoresce under infrared light but remained white and opaque. Radiography was therefore undertaken.1

Information obtained from the radiography The radiographs were taken to highlight the metal elements of the embroidery (Figure 18.3). After digitisation and digital image processing the card, and junctions in the card, can just be seen but textile and

organic stitching threads are not visible. In terms of imagery, attributes such as the book and cross in the cartouche showing Faith could be read more easily. The radiograph (Figure 18.4) of the putto in the Charity panel illustrates the wide variety of different metal threads including purl, a ‘trim’ made from a helical wire with a beaten edge, and one type of sequin with a new clarity. For instance, these images emphasise the pairs of metal threads which are linked by a spiralling metal strip. This produces the effect of diamond hatching on the putto and in the flesh and drapery of Faith (Figures 18.4 and 18.5).

234

Case studies

Figure 18.3 Hanging pockets, (a) radiograph, (b) line drawing. (Photograph © Textile Conservation Centre; reproduced by kind permission of St John’s Museum, Warwick; line drawing by Lynn Grant.)

Information about the construction of some of the metal threads also becomes visible. For example, in the cloak of the putto (Figure 18.4), one of the metal threads becomes gradually brighter and then abruptly returns to the normal level of intensity. This is also visible on Figure 18.6a where it highlights the path of a thread as it turns at the edge of the motif. This suggests that these are joins in the metal wrapping of the fibre core but that the start and finish of the wrappings are different. Further mapping of such overlaps may

enable the start and finish of the metal wrapping to be differentiated. Variations in the size, shape and thickness of the sequins scattered all over the piece are evident. The off-centre hole and curved line running to the edge suggests that the round sequins were made from a spiralled wire, cut down one side to produce individual links which were then beaten flat. This is a similar technology to that which was used to make the off-centred looped metal wire ‘trim’.

The use of X-radiography in the analysis and conservation documentation 235

Figure 18.4 Detail of radiograph showing a putto from the Charity top panel. (Radiograph © Textile Conservation Centre; Digitisation and image processing by Sonia O’Connor, University of Bradford.)

Figure 18.5 Detail of radiograph of the Faith pocket. (Radiograph © Textile Conservation Centre; digitisation and image processing by Sonia O’Connor, University of Bradford.)

Embroidery techniques

turning and returning at the edges to form the desired shape. However, it is hard to achieve a smooth curve when laying down metal thread. The edges were therefore defined using a thicker plied thread. In some areas, a smooth curve has been achieved by overlaying the curved edge with another ‘paisley’ shaped motif (Figure 18.6a). In some places, cut ends of the metal threads forming the cartouche were concealed beneath the overlaying motif; these are now visible in damaged areas (Figure 18.6b). Thread ends can be seen secured beneath the parallel threads of the cartouche in Figure 18.6a and the proper right leg of the putto in Figure 18.4. This radiograph also shows that the couched metal threads forming the cloak overlay those forming the body. This suggests that being economical in the use of metal threads may have been less important than the efficient use of time.

The varied poses of the Virtues are created by skilful use of padding. The radiographs highlight these contoured areas where the threads are angled over the padding. The X-ray beam is more attenuated here because it passes through more than one thread producing a brighter edge to the feature (Figure 18.5). It is noticeable that the metal threads are not joined mid-way along a line of couching but are always terminated at an edge to avoid disruption of the smooth and continuous surface of the motif. It is difficult to see where the stitching holding the metal threads down actually is except where the threads are bent at a relatively sharp angle, as at the flexed limbs of the figures; for example, the putto in Figure 18.4. Here the effort of positioning the metal threads has left gaps between the paired threads. The cartouches were created by laying down the metal thread in a spiral around the inner foundation line of the oval. In the case of the pocket showing Faith, eight threads were couched down parallel to each other (Figure 18.6a). The scrolls and flourishes of the cartouche were then built up individually with another thread, passed backwards and forwards,

Damage and repair Areas of wear show up where the metal wrapping has been abraided, become broken and lost. This is particularly evident in raised areas (Figures 18.2 and 18.5). These radiographs also show a small area of repair to Faith on her upper left thigh.

236

Case studies

(a)

(b)

Figure 18.6 Detail of radiograph of the cartouche surrounding Faith, (a) edges, (b) cut ends of wrapped metal threads. (Radiograph © Textile Conservation Centre; digitisation and image processing by Sonia O’Connor, University of Bradford.)

Conclusion

Note

In this example, the radiography was helpful in enabling greater understanding of the metal thread embroidery. It contributes to the sense that this piece was made professionally with the emphasis more on efficient use of time than very fine and detailed work. Nonetheless, the pockets were created with great skill and understanding of the materials. This would have been an expensive and high-status artefact. This study has also provided a glimpse of the huge potential for understanding metal thread technology and embroidery techniques through radiography.

1.

Acknowledgements Thanks are due to Lynn Grant and her supervisor for their pioneering use of radiography and St John’s Museum, Warwick.

The radiographs were taken at the Technology Department, Courtauld Institute of Art, and digitised and prepared for publication by Sonia O’Connor.

Reference Grant, L. (1983). The Conservation of a Set of Seventeenth Century Hanging Pockets TCC 0499. (Unpublished Diploma report, Textile Conservation Centre.)

19 ‘In needle works there doth great knowledge rest’: the contribution of X-radiography to the understanding of seventeenth century English embroideries from the Ashmolean Museum of Art and Archaeology, Oxford Mary M. Brooks and Sonia O’Connor

Introduction The large numbers of surviving seventeenth century English embroidered panels, boxes and mirrors are ample evidence of a very specific and feminised material culture. Affluent schoolgirls and women evidently dedicated – or were made to dedicate – considerable time and energy, not to mention money, to producing these embroidered artefacts. Preachers, educationalists and moralists stressed the links between needlework and writing skills with the development of appropriate religious and social behaviour. John Taylor’s poem The Praise of the Needle in James Boler’s The Needles Excellency (1631) makes an explicit statement about the moral benefit of ‘work’, that is embroidery. Profit here bears a spiritual meaning: So Maids may (from their Mistresse, or their Mother) Learne to leave one worke, and to learne another… Untill, in time, delightful practice shall (With profit) make them perfect in them all. Thus hoping that these workes may have this guide To serve for ornament, and not for pride: To cherish vertue, banish idlenesse… (Epstein, 1995: 24) Embroidery was therefore a means by which women could demonstrate not just genteel skills but also virtue. William Barley, in his A Book of Curious and Strange Inventions, called the First Part of Needleworkes,

actually a translation of Giovanni Ciotti’s collection of patterns, argued that embroidery was the feminine equivalent to the male study of theology, alchemy or the liberal arts: In needle works there doth great knowledge rest. A fine conceit thereby full soone is showne: A drowsie braine this skill cannot digest, Paines spent on such, in vaine awaie is throne: They must be carefull, diligent and wise, In needle workes that beare away the prise. Radiography enables a greater understanding of these complex embroideries and the feminine knowledge and ‘paines’ which created them. This paper focuses on seventeenth century examples in the collection of the Ashmolean Museum of Art and Archaeology, Oxford. Typically, these embroideries are made using an extraordinary range of organic and inorganic threads and decorative elements in techniques ranging from flat stitching to elaborate padded sculptural elements. Their possible social and religious significance has been discussed elsewhere (Brooks, 2004).

Radiography techniques These embroideries were radiographed at the Ashmolean Museum on their radiography equipment, a Victor X-Ray unit.1 This long-established facility is used for a wide range of artefacts from the Museum’s collection in materials as diverse as stone, 237

238

Case studies

metal and mummies but was not designed for lowenergy radiography. This X-ray unit has filters which cannot be removed, so truncating the lower end of the beam spectrum most useful for imaging textiles. However, the results obtained demonstrate that it is possible to achieve useful images with such equipment. Kodak MX125 film enclosed in flexible plastic cassettes was used as this film stock was not prepacked in lightproof envelopes. Exposures were taken at the lowest possible energy, about 30–35 kV. The resulting radiographs were necessarily rather low in contrast. After digitising using the dedicated X-ray film scanner at the University of Bradford, the contrast was adjusted to facilitate image interpretation.2

The contribution of radiography to understanding materials, condition and construction The following discussion of the evidence provided by radiographs of selected embroideries aims to show how radiography adds to the understanding of their materials, embroidery techniques and condition. It is worth noting that the mounting of these fragile pieces on conservation grade fabrics over acid-free card or Correx3 supports enables them to be handled safely for research and display but also present a challenge for radiography. Images of such boards and support fabrics are superimposed on those of the embroidery, further reducing contrast and degrading the detail. The Sacrifice of Isaac (WA OA.414) In this version of Abraham’s sacrifice of his son Isaac, the central biblical scene is framed within an elaborate raised cartouche surrounded by large flowers and, to the left and right, exotic birds perched in trees (Brooks, 2004: 36–39; Figure 19.1a). Both flat and raised embroidery techniques have been used, employing silk, metal threads, pearls and beads. Detached needlepoint sections were used for many of the most three-dimensional, almost free-standing, elements including the birds and the figures’ heads. Small pearls have been used to work the date, which may be 167(3) and the initials I and E or Y. This embroidery is adhered to a wooden panel. As expected, radiography provided information about the metal threads and the supporting wires and pins used in the raised work embroidery (Figure 19.1b; Brooks and O’Connor, 2004a: 172–173;

Brooks and O’Connor, 2004b). There are two thicknesses of wires. The finer wire has ripples in it. This may be due to the wire having been stored tightly coiled and then straightened or it could be a sign of reuse. Alternatively, the shaping of the wire may have been deliberate to help anchor the detached needlepoint stitches during construction or, perhaps, is a deformation resulting from tension in the needlepoint (Figures 19.2a and b). Some wires have halos suggesting surface corrosion. Breaks are evident in some wires, indicating brittleness. The different types of metal threads can be easily distinguished (see Chapter 10, p. 154, Figure 10.6, p. 155). The glass beads show up clearly, including a previously unsuspected bead hidden from view on the concealed side of the right-hand bird’s head (Figures 19.2a and b). The left-hand bird also has two eyes. In contrast, the wire in the leopard’s head passes through to the far side of the head but there is no bead for the second eye. The radiograph also reveals there is surface corrosion on this wire. Tracing the drilling holes in the pearls used for the date provided corroborative evidence for the identification of the final number as ‘3’. An unexpected revelation was the discovery that both birds have beaks made from those of real birds. One beak is long and slender. One is shorter and wider and may come from a finch or linnet (Figures 19.2c, d and e).4 The keratinous sheath of this beak and the underlying bone are visible against the image of the wood grain of the backboard. The presence of these beaks immediately raises the question of whether this is typical or atypical and, if the former, how such beaks were obtained and prepared for embroidery purposes. Further research is required here. The wooden support is made of two planks which appear to be radially cut as the growth rings, alternating rings of more and less dense wood, appear as fine lines running along their length. This provides a potential for dating by dendrochronology. They are approximately 180 mm wide with a horizontal join running across the centre of the embroidery, and a third narrower board, starting approximately at the level of the lion and leopard. The narrow upper fillet is a later addition, as is the modern screw (Figure 19.1b). The middle joint is opening up on the righthand side underneath the bird. Two sorts of adhesive can be identified in the radiograph. One, which was used to stick the boards together, has ‘dribbled’ slightly over the face of the board in some places. The other is the adhesive used to attach the textile to the board. The cracquelure in this adhesive can be seen in this radiograph when magnified (Figure 19.2d).

The contribution of X-radiography to the understanding of seventeenth century English embroideries 239

(a)

(b)

Figure 19.1 The Sacrifice of Isaac, (a) photograph, (b) composite radiograph. (WA OA.414; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

240

Case studies

(a)

(c)

(b)

(d)

(e)

Figure 19.2 Embroidered bird with pears, right-hand, detail The Sacrifice of Isaac, (a) photograph, (b) radiograph, (c) photograph detail, left-hand bird’s head, (d) radiograph detail, left-hand bird’s head (WA OA.414; reproduced by permission of Ashmolean Museum, Oxford; digital manipulation and photography © Sonia O’Connor, University of Bradford), (e) skull of a finch for comparative purposes. (© Experimental Zoology Group Wageningen University.)

Pastoral Scene (WA 1954.90) A shepherdess with her sheep is shown seated next to a shepherd piping a tune to his dancing dog (Figure 19.3a; Brooks, 2004: 64–65). This bucolic couple is enclosed within an elaborate oval cartouche which is further surrounded by exotic birds, flowers, a lion and a leopard and a dog chasing a hare. It is worked in raised and flat embroidery stitches on a silk satin ground fabric; the large flowers are ‘slips’ worked separately and then applied. This piece has a double layer of Correx as a support

board. The walls of the channelling in the Correx have produced a grid of bright lines across the radiograph which can be visually very distracting. However, because the lines recurred at regular intervals it proved possible to remove them digitally from the image using Fourier transforms (see Chapter 4, p. 71, Figure 4.9, p. 73). The weave evident in the radiograph is the support fabric used in the conservation treatment. The radiograph provides additional clues about the construction technique of the three-dimensional

The contribution of X-radiography to the understanding of seventeenth century English embroideries 241

(a)

(b)

(c)

Figure 19.3 Pastoral Scene, (a) photograph, (b) photograph detail, left-hand side of the cartouche, (c) radiograph of the same detail. (WA 1954.90; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

242

Case studies

(a)

Figure 19.4 Solomon and the Queen of Sheba detail of the Queen and her maid, (a) photograph, (b) radiograph detail. (WA 1994.142; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

cartouche (Figures 19.3b and c). The foundation, otherwise invisible, appears to be a bundle of fibres, sewn down with a thread which passes over them and then through the original supporting fabric in a helical fashion. The overlying laid and couched stitches in silk thread are not evident on the radiograph although the black thread used for the outlining stitches is visible. This highlights the importance of interpreting what can be seen on the radiograph with what is evident on the object in order to interpret data correctly. Contrary to expectations, neither the decayed black satin inlay fabric nor the replacement plain weave at the base of the fleur-de-lys on the cartouche is visible on the radiograph. However, as on the other embroideries, the black silk outlining thread can be seen. Presumably, this has implications for the types of mordants and dyes present. Different types of infilling stitches show up with different levels of intensity; the French knots show up more strongly than the long and short stitches, reflecting their relative thickness.

Solomon and the Queen of Sheba (WA 1994.142) This crowded embroidery is filled with trees, birds, animals and plants around the central figures of Solomon greeting the Queen of Sheba and her maid (Figure 19.4a; Brooks, 2004: 50–51). A mermaid with her mirror occupies a pond in the foreground. Even though, like the other radiographs, this image was very lacking in contrast it was possible to use digital adjustment to produce images that were very revealing and enable effective mapping of different features. It is possible to distinguish areas of the canvas which have been embroidered and those, under the appliqué satin motifs, which have not been embroidered. The faces of the Queen and her maid, the bird, metal thread work, distribution of the pearls and the outline stitching of motifs such as the squirrel are all visible (Figure 19.4b). Different colours of embroidery threads in the shading of the striped dress fabrics and the floral motifs on Sheba’s skirts can be distinguished. Different embroidery stitches,

The contribution of X-radiography to the understanding of seventeenth century English embroideries 243

(b)

Figure 19.4 (Continued )

such as in the rays of the sun worked in braid stitch and laid and couched work and the thread passing from one ray to the next, can be seen. Details of the underlying embroidery beneath the lace collars are also visible. The radiograph made it possible to explore the construction of the faces. These are made of appliqué satin fabric stitched down over bundles of threads which form the three-dimensional features. The radiographs show the threads crisscrossing to form the forehead, circular bundles forming the cheeks and tightly compressed bundles of threads for noses. The dark threads used for

the eyes show up but other sewing threads do not register. The mermaid’s face is made up in a similar manner. Her breasts and stomach are formed from stitched down circles of fabric under the satin fabric appliqué (Figures 19.5a and b). The arms are also worked over a supporting material. The patch on the reverse with the inked image of a mermaid does not appear on the radiograph suggesting that the ink does not contain metal. Not all appliqué motifs are padded. The radiograph of the mermaid in The Proclamation of the Young Solomon (WA 1947.191.

244

Case studies

(a)

(b)

(c)

(d)

Figure 19.5 (a) Photograph, mermaid, Solomon and the Queen of Sheba, (b) radiograph, mermaid, Solomon and the Queen of Sheba, (c) photograph, mermaid, The Proclamation of the Young Solomon, (d) radiograph, mermaid, The Proclamation of the Young Solomon. (WA 1994.142 and WA 1947.191.313; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

The contribution of X-radiography to the understanding of seventeenth century English embroideries 245

(a)

(b)

313) shows that there is no underlying fabric or thread filling (Figures 19.5c and d). Charity (WA 1975.13) The allegorical figure of Charity with three children is shown in a landscape surrounded by a border filled with plants and animals (Brooks, 2004: 62–63). The

Figure 19.6 Leopard, Charity, (a) photograph, (b) radiograph. (WA 1975.13; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

radiograph provides relatively little additional information. The faint outlines of the black thread used for the snail can be seen. More information can be obtained about the leopard (Figures 19.6a and b). Here, the radiograph provides evidence that padding together with a regular pattern of discrete groups of three ‘dots’. Looking at the evidence from the embroidered leopard, disruption of the long and

246

(a)

(c)

Case studies

(b)

Figure 19.7 Nutmeg tape measure, (a) photograph, (b) radiograph. (WA 1947.191. 325; reproduced by permission of the Ashmolean Museum, Oxford; digital manipulation and photograph © Sonia O’Connor, University of Bradford.)

The contribution of X-radiography to the understanding of seventeenth century English embroideries 247

short stitching suggests that the embroidery thread used to work these spots has been lost. The claws are worked in black thread and the radiograph shows how the thread is worked from one area to another. However, this is not the case for the leopard’s spots. First, there is no passing thread between the groups on the radiographic image and, second, the spots are much darker than the image of the threads in the claws and have much sharper outlines. This contrasts with the diffuse outline of the embroidery threads. Are these spots the radio-opaque ink underdrawing? Looking at other areas where underdrawing is known to exist, these cannot be seen on the radiograph suggesting that the leopard’s spots have been made using a different medium. Nutmeg tape measure (WA 1947.191. 325) This tape measure consists of an unmarked parchment tape within a nutmeg (Figure 19.7a; Brooks, 2004: 78–81). At the kV used for this exposure, the textile element is only visible on the radiograph as faint images where the threads at the end of the tassels have been tightly knotted (Figure 19.7b). The metal thread bobbles and pear-shaped structures appear to be entirely empty. This does not mean that they do not have a fabric core, which would be difficult to see through the interstices of the structures. It is, however, clear that there is no underlying core of material such as wood or bone. The structure of the nutmeg can be viewed in the gap between the upper and lower metal thread work. A crack and an area of loss can be seen on the left. There are dark, vertical lines on both sides, just in from both edges, which delineate the gap between the edge of the hollowed nut and the spooled tape. The dark and light banding, parallel to these lines, particularly visible on the left, is formed by the layers of parchment rolled around the central pair of wires.

There is huge potential to advance understanding of these complex pieces through radiography particularly when this is coupled with detailed observation and further supported by microscopy and material analysis. It is worth noting that it is much easier to achieve this radiographic evidence before such embroideries are mounted. This avoids problems with loss of contrast due to the presence of mount boards and enables immediate access, with appropriate care, to the backs of the embroideries. This radiographic experiment thus validates the move towards more minimal conservation treatments. Radiography can be an excellent tool for monitoring changes – such as the loss of pearls, new corrosion on hidden pins or cracks in wooden boards. There is also potential for exploring seventeenth century attitudes to the natural world, not just in the depictions of animals and plants in the embroideries, but through their literal incorporation into the pieces in the form of beaks and feathers. Radiography thus reveals the ‘fine conceits’ which informed the development of these embroideries.

Acknowledgements Thanks are due to colleagues at the Ashmolean Museum of Art and Archaeology, especially Dr Catherine Whistler, Curator, Department of Western Art, Mark Norman, Head of Conservation and Susan Stanton, Textile Conservator.

Notes 1. 2. 3.

Radiography was undertaken by Mark Norman, Head of Conservation, Ashmolean Museum. Digital image processing was undertaken by Sonia O’Connor, University of Bradford, using PaintShop Pro, a graphic software package produced by JASC. Correx is an extruded polypropylene rigid corrugated board. Thanks to Professor Terry O’Connor, Department of Archaeology, University of York, and Jan Jansen, Experimental Zoology Group, University of Wageningen; http://www.skullsite.com/

Summary

4.

The Sacrifice of Isaac provided a mass of detail about different types of materials and techniques but each of these pieces added further elements to the understanding of seventeenth century embroidery as well as challenges in taking and interpreting radiographs. These radiographs were taken to address specific questions that had arisen from previous research. Some of these questions could be answered while others will have to wait for the opportunity for further radiography in more ideal conditions.

References Barley, W. (1596). A Book of Curious and Strange Inventions, called the First Part of Needleworkes. Brooks, M. M. (2004). English Embroideries of the Sixteenth and Seventeenth Centuries in the Collection of the

248

Case studies

Ashmolean Museum, Oxford. Ashmolean Museum/ Jonathan Horne Publications. Brooks, M. M. and O’Connor, S. A. (2004a). New insights into textiles. The potential of X-radiography as an investigative technique. In Scientific Analysis of Ancient & Historic Textiles, Informing Preservation, Display and Interpretation. Post-prints of the AHRB Research Centre for Textile Conservation & Textile Studies, 13–15

July 2004 (R. Janaway and P. Wyeth, eds), pp. 168–176, Archetype Press. Brooks, M. M. and O’Connor, S. A. (2004b). Revealing the hidden: the X-radiography of textiles. National Preservation Office e-Journal, 2. Epstein, K. (1995). The Praise of the Needle. John Taylor’s Poem of 1631 with Annotations by Kathleen Epstein. Curious Works Press.

20 X-radiography of dolls and toys Mary M. Brooks, Sonia O’Connor and Josie Sheppard

Introduction Dolls and toys are invested with memories for their owners.When people give their favourite doll or teddy bear, sometimes a family heirloom, to a museum they become public objects, serving to remind visitors of the comfort of a silent friend in their own childhood. Some are even icons, such as Winnie the Pooh and Piglet, immortalised by A. A. Milne, and now in the collection of the New York Public Library (Central Children’s Room, Donnell Library Center). Museums collect these emotive artefacts both to record important aspects of personal and social histories and also to understand their manufacture and materials. However, they can be difficult to understand as their inner workings are concealed and the mechanisms of moving limbs and eyes may have been broken or lost. This chapter links curatorial and conservation perspectives to explore the contribution that radiography can make to the understanding of dolls and toys. It does not attempt a complete survey but aims to indicate how radiography can provide information that may aid dating and inform conservation decisions. It also highlights the challenge of radiographing complex threedimensional, mixed-media artefacts. In 1989 a group of dolls and toys from the York Castle Museum collection were selected for radiography with the goal of addressing the question ‘what information can radiography show which cannot be obtained through other means?’ Nine of these are illustrated in Figures 20.1a–i; see Table 20.1.These letters will be used to enable the dolls to be easily identified.

Materials and manufacture of European dolls: a brief overview Modern dolls come in a bewildering range of varieties, designed to appeal to children through the

medium of current fads and fashions. The same is true of historic dolls with one significant difference: today’s mass-produced dolls are almost always made of the same material – flesh-coloured, moulded plastic. Their predecessors were made from substances as disparate as wax, china and wood as well as man-made materials. Doll makers have always sought to develop new techniques, both to improve dolls’ life-like appearance and to enhance their appeal through sound and movement. Thus even a small collection of dolls may include many different materials and mechanisms, some of which are easily identifiable, others less so. Until the nineteenth century, dolls were usually made of wood, which was cheap, hardwearing and readily available. Limbs were often crudely attached and simple peg-joints were used to give movement. At the most basic level, features were crudely painted. Dolls like this could be bought to be dressed at home. More sophisticated dolls had inset glass eyes, layers of paint and varnish, intricate wigs and fashionable clothes. The simple wooden doll remained popular in the early nineteenth century but was gradually supplanted as manufacturers produced dolls with more realistic features. Wax was ideal for this. Two techniques were used to produce dolls’ heads and, sometimes, limbs.‘Poured heads’ were made by pouring wax into moulds, tipping out the excess once the outer shell had set and reinforcing the interior with a plaster composition.This gave a beautiful life-like effect but was very expensive.‘Dipped heads’ were cheaper and were made by dipping a moulded composition head into wax.This method also gave a life-like, translucent finish but neither process could overcome the inherent fragility of wax, so many surviving wax dolls are in poor condition.‘Composition’ is the term used for a variety of easily mouldable materials including papier-mâché or mixtures of glue with wood pulp, plaster, rags, scraps of leather 249

250

Case studies

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

Figure 20.1a–i A Armand Marseille ‘My Dream Baby’ doll (YORCM: BA 3215) B Armand Marseille girl doll (YORCM: BA310) C Shut-eye doll (YORCM: BA4844) D Double-faced jester doll (YORCM: BA3317)

E F G H I

‘Walking’ doll (YORCM: BA4855) Pedlar doll (YORCM: BA4843) Fixed-eye baby doll (YORCM: BA4845) Caribbean doll (YORCM: BA179) Jockey doll (YORCM: 31.53)

(© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

X-radiography of dolls and toys 251 Table 20.1 Radiography of dolls from the York Castle Museum collection Descriptions and date

Museum number YORCM

Makers, materials and mechanisms

Radiographic exposures to reveal organic components (textile, hair, stuffings)

Additional radiographic exposures

Figures

A

‘My Dream Baby’ 1925 onwards

BA3215

20–30 kV

–

20.1a 20.2

B

Girl doll 1890–1910

BA310

20–30 kV

–

20.1b 20.3

C

Shut-eye doll 1860s–1870s

BA4844

20–30 kV

–

20.1c 20.4

D

Double-faced jester doll 1870–1890

BA3317

20–30 kV

–

20.1d 20.5

E

‘Walking’ doll post 1862

BA4855

Armand Marseille bisque head, shut-eye, squeaker Armand Marseille bisque head, shut-eye, cup and ball strung joints Wax/composition head, composition limbs, straw filled body, shuteye Motschmann-type joints, squeaker Solid wax head, Motschmann-type joints, turning head, squeaker Wax/composition head, metal ‘walking’ mechanism

20–30 kV

20.1e 20.6

F

Pedlar doll 1830–1840

BA4843

20–30 kV

120 kV with lead screens to image metal structures –

G

Fixed-eye baby doll c. 1840

BA4845

20–30 kV

–

20.1g 20.8

H

Caribbean doll 1890–1920

BA179

20–30 kV

20.1h 20.9

I

Jockey doll, part of pull-along toy 1810–1900

31.53

40 kV to image the body –

Composition/wax(?) head, wood body, dowelled joints Wax/composition head and torso sections, Motschmann-type joints, squeaker Nut head, wood body Bisque-coated head, musical box in base

and ground-up bone. Composition mixtures were used alone to produce heads and limbs and later, in the nineteenth century, dolls’ torsos. It was ideal for mass production as well as being quite hardwearing. With a suitable painted finish, it produced an acceptably realistic doll which was cheaper than its wax counterparts. Ceramic heads and limbs were increasingly popular from the middle of the nineteenth century.Again, the manufacturing process lent itself to mass production. Ceramic factories could make dolls’ heads alongside their other wares without the cost of setting up a new operation. German makers were the most prolific and exported thousands of their products to Europe and America. At first, glazed china

20–30 kV

20.1f 20.7

20.1i 20.10

was used but by the 1870s bisque – unglazed porcelain – had taken over.This remained the most popular substance for dolls’ heads until it was replaced by the development of reliable plastics in the early twentieth century. Bisque has a matt surface and, when suitably coloured, has a delicate, life-like finish. This can be seen on Armand Marseille bisqueheaded dolls such as Doll A (Figures 20.1a and 20.2a) and Doll B (Figures 20.1b and 20.3a). Doll production was a piecemeal business since the makers of heads and limbs – which were the most important elements – were not usually responsible for the bodies.These were also mass produced but as a separate operation so the finished doll would be a combination of parts and materials from many

252

(a)

Case studies

(b)

(d)

(c)

(e)

Figure 20.2 Doll A, Armand Marseille ‘My Dream Baby’, (a) photograph of head, (b) radiograph, head front view, (c) radiograph, head side view, (d) radiograph of hand, (e) radiograph, squeaker (see Figure 20.1a). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

sources. Doll A is a good example of this (Figures 20.1a and 20.2). The bisque head was produced by Armand Marseille, one of the largest German manufacturers, specifically for the Arranbee Doll Company of New York who made the soft bodies. First sold in 1925 under the name ‘My Dream Baby’, this doll became a best-seller. As with the heads, bodies came in a range of materials. Until the 1880s, stuffed bodies of cloth or kid predominated. Stuffing materials might include sawdust, bran, ‘straw’, kapok, wood-wool and cotton. From the 1880s, composition torsos were more usual although these would often include wooden components in the joints; Doll B is a good example of this type (Figures 20.1b and 20.3). Doll makers went to great lengths to develop bodies that were life-like in posture. Although the results are not often aesthetically pleasing to modern eyes, it should be remembered that realism did not extend to the appearance of the undressed torso as these dolls were meant to be seen clothed. Many patents were taken out for different jointing systems. Charles Motschmann developed a type of ‘floating’

joint for baby dolls where limbs were jointed by means of fabric tubes and he patented this in 1857. This method had been tried before but the tubes and the limbs they were meant to connect usually parted company. Motschmann solved this problem by securing the fabric into grooves in the wooden ends of lower limbs. From the 1880s onwards, strung ball and socket joints were those most commonly used; Doll C is a good example here (Figures 20.1c and 20.4). The search for ever-more realistic dolls led naturally to the development of movement and sound. The simplest moving mechanisms were those that controlled a doll’s eyes. These were certainly in use for mass-produced dolls by the middle of the nineteenth century. At first, a wire or rod protruding from the torso was used to open and shut glass eyes. This was replaced by shut-eye systems where a bar connecting the eyes was fixed to a weight inside the head which would move when the doll was laid down, thus closing the eyes. The process would be reversed when the doll was moved into a sitting or standing pose. Shut-eyes were very popular but their use in dipped wax heads often caused damage since

X-radiography of dolls and toys 253

(a)

(b)

(c)

(d)

Figure 20.3 Doll B,Armand Marseille girl doll, (a) photograph of head, (b) radiograph of arm; arrow indicates corrosion stained section of cord, (c) radiograph of head front view, (d) radiograph of head side view (see Figure 20.1b). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

the material was not always strong enough to support the moving weight. Bisque heads were stronger and many surviving bisque dolls have shut-eyes. ‘Crying’ or ‘talking’ dolls had been made since the early eighteenth century.The same simple device used then – a combination of bellows and a reed – continued in use for mass-manufactured dolls in the nineteenth century. The bellows were housed in the doll’s torso and could be made to work by squeezing,pulling

a cord or moving a limb. Doll D, a jester doll, has a dipped wax double-faced head which rotates when a cord is pulled (Figures 20.1d and 20.5b and d).Another cord operates internal bellows producing a ‘cuckoo’ sound. Manufacturers were keen to produce a doll that could move.The first mass-produced ‘Autoperipatetikos’ or walking doll was patented in 1862. Doll E, a walking doll, is effectively a cone-shaped container,

254

Case studies

(a)

(b)

(c)

(d)

Figure 20.4 Doll C, shut-eye doll, (a) photograph of head, (b) radiograph, front and side views, (c) radiograph, detail of composition shoulder plate, (d) radiograph of head, front and side view (see Figure 20.1c). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

housing a clockwork mechanism on wheels, with the head, arms and torso above (Figures 20.1e and 20.6). Dolls were patented with various moving parts: turning heads, lifting arms, and even one that could ‘swim’. More sophisticated mechanisms were clockwork but some, especially on cheaper dolls, were operated by strings or wires.

This remarkable array of materials, designs and mechanisms can provide the curator or researcher with many clues about date and, possibly, manufacturer. These need to be carefully analysed, especially since it should not be assumed that external features such as clothes, makers’ marks and materials will act as accurate guides.

X-radiography of dolls and toys 255

(b)

(c)

(a)

(d)

Figure 20.5 Doll D, double-faced jester, (a) radiograph, (b) photograph, moustachioed face, (c) radiograph, head, (d) photograph of alternate face (see Figure 20.1d). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

256

Case studies

(a)

(c)

(b)

(d)

(e)

Figure 20.6 Doll E,‘walking’ doll, (a) radiograph taken to reveal metal elements, (b) radiograph taken to reveal head and torso, (c) radiograph, head detail, (d) radiograph, detail of impregnated textiles from the head, (e) detail of impregnated textiles from the torso (see Figure 20.1e). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

The value of radiography for curation and conservation Radiography is a valuable additional investigative tool as it can provide corroborative information about materials and features, such as eye mechanisms, joint

construction and sound mechanisms, which may help to establish a doll’s origins and, possibly, date. Mapping both external and internal features can be undertaken without dismantling toys or undressing dolls. This is particularly useful in situations when dolls’ clothes are permanently attached, too fragile to handle or when

X-radiography of dolls and toys 257

(b)

(c)

(a)

(d)

Figure 20.7 Doll F, pedlar doll, (a) radiograph, full view, (b) photograph, head front view, (c) photograph, head side view, (d) radiograph of head and shoulder plate (see Figure 20.1f). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

258

Case studies

removal is considered ethically unacceptable. For example, the pedlar doll (Doll F) had been sewn into impenetrable layers of concealing clothing (Figures 20.1f and 20.7a). In the case of the ‘Autoperipatetikos’ (Doll E), whose very fragile silk dress is glued and pinned in place, the card cone would still conceal the mechanism even if the clothes could have been removed (Figures 20.1e and 20.6a and b). The fact that dolls were mass produced does not always aid identification.There may be no evidence of a maker’s mark. One type of head or construction feature could be used for many years with little or no change and the copying of successful designs was rife.The so-called Motschmann-type baby doll was widely copied (Doll G; Figures 20.1g and 20.8a). European collections may also include dolls that were made elsewhere, perhaps for the tourist market, and so contain other, unfamiliar materials.This is the case with the Caribbean doll which was crudely made from a range of locally abundant organic materials including wood, nuts, shells, seeds and beetle wing cases (Doll H; Figures 20.1h and 20.9). Dolls are primarily playthings and over time, clothes or limbs may be removed, altered and replaced.Through radiography unexpected and previously unsuspected features may become evident, such as the chicken wishbone in ‘The Old Woman in the Shoe’ toy (see Figure 1.1, p. 4). Radiography also enables accurate recording of damage and repairs. Conservators have long recognised this potential of radiography for understanding, documenting and aiding decision making when developing strategies for the treatment of dolls (Finkelstein, 1993; Noël, 2004).

Summary of radiography methods Experimental radiography of the dolls was carried out by the authors using the facilities of York Archaeological Trust’s (YAT) conservation laboratories.1 The range of data that would be revealed could not be anticipated.The project provided a great deal of information about more appropriate exposure parameters. In addition to low kV exposures to reveal textile layers, hair, stuffings and other organic components, higher energy exposures were made, some with lead intensifying screens, to study metal structures and other materials (see Chapter 3, pp. 41–43). For example, taking both types of exposure of the walking doll was essential to allow full understanding of the relationship of thedifferent components.The lead screen

intensification technique seemed to capture the most information in a single exposure of these disparate components but a low kV image is still necessary to complete the coverage (Figures 20.1e and 20.6a and b). Lower energy images of some of the dolls, for example about 10 kV, might have given more information about the textiles where they were not ‘blocked’ by more substantial components. Higher kV exposures would have provided better penetration but with the loss of image contrast and detail of the more ephemeral materials. The benefits of increasing exposure latitudes with lead intensifying screens are demonstrated by the range of materials revealed in the radiograph of the jockey doll’s head (Doll I; Figures 20.1i and 20.10a and b). As a result of these experiments, it is clear that taking different energy exposures maximises the information obtained for mixed media artefacts. YAT’s stock film was used (Agfa Structurix D7, 180 mm by 240 mm) in an industrial cabinet-type X-ray unit, a Hewlett Packard Faxitron (model 43855A option A04,0–130 kV range).This is designed to give high output of ‘soft’ X-rays and was ideal for low energy, high resolution imaging. However, the unit was located in an area lit by roof lights, making it impossible to work with unencapsulated film. To allow daylight handling, the films had to be packed into thick gauge, black plastic bags in a darkroom. It became clear that it is important to plan the dolls’ position during radiography carefully. In this experimental series, the limited space inside the X-ray cabinet restricted the positioning of the larger dolls so not all areas could be brought under the beam centre. As the aim of the project was to explore the potential of radiography for this type of object and not necessarily to produce complete radiographic surveys of each item, this shortcoming was not particularly serious.Where several exposures are required to give full coverage, the relative position of limbs and clothing needs to be carefully maintained to provide continuity across the images. It proved vitally important to have the objects to hand when interpreting the images. Features from external and internal surfaces and structures from the inside of an object are superimposed on the image and this may be very confusing.Viewing the image and object side by side allowed variations in densities produced by external features to be identified. Images of those furthest from the plane of the film are often unsharp, magnified, relatively displaced or distorted due to the conical projection of the X-ray beam (see Chapter 2, p. 17). Only those features in close contact with the film are reproduced at life size.

X-radiography of dolls and toys 259

(b)

(c)

(a)

(d)

Figure 20.8 Doll G, fixed-eye baby doll, (a) radiograph full view, (b) photograph of head, (c) radiograph of head, (d) photograph, detail of lower torso and legs (see 20.1g). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

260

Case studies

(a)

(b)

Figure 20.9 Doll H, Caribbean doll, (a) radiograph2, full view, (b) radiograph, detail showing the nut which forms the head (see 20.1h). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

Due to their depth and sculpturing, this applies to very few features of dolls and toys. The pairs of bright, curved lines in the middle of the radiograph of ‘My Dream Baby’s’ head were produced by the ears and are not part of the shut-eye mechanism (Doll A, Figure 20.2b). The ears are both the same size but the image of that furthest from the film plane (approx. 100 mm) appears to be noticeably larger and fuzzier.

Digitising these radiographs demonstrated how much information had been captured which could not be seen when viewing the films on a light box. For example, there appeared not be any penetration of the head in the radiograph of Doll C, a shut-eye wax and composition doll (Figure 20.4b). Adjusting the contrast through digital image processing (DIP) of the greyscale values revealed a wealth of hidden detail (Figure 20.4d).

X-radiography of dolls and toys 261

Information from radiography

wood features in the lower arms and legs are hidden by the more radio-opaque gesso-like surfacing layer.

Materials and construction Textiles In most of these radiographs, little is visible of weaves even after digital image contrast adjustment. This is because the beam energies were selected to highlight more substantial – and hence more radio-opaque – materials. Other exposure levels would have provided information but, in these exposures, textiles are generally visible where they have been caught on edge or are folded. In some cases, radiography did provide further information. In the case of Doll F, the pedlar doll, the woven structure of her red cloak became visible, confirming it is a felted fabric rather than felt. Her gathered skirts beneath the apron and the extent of the legging on the proper left leg can also be mapped.The stronger image of the checked pattern of the patch on the proper right side of her apron suggests the presence of a metal mordant (Figure 20.7a). Surprisingly strong images of the clothes worn by Doll I, the Caribbean doll, appear on the radiograph which was taken at 40 kV. This also suggests the presence of a metal mordant or a finish on the rather stiff cotton garments (Figure 20.9a).The attachment of the arms to the body by a textile tube glued to the upper arm and then nailed to the upper back can only be seen through radiography. Wood The higher exposure level used for the Caribbean doll confirmed her head to be a nut (Doll I, Figures 20.1i and 20.9a and b). It also showed that, under the firmly attached clothes, her body is a billet of wood, hollowed down the centre and split to make the legs (O’Connor and Brooks, 2007). Wood with prominent growth ring structures is easily recognisable depending on the direction of the radiograph relative to the grain of the wood. In certain directions the differences in wood density produce irregular dark and light banding. This is seen in the radiograph of the upper forearm of Doll B (Figure 20.3b).These structures are absent in the radiograph of the Caribbean doll but close-up examination shows longitudinal subtleties in image density which are typical of wood’s anisotropic structure (Doll I, Figure 20.9a). Her shell, seed and nut jewellery is visible on the radiograph but the decorative iridescent green beetle’s wing does not show up as it lies over the more radio-opaque wood. Radiographs show that the pedlar doll is actually a wooden peg-doll with a composition head and shoulder plate (Doll F, Figure 20.7a). A wood knot was observed in the torso but all hints of

Wax The radiograph of the wax-coated head of Doll D, the jester doll,indicates it is solid composition over a metal rod. The wax layer contains different sized coarse and angular particles, possibly mineral derived (Figure 20.5c). In contrast, the radiograph of the walking doll’ s head shows it to be hollow (Doll E, Figure 20.6a). It appears to be made up of small fragments of textiles set in a mineral-based substance, possibly plaster of Paris. This provides a strong foundation for the surface wax coating; note that the nose appears to be made totally of wax (Figure 20.6c).The weave structure of the fragments in the head is evident as a jumbled pattern on the radiograph (Figure 20.6d).This type of pattern is produced when textiles are set into, or impregnated with, a more radio-opaque matrix. Bisque Dolls A, B and I all have bisque heads (Figures 20.2, 20.3 and 20.10). The radiograph of the bisque used to make the head and hollow lower arms of Doll B is typical of this substance, appearing rather featureless as might be expected of this very fine-grained and homogeneous material. The teeth, which also seem to be bisque, were made separately and then applied behind the lips (Figures 20.3c and d). In contrast, the hollow body is more variable in its radio-opacity and is probably moulded composition. Composition and coatings Composition can be the foundation for many different types of surface coatings. The damage to the proper left side of Doll C’s face reveals the three layers making up the head and shoulder plate – the grey fibrous composition, a fine pink gesso-like layer and the yellowed wax coating (Figure 20.4a). This gesso-like material was used on a number of the dolls to give a smooth finish to wood and composition surfaces. It is often white in unexposed areas but, where it was meant to be visible, it is coloured pink and may be coated in wax to give a skin-like effect. Numerous bubbles of varying sizes are visible in this coating.These can also be detected in the radiographs; for example, in Doll B’s composition body, Doll C’s shoulder plate and Doll F’s wooden legs (Figures 20.3b, 20.4c and 20.7a). Radiographs can also indicate the junction of seams in composition where external evidence has been removed by careful working or surface coatings. The radiographs of the hollow composition hands of

262

Case studies

‘My Dream Baby’ show that they were cast in a twopiece mould with seams down the sides (Doll A, Figure 20.2d). Discontinuities evident on a radiograph can raise new questions. The head of the pedlar doll was thought to be solid composition with a wax coating (Doll F, Figures 20.7b and c). The cracks visible to the naked eye at the neck and on the side of the head were interpreted as damage. However, after digitisation, further exploration of the radiographs suggested an alternative interpretation of these features. It seems the construction is more complex and that the wax face may have been moulded separately and applied using similarly coloured molten wax to a relatively featureless composition head. The wax layer appears to have been carried down onto the shoulder plate. It shows up as an external layer over the neck and shoulders in the radiograph (Figure 20.7d). Her eyes are partly hidden by heavy soiling but are very distinct on the radiograph; their radioopacity suggests that they may have been painted with lead white. Stuffings Stuffings including wood-wool and, possibly, horsehair which can be seen on radiographs of several of the dolls. Doll C has a ‘straw’ stuffing.This is visible through a small hole, possibly made by a rodent, in the corner of the cotton bag forming the body.The radiograph provides additional information about the method by which the stuffing was introduced into the body, showing that the long, flat, ribbonlike strands have been formed into small rolls or mats, some more densely packed than others (Figure 20.4b). These lie in a longitudinal direction across the body. Doll G has textile strips wound around the squeaker (Figure 20.8a). Wig attachment Dolls’ wigs can be human hair or other fibres such as mohair. Poured wax heads may have hair inserted directly as individual fibres or tufts while heads with a wax coating over another material often have a slit for the insertion of a wig. Open head dolls may be either moulded complete and the crown cut off in order to insert their eyes or made with no crown as in the case of Doll B (Figures 20.3c and d).This opening is generally covered with a ‘pate’ of cardboard, cork or other material before the wig is attached. Much damage may be caused by lifting wigs in order to establish

the attachment method but this can be avoided with radiography. Doll I is a jockey doll which forms part of a nineteenth century German pull-along musical toy.The doll’s bisque head with a large crown hole for the pate (Figures 20.10a and b).The holes for the attachment of the wig and the faint line of a cord running up from a hole in the neck post, through the attachment holes and the wig and down again, can be seen on the radiograph (Figure 20.10b). Articulations and mechanisms Heads and limbs The radiograph of the pedlar doll’s peg body and limbs shows that she has dowelled elbow, hip and knee joints (Doll F, Figure 20.7a). Doll C (Figure 20.4b), Doll G (Figures 20.8a and d) and the jester doll (Doll D, Figure 20.5a) all have Motschmann-type limbs.Doll G has the composition head and separate upper and lower torso sections which are typical of a Motschmann-type doll, with feet and hands which articulate at the ankles and wrists.This construction results in a remarkable degree of articulation. The radiograph shows that the head pivots on a post extending up from the upper torso section, in a manner similar to that of the jockey doll (Doll G, Figure 20.8c and Doll I, Figure 20.10b). Doll B has cup and ball joints (Figure 20.3b).This radiograph shows the wooden upper arm and hollow composition lower arm and torso. The hollow lower arm has a wooden ball set in the top; a metal bar can be seen fixed across the cavity of the arm. This ball articulates with the cupped surface of the lower end of the upper arm. Even though the hidden cord securing the arm sections to the body is too radio-lucent to show in these exposures, evidence revealed by the radiographs make it possible to infer its path.The cord evidently passes round the bar at the elbow joint and then up to the shoulder through a hole drilled down the length of the upper arm. This hole becomes visible only in the upper section of the arm where the space around the cord is plugged with a radio-opaque filler or adhesive. At the top of the arm, it loops around the hooked end of the thick wire connecting the two arms through the torso. It becomes visible on the radiographs (see arrow), probably because it has been stained by corrosion products from the wire. Two further wires come down from inside the head through the neck and diverge, one towards each leg. These wires pass behind the weight for the shut-eye mechanism and hook over the back of the head through the crown hole (Figure 20.3d).

X-radiography of dolls and toys 263

(a)

(c)

Squeakers, criers and other sound mechanisms The presence of sound mechanisms can sometimes be detected because they still make noise when a doll is moved but can be harder to establish if they have ceased to work or require pressure to operate. Radiography can reveal their form and detail, whether they are working or not. If it is possible to develop sufficient data about the range and type of these mechanisms through radiography, this information could be useful in identifying makers or in dating. The radiograph of Doll C shows where its squeaker is located and that it is activated by a coil

(b)

Figure 20.10 Doll I, jockey doll, (a) photograph, doll’s head, (b) radiograph, doll’s head; arrows indicate location of the displaced eyes, (c) photograph of base of neck revealing an eye (see 20.1i). (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust,York Castle Museum.)

spring fixed between front and back boards of wood; the pin relates to the clothing (Figure 20.4b). The reed, which produces the sound when the boards are compressed and released, is seen in the front view.The side view confirms that the spring is in working condition although, under magnification, its outline is granular in places suggesting there is some metal corrosion. A membrane, perhaps of fabric or leather, encapsulates and restrains the expansion of the squeaker.This itself cannot be seen in the side view but is clearly present as it has prevented the ingress of the stuffing.The radiograph of the reed and wooden board squeaker inside Doll G

264

Case studies

is very similar except for the form of the spring (Figure 20.8a). This squeaker seems to be secured into the rigid upper and lower sections of the doll. In contrast, the radiograph of ‘My Dream Baby’ shows a circular crier with a perforated metallic disc and fabric bellows which would have been operated by tilting the doll (Doll A, Figure 20.2e).The jester doll has oblong bellows with an external sprung hinge, seen on edge in the radiograph (Doll D, Figure 20.5a). The cord to operate this passes in through a metal eyelet on the proper left of the body tube and is knotted behind a metal lever extending from the nearest board of the bellows. Shut-eye mechanisms Identifying details of shut-eye mechanisms may be helpful in dating dolls’ heads. It might be expected that the eyes in static eye dolls would be lenticular. The jester doll has two pairs of such eyes in its solid head (Doll D, Figures 20.5b, c and d). Doll G appears to have a hollow head and also has lenticular fixed eyes (Figures 20.8b and c). However, the radiograph shows the fixed eyes in the hollow head of the walking doll are blown spheres with stems (Doll E, Figure 20.6c).These eyes are the same shape as those in Doll B (Figures 20.3c and d). Here the hollow stems house the wires from the shut-eye mechanism weight which opens and closes the eyes as the doll is tilted forwards and backwards. Shut-eye mechanisms can also be seen in the radiographs of Dolls A and C (Figures 20.2b and c and 20.4d). Each has a weighted counterbalance and, in each, the weight is on a wire which arches above the level of the eyes and then passes down into the back of the eye support. However, the shape and position of the mechanisms are different.The internal sculpturing in the two bisque Armand Marseille heads (Dolls A and B), which provides the sockets in which the glass eyes pivot, is similar (Figures 20.2b and c and 20.3c and d). In Doll C’s wax and composition head, the shape and arrangement of this support seems to be proportionally more substantial (Figure 20.4d). The thickening also extends across the face below the eyes and above the level of the mouth, perhaps reflecting the weaker nature of this material. Novelty mechanisms Cords and rods may break off, leaving mysterious holes in the torso, and moving parts seize up, leaving no clue as to what functions the toy originally performed. Taken specifically to image unusual mechanisms, radiography may establish their purpose without the risks implicit in either activating them or opening up casings.

The radiograph of the walking doll shows that the body is made of two interlinking card cones (Doll E, Figure 20.6b). The top cone forming the upper body is inverted over the lower one housing the clockwork walking mechanism which appears to be in working order (Figure 20.6a). The construction technique is crude and cheap, secured by many nails and roughly finished.The nails appear to be standing proud of the cones in the radiograph but this is because the doll’s clothing is not visible at these exposures.The lower cone is closed by a card disc with slots for the metal feet. Both cones are covered with stiff sized cotton but this is only visible on the radiographs just below the edge of the shoulder plate where a radio-opaque substance is visible in the interstices of the weave (Figure 20.6e).This is similar to the effect noted by Wettering when canvas supports are made visible in radiographs of oil paintings (see Chapter 29, pp. 325–6). Perhaps this textile has been pressed into a thick layer of adhesive used to help secure the body form to the shoulder plate. Originally the jester doll’s hood was integral with the composition shoulder plate (Doll D, Figure 20.5d). This has broken away revealing the wooden pulley and cord which turned the head (Figure 20.5b). As would be expected, metal elements show up strongly including the eyelets surrounding the holes for the operating cords, internal spring mechanisms, connecting wires, rods, metal threads, buttons and bells. A horizontal rod inside the shoulder plate connects with a vertical rod which runs through the centre of the pulley and connects the head with the body (Figure 20.5a). The rest of the mechanism and the squeaker are housed in the double card tube forming the body. This radiograph was, however, taken at low kV so many of the more ephemeral elements are also visible, including the routes of the cords which provide a context for the mechanism. Damage and repair External features are easily damaged on dolls made from fragile materials, and sometimes they disappear completely. Eyes are a common casualty. In the case of the bisque-headed jockey doll, the location of the doll’s ‘missing’ eyes was revealed (Doll I, Figures 20.10a and b). One of these lenticular glass eyes was lodged at the intersection of the head and the body while the other remained inside the head (Figure 20.10c). Radiography can provide information about condition. The wax and composition head and shoulder plate of Doll C have suffered considerable damage (Figures 20.4a, c and d). Long horizontal

X-radiography of dolls and toys 265

cracks typical of composition heads with shut-eye mechanisms have formed on either side of the eye sockets (Figures 20.1c and 20.4a). The proper left eye socket has also been enlarged by a gnawing mouse.3 Some of this damage is visible in the photograph but the radiograph shows that there is other cracking below the wax layer; this may partly be due to shrinkage of the composition.This has led to the pink gesso-like coating below the wax on the proper left side of the doll’s face becoming detached from the underlying composition. Damage not yet visible on the surface is thus highlighted by the radiograph. Many of the cracks around the face of the pedlar doll’s head seem to relate to its construction; these may have been exacerbated through handling. The crudely applied thin sheets of unpigmented wax may be part of the original construction, possibly to attach the bonnet and the hair strip, or repairs (Doll F, Figure 20.7c). The red sealing wax on top of the head may have been an attempt to reattach the bonnet (Figure 20.7b). All these layers are visible – and visibly different – on the radiograph (Figure 20.7d). Adjusting the digital images allows these layers to be differentiated so giving the opportunity to link the cracks and the layers, providing an additional insight into the history of a complex structure.

appeal. For example, the Museum of London (2006) includes the radiograph of a pedlar doll as part of its on-line Object Handling Workshop for children. Seeing inside a doll or toy means that radiography can be a powerful tool for visitors’ education and enjoyment.

Acknowledgements The authors would like to thank York Museums Trust and York Archaeological Trust.

Notes 1.

2.

3.

In 1989, Sonia O’Connor was Deputy Head of Conservation, York Archaeological Trust, and Mary Brooks was Assistant Keeper of Textiles (Conservation),York Castle Museum. Josie Sheppard continues as Curator of the York Castle Museum Costume and Textiles collection, now part of York Museums Trust. This radiography was carried out in 2005 using the Faxitron X-ray unit at the Department of Archaeological Science, University of Bradford, with Agfa Structrix D4 film. Identification of tooth marks by Professor T. P. O’Connor, Department of Archaeological Sciences, University of York.

Summary The range of information revealed in the experimental radiography of these dolls and toys is quite remarkable despite the fact that it was not always possible to anticipate the materials in advance. Dealing with solid and hollow structures and the wide variation in thickness also made these objects particularly challenging to image. In all cases, it was necessary to take more than one exposure both to cover the range of materials involved adequately and also to gain different views of the dolls to aid image interpretation. Some components remained masked by other more radio-opaque materials, no matter what beam energy or imaging technique was selected. Radiography can help minimise handling damage while informing conservation interventions and provide vital information for collectors and curators. Features such as shut-eye mechanisms or limb attachments can be understood and mapped and there is potential for building comparative databases to aid dating and attributions. Radiography should, therefore, form part of the systematic documentation of dolls and toys, whether or not the former can be undressed. In addition, radiographic images can have a broader

References Finkelstein, L. (1993). Preservation of dolls in the collection at the Colonial Williamsburg Foundation. In The American Institute for Conservation of Historic and Artistic Works. Abstracts of Papers presented at the Twenty First Annual Meeting Denver, Colorado May 31–June 6, 1993, p. 63. American Institute for Conservation of Historic and Artistic Works. Museum of London. (2006). http://www.museumoflondon.org.uk/learning/features_facts/voh/voh_kit/obje ct03.htm (accessed 2 March 2006). Noël, G. (2004). La Collection de Poupées Koenig: son histoire la restauration de quatre poupées costumées. Etudes et réalisation de supports de Conservation Presentation (unpublished thesis). Institute National du Patrimoine. Department des Restaurateur du Patrimoine, Textiles, Saint-Denis La Plaine. O’Connor, S. and Brooks, M. M. (2007). Looking into the past: the potential of X-radiography as an investigative technique for archaeological and ethnographic textiles. In Recovering the Past:The Conservation of Archaeological & Ethnographic Textiles. Mexico. National American Textile Conservation Committee Conference Proceedings (S. Thomassen-Krauss, ed.), National American Textile Conservation Committee.

21 X-radiography of teddy bears and other textile artefacts at the Victoria & Albert Museum Marion Kite

Introduction Textile conservators have infrequently required radiography of objects as part of routine documentation and assessment. However, it is worth noting that this may also be a useful tool in the range of available investigative techniques when surface examination and handling does not provide sufficient information about an object’s construction. Radiography has been used on particular occasions at the Victoria & Albert Museum (V&A) for fifty years, and possibly longer, in order to reveal unseen elements and details of internal construction in three-dimensional textile objects such as dolls and soft toys as well as for compacted and layered archaeological material.

Early radiography of textiles and dolls at the Victoria & Albert Museum The earliest surviving record of radiography associated with the conservation of textile artefacts at the V&A is dated 1968. The archaeological remains of a textile found in the tomb of Archbishop Walter de Gray from York Minster were radiographed at the British Museum (King, 1971; Werner, 1971: 40). On this occasion, the information gathered enabled informed choices to be made regarding both the further investigation and the possible conservation of these fragments at the V&A (Werner, 1971: 139). During the 1970s, radiography was used to examine a collection of seventeenth and eighteenth century dressed dolls, including Lord Clapham. The purpose was to reveal, among the hidden layers of clothing, the original pins which were believed to be holding that clothing in place. Methods used for the construction of the dolls’ bodies, joints and the 266

means of articulation of various limbs, the presence of other pins, nails or fixings and sundry other details were also revealed (Figure 21.1). Some of the dolls had not been undressed since their acquisition by the museum and it was not proposed to do so unless this became absolutely necessary because urgent interventive conservation was required. Radiography has been used from time to time since the 1970s on objects from the textile collection but it has been used most frequently for the examination of toys and dolls.

Radiography of teddy bears The year 2003 marked the centenary of the teddy bear and a large exhibition, The Teddy Bear Story: 100 Years of the Teddy Bear, was held at Liverpool Museum and V&A’s Museum of Childhood at Bethnal Green, London. In preparation for this exhibition, 318 bears from the Museum of Childhood collection were examined and their condition checked. Forty-eight bears from this collection required cleaning and conservation. Six teddy bears were radiographed. These were selected because more information about their construction was required than could be seen or determined from examining them externally. Information required related mostly to fixing methods used for heads and limbs and other internal mechanisms. It was particularly needed in cases where limbs or heads were loose or detached. It was required in order to assess the nature of these fixing methods and to help plan the conservation procedure to be followed. It was useful minimising the extent of unpicking required in order to gain access for treatment.

X-radiography of teddy bears and other textile artefacts 267

(a)

(b)

Figure 21.2 English Teddy bear c. 1930–1959 with detached head (Misc. 1237-1991), (a) photographed during conservation by Marion Kite, (b) radiograph by Paul Robins, Victoria & Albert Museum.

Figure 21.1 Radiograph of Lord Clapham, c. 1700, showing details of the doll’s layers of clothing, buttons, pins, and other fixings (Victoria & Albert Museum T.8491974). Composite image from two radiographs; due to movement between the two exposures these cannot be completely joined. Archive image from Textile Conservation, Victoria & Albert Museum. See http:// www.vam.ac.uk/images/image/13894-popup. html for a photograph of the dolls known as Lord and Lady Clapham.

In general, damage had come about through use and wear, poor storage and inappropriate repairs carried out in a domestic situation while the bears were still functioning as toys and companions and before they became part of the museum collection (Kite, 2003). In the case of one English teddy bear (Misc. 1237-1991), made between 1930 and 1959, the head had become detached (Figure 21.2).

268

Case studies

Figure 21.3 A worn and sagging English teddy bear (B225-1999). (Photographed during conservation by Marion Kite, Victoria & Albert Museum.)

There were other cases where limbs had become loose, joints stretched, heads were nodding and falling forward while stuffing had degraded, powdered and become compressed, resulting in a distorted and sagging appearance (Figure 21.3). X-raying these bears not only helped assess specific questions about their condition but also aided in the understanding of their construction.

Taking and interpreting radiographic images It is important that radiographic images are taken with an understanding of the artefacts themselves. The conservator needs to brief the radiographer with regard to the nature of the information being sought so the images can be taken accordingly. In this case, stuffing and some interior features are easier to see on the radiographs taken at lower energies while radiographs taken with more penetrating higher energy X-rays provide more information about wires and armatures. In the V&A, this work falls to a photographer trained to undertake radiographic imaging for all

conservation disciplines. The X-ray images of teddy bears illustrated here were made using Agfa Structurix D7 daylight wrapped film (350 mm 430 mm) with the bears laid directly upon the encapsulated film. Exposures were in the order of 30 seconds using an X-ray tube at a focus to film distance (FFD) of 500 mm at 50 kV/5 mA. No filtration was used. This relatively high kV has produced rather lowcontrast grey images but was selected for a single exposure because of the range of materials (metal, card, wood, textile, leather, stuffing fibres, etc.) expected to be encountered in these objects. The films were processed on an automatic processor. For this chapter, the images were digitally manipulated to improve the contrast for ease of interpretation and improved print quality.1 No information is available concerning the methods used over thirty years ago to produce the radiographic images which were discussed earlier. An X-ray image can be very straightforward to read and is excellent for interpreting methods of construction and the order of assemblage of a bear, providing common sense is used together with an understanding of pattern cutting and toy making. Although the method of attachment of arms, legs and head to the torso of a bear may be straightforward, there can be variations in the way all the parts of an articulated bear are joined together. It is useful for the conservator to understand which method of construction has been used in order to identify which seam would be the most appropriate for unpicking should this be necessary to facilitate repair.

Stuffings, squeakers and structures It is often possible to determine the nature of the stuffing used in a particular teddy bear by carefully ‘feeling’ and manipulating the bear and by balancing what can be ‘felt’ with an understanding of which type of stuffing may be found at particular dates. For example, traditional stuffings included kapok, woodwool, textile waste (described as ‘sub’) and a mixture of wood-wool and ‘sub’ which was used during World War II (Cockrill, 1993). It will, of course, be much easier to see what has been used if the fabric of the bear is torn or a seam has failed. However, when an older example of a teddy bear is fragile and it has been made using mohair plush fabric,2 handling must be kept to a minimum. Degraded mohair plush will tear easily and so it may not be possible to squeeze, manipulate or ‘feel’ the bear to find out what type of stuffing is inside. For the conservator, unpicking an

X-radiography of teddy bears and other textile artefacts 269

original seam is a radical intervention which is only carried out when other means of gaining access are not available and when it is essential in order to carry out remedial conservation treatment. It is often possible to determine the nature of the stuffing used by looking at the ‘swirl’ and ‘cloud’ patterns which can sometimes be seen on the radiograph. These are caused by the tangle of thinner or thicker fibres of the stuffing materials. It must be remembered, however, that materials other than traditional ones may also be found inside bears. One bear (Misc. 1237-1991) is stuffed, not only with wood-wool but also has scraps and off-cuts of lambskin with the fleece attached – the same material from which it was made (Figures 21.2a and b). The pattern caused by these types of stuffing is of a more solid form than that created by some of the traditional materials. X-ray imaging is also a useful and efficient method of gaining information on the internal structure of a teddy bear including details of fixings and other mechanisms. Many bears have internal features such as growlers, which come in a variety of forms, squeakers, musical boxes, fixings for eyes, bars to connect various limbs to allow for combined or separate movement and a range of other complex devices for a variety of functions. Radiography reveals one such mechanism inside a Schuco bear known as Baby Bar (Misc. 6-1978), dating from the 1920s to the 1930s (Figures 21.4a and b). This mechanism connects the head and tail, facilitating the movement of both parts in conjunction, and allowing the head to rock in a controlled manner. Sometimes hidden repairs or alterations may also be revealed. Figure 21.5 shows a wire hidden within the soft parts of the head and upper torso of a German teddy bear dating from 1908–1913 (B10-1998). This has been used as a crude repair to reattach the head to the body.

(a)

(b)

Figures 21.4a and b Large Schuco teddy bear known as Baby Bar (Misc. 6-1978). Two radiograph plates were required to show full details. The position of the bear differs between the two exposures, giving a different view of the mechanism. (Radiography by Paul Robins, Victoria & Albert Museum.)

Construction and assembly methods In the case of some bears, the head and limbs were constructed with the washers and fixings in place and were stuffed and finished first in order to ensure that their fixing components were held firmly in place by the stuffing (Misc. 1237-1991, B10-1998 and Misc. 13-1968). The head and limbs would next be fixed onto the hollow body cavity. Small metal washers were usually used on either side of the large card washers (Figures 21.5 and 21.6). A small opening could possibly have been left at the

fixing point of each limb to the body to allow for access and adjustment of the joint. The final securing mechanism of a bent nail, nut – where a threaded bolt was used (Figure 21.2b) – split pin or cotter pin (Figure 21.5), inside the body cavity ensured the close fit of head and limbs. At this point any access openings left at the top of the limbs would have been sewn up. Next, the body cavity could have been fitted with a growler and any other internal mechanisms and then stuffed and sewn up. The final stitch line is usually in the centre front or

270

Case studies

Figure 21.5 Radiograph showing a wire hidden within the head and upper torso of a teddy bear, which rejoins the head to the body (B10-1998). (Radiography by Paul Robins, Victoria & Albert Museum.)

centre back. The seam is often crude but it is lost within the pile of the fabric although it can be clearly visible on a radiograph. In Figure 21.5, the spring from the squeaker can clearly be seen to have become detached from the two card ends of the squeaker (B10-1998). The spring has moved from the bear’s tummy and come to rest against the leg washer but the card ends remain in the tummy. A lighter shadow on one of the card ends indicates the position of the reed for the squeaker. The placing of metal washers against the card washers is clearly visible. The squeaker in a Schuco teddy bear made in the 1930s (Misc. 531964) is seen from the end view only but the wire curls of the spring, the fabric of the bellows and the reed may be discerned (Figure 21.7). This squeaker does not appear to have disintegrated. One bear (Misc. 1237-1991) does not have a squeaker or growler (Figure 21.3). The eyes in this bear have long wire attachments which were fixed by coiling

Figure 21.6 Radiograph of a German teddy bear, c. 1908 (Misc. 13-1968). Note the stretched pin holding the proper left leg. The leg is still secure but not held in close proximity to the body. (Radiography by Paul Robins, Victoria & Albert Museum; composite image from two radiographs.)

these round inside the head. The radiograph of the German teddy bear, dated about 1908 (Misc. 13-1968), shows the pin holding the proper left leg still secure but no longer in close attachment to the body (Figure 21.6). Another method was to complete the body first, stuffed and finished with all internal mechanisms, such as growlers, in place. The head would be fitted and secured before the body was sewn up with card washers at the base of the head and the top of the

X-radiography of teddy bears and other textile artefacts 271

Figure 21.7 Radiograph of Schuco teddy bear, c. 1930 (Misc. 53-1964). Note the stretched metal loop fixing on the proper left eye. (Radiography by Paul Robins, Victoria & Albert Museum.)

body. Finally the head was completed and sewn up. Depending on the fixing method chosen, washers may or may not have been used in the fixing of the limbs. Whatever construction method was chosen, a bent pin or wire was used to secure the arms or legs. This passed from one side of the body to the other and the ends were ‘lost’ in the upper section of the limbs of the bear. The final portions of stuffing would then be inserted into the limb to cover and mask the fixing prior to being sewn up. Again, the stitching might be crude but it would disappear in the plush fabric. One bear (Misc. 53-1964) clearly shows one variation of this method (Figure 21.7). There is a nut and bolt fixing at the neck with a small metal washer between the head of the bolt and a large card washer in the head as well as a large card washer and a small

metal washer and nut in the body. An indication of the kapok and wood-wool stuffing can just be seen in the swirling pattern in the body. No washers have been used at the joins of the arms and legs to the body but the ends of the wires securing the arms and legs have been bent round into the stuffing of the limbs. The metal loop fixing on the proper left eye has stretched, indicating that this eye is at risk of falling out. A third approach was to complete the head first with the mechanism for the movement secured in place, as seen in Figures 21.4a and b, in Baby Bar (Misc. 6-1978). The body is fitted onto this and the other mechanisms, such as the growler, were then inserted. The mechanism to move the head is connected to the tail so both move together. A wire armature therefore goes through the body from the complex head mechanism into the tail section. The wire is turned back to fit snugly into the tail, where it makes a substantial metal loop. This serves to strengthen the tail in order to withstand repeated handling. In this bear, the final stage of assembly was fitting the limbs. The pins holding these pass through the washers located in the limbs, then through the washers located in the body cavity where the securing pins are bent back. The body cavity is well packed with stuffing and then sewn up. Once again, the stitching is lost in the plush fabric. For the purposes of taking the X-ray image, this bear was placed upon its back with the tail therefore lying to one side. Reading the resulting image could cause confusion as it could be seen as suggesting that the armature wire from the head is also connected through to the leg of the bear. This is not the case. The tail is lying directly behind the top of the leg so the image of the metal leg fixing pin is overlaid onto the tail wire. Further variations on any of these methods may occur. The head of a German cloth teddy bear (Misc. 10-1986), dating to c. 1910, is not articulated. However, the proper right arm and leg are connected to the tail and move in conjunction (Figure 21.8). This bear does not appear to have a growler. The metal plate insert in the proper right leg is probably part of the mechanism for lifting the leg but it is no longer attached to the wire projecting through into the top of that leg. The proper left arm is held in place by means of a button fixed through to the outside of the arm. Clearly the arm was once fixed onto the projecting wire but at some time it had broken away. The button method was used as a quick, probably domestic, repair to prevent the loss of the limb. The button provides a solid structure through which to pass the securing thread.

272

Case studies

in a different material to the rest of the body (Figure 21.5). In the English teddy bear (Misc. 1237-1991), the characteristics of the outline of the head and the body are quite different to that produced by the fabric of the limbs (Figure 21.2b).

Conclusion Radiography has thus demonstrated its value as an investigative tool to aid in both the understanding of textiles and in the development of effective conservation decision making. However, it is important that the nature of the artefact is taken into consideration when planning the radiography of textiles and soft toys.

Acknowledgements I would like to thank Paul Robins for radiography of the bears and Sonia O’Connor, University of Bradford.

Notes 1. 2.

Figure 21.8 Radiograph of a German teddy bear, c. 1910 (Misc. 10-1986). (Radiography by Paul Robins, Victoria & Albert Museum.)

It acts to prevent a concentrated strain point on the mohair plush fabric which would quickly give way and again cause the arm to tear away. Different arm fixings may also be seen in other bears (Figure 21.7).

Threads and fabrics Although in the German bear (Misc. 10-1986) the thread securing the button to the arm is not visible, the thread used in the seaming of the plush fabric shows up exceptionally brightly on the radiograph (Figure 21.8). In this bear (B10-1998), the radiograph also shows that the pads of the feet are clearly

Digital manipulation was carried out by Sonia O’Connor, University of Bradford. Mohair plush or ‘Teddy Bear Cloth’ was also known as Yorkshire Cloth on account of its place of manufacture. It was usually a mixture of mohair, wool and cotton (Cockrill, 2001: 5).

References Cockrill, P. (1993). The Teddy Bear Encyclopaedia. Dorling Kindersley. Cockrill, P. (2001). Teddy Bears and Soft Toys. Shire Books. King, D. (1971). The textiles. In The Tombs of Archbishop Walter de Gray (1216–55) and Godfrey de Ludum (1258–65) in York Minster and their Contents (H. G. Ramm, ed.), pp. 127–131, Society of Antiquaries of London. Kite, M. (2003). 100 Years of the Teddy Bear. V&A Conservation Journal, 43, 4–6. Werner, A. E. A. (1971). The scientific examination and treatment of objects from the tombs. In The Tombs of Archbishop Walter de Gray (1216–55) and Godfrey de Ludum (1258–65) in York Minster and their Contents (H. G. Ramm, ed.), pp. 136–140, Society of Antiquaries of London.

22 X-radiography of patchwork and quilts Mary M. Brooks, Sonia O’Connor and Josie Sheppard

Introduction The decorative and colourful surfaces of historic quilts and patchwork can conceal more than layers of fabric. Examples of patchworking and quilting are found in many museum collections in Britain, very often in the form of bedspreads although these techniques may also appear on a range of smaller domestic items, such as table and cushion covers, mats, tea cosies and garments. The range of materials and techniques is as varied as the uses to which these objects were put. Some are homely pieces thriftily made from cotton or wool,

and used primarily for warmth. Others are sophisticated works in silk or velvet, intended to show off their makers’ sewing skills, while adding colour and texture to a domestic interior. This chapter does not attempt a complete survey of different types of quilts and patchwork but seeks to show how radiography can provide new information to enhance understanding and contribute to conservation decision making. A range of quilts and patchwork covers, from the early eighteenth to the twentieth century, were X-rayed; see Table 22.1. The reasons for radiography were varied: some were X-rayed to answer particular conservation

Table 22.1 Radiography of patchwork and quilts Type and date

Museum number

Techniques and makers (where known)

Radiographic exposures

Figures

Beaded table cover 1880–1890

YORCM: BA239

Clamshell patchwork

14 kV; 5 mA; 4 mins; Agfa D7 Daylight Wrapped film; 1 m FFD

22.1

Doll’s bedspread 1830–1840

YORCM: BA622

Hexagon patchwork. Sarah and Mary Pritt

14 kV; 5 mA; 4 mins; Agfa D7 Daylight Wrapped film; 1 m FFD

22.2

Quilt 1890–1910

YORCM: BA3019/1

Hand stitched, single cloth. Reused quilt as wadding. Mary Burnett

15 kV; 3 mA; 4 mins; aluminium filter; Kodak AA 400 Ready Pack film; c. 1 m FFD

22.3 22.8

Eiderdown 1950s

Private collection

Machine stitched, single cloth

15 kV; 2 mA; 4 mins and 8 mins; aluminium filter; Agfa D4 film, unencapsulated; c. 0.6 m FFD

22.4

Coverlet 1880–1900

YORCM: BA3034

Log cabin. Hand and machine stitched

14 kV; 5 mA; 4 mins; Agfa D7 Daylight Wrapped film; 1 m FFD

22.5

Bedcover 1880–1900

YORCM: BA607

Crazy and chevron patchwork, embroidery, appliqué. Hand and machine stitched

14 kV; 5 mA; 4 mins and 20 kV; 5 mA; 4 mins Agfa D7 Daylight Wrapped film; 1 m FFD

22.6

(Continued ) 273

274

Case studies

Table 22.1 (Continued ) Type and date

Museum number

Techniques and makers (where known)

Radiographic exposures

Figures

Durham quilt 1900–1930

YORCM: BA715

Single cloth. Hand and machine stitched. Made in Hartlepool

14 kV; 5 mA; 4 mins; Agfa D7 Daylight Wrapped film; 1 m FFD

22.7

Quilt 1819

YORCM: BA984

Patchwork, quilting and appliqué. Elizabeth Watson

15 kV; 3 mA; 4 mins; aluminium filter; Kodak AA 400 Ready Pack film; c. 1 m FFD

22.9

Quilt 1805–1816

YORCM: BA1205

Patchwork and quilted (waves, wineglass and diamonds). Initialled ‘YGF’

15 kV; 3 mA; 4 mins; aluminium filter; Kodak AA 400 Ready Pack film; c. 1 m FFD

22.10

Oblong panel, ribbon binding 1690–1720

YORCM: BA3058

Cord quilting and embroidered floral motifs

14 kV; 5 mA; 4 mins; Agfa D7 Daylight Wrapped film; 1 m FFD and 15 kV; 2 mA; 4 mins; with and without aluminium filter; Agfa D4 film, unencapsulated; c. 0.6 m FFD

22.11

or curatorial questions while some radiographs were taken specifically for public display.

Quilting and patchwork: a brief overview At its most basic, a patchwork cover consists of a single layer made from small pieces of fabric cut to shape and then stitched together, usually forming a pattern. A cover of this type is decorative but does not provide weight or warmth when used on a bed. These are supplied by the addition of a backing fabric which may simply be an old sheet or some plain cotton or, on some covers, another piece of patchwork. A third inner layer is often added, usually thicker than the two outer covers. This may be cotton or wool wadding, an old blanket or even an old bedcover which has seen better days. Once this textile ‘sandwich’ has been assembled, the layers are anchored together by quilting. Basic quilting employs a simple running stitch which can be used to create patterns as simple, or as complex, as the maker’s skill – and time – will allow. It should not always be assumed that a cover was the work of one individual as it is known that large pieces could be made by two or more people. Neither were all quilted pieces made for the home, as many women – and some men – made their livings by quilting. In some old quilts, yet another layer will sometimes be found inside due to the practice of piecing

over papers which was popular in the nineteenth century. Before being stitched together, cut-out fabric pieces were tacked down over corresponding paper shapes, a method which gave sharp definition to edges and thus helped with the neatness and accuracy of the patchwork. Sometimes these papers were removed when the work was finished, but very often they were not. In typically thrifty fashion, the papers were cut out from any scraps that came to hand. Pages from copy books, old bills, letters, and even railway timetables, were used in this way. If such paper templates can be examined, they may sometimes provide handy clues about the age of a piece. The imaginative needlewoman could also add further surface decoration to her work, often in the form of appliqué patterns, embroidery, or beading. In the eighteenth century the technique of corded quilting was frequently used. Narrow channels were produced by stitching through two layers of fabric and thick cord was threaded through these from the back, creating a raised surface pattern which could be remarkably dense and elaborate.

The value of radiography for curation and conservation In many cases quilts and patchwork come into collections with little or no provenance so museum curators

X-radiography of patchwork and quilts 275

and conservators need to be able to explore and understand them as fully as possible. Understanding a single layer of patchwork does not usually present too many difficulties because all of the work can be seen and assessed. However, even a small patchwork or quilted cover may have a complex structure, involving the use of several different materials and sewing techniques. Even when covers are damaged, only limited access may be possible and so tantalising puzzles may remain as it is usually not possible to examine all of the interior. The varied colours and textures of quilts and patchwork can be so visually exciting that details of construction may be obscured. Black and white photography can help by rendering colours in shades of grey but constructional details can still be obscured by patterns and shadows. Radiography shifts the balance in favour of the constructional information. The fundamental elements of the patchwork and quilting – inside and out – become visible and may be mapped and recorded. Such investigation of fillings, concealed seams and other stitching may help to establish materials, structure, techniques and condition. Understanding the hidden elements of a quilt through radiography can help with conservation decision making for display, storage and interventive treatments. For example, knowing whether a filling is worn and weak or continuous could influence whether a quilt is displayed hanging or whether an angled slope is required. Such knowledge could also help in the location of support fabrics and conservation stitching. Equally, knowing the condition and location of paper templates in a patchwork piece could affect decisions as to whether it should be stored boxed or rolled.

Special requirements for radiography of quilts and coverlets An obvious, but significant, challenge when radiographing these textiles is often their size, particularly that of bed quilts and coverlets. The type of X-ray facility most suited for this purpose is that designed for radiography of paintings, particularly where quilt could be supported horizontally with the X-ray tube below and the film placed on top (see Chapter 3, Figure 2d). The X-ray units used in paintings radiography have a kV range suitable for textiles and a good output of low energy X-rays. The only drawback is that these facilities are rarely arranged to be used with unencapsulated film. As a result, the quilts discussed below were all radiographed using film pre-packed

into light-tight paper envelopes.1 This slightly reduced the contrast of the images compared with what might be achieved with unencapsulated film and the image of the paper will be superimposed on that of the textile. However, the images obtained of the quilts were highly detailed and this arrangement proved a satisfactory practical compromise. Each quilt was X-rayed in one, or occasionally two areas, giving individual radiographs of 350 mm 430 mm. Each image is therefore only an indication of the information that might be obtained if full mapping was undertaken; for the practical issues involved in such mapping, see Chapter 3, pp. 43–5.

Information from radiography In every quilt, radiography revealed new information and prompted new lines of enquiry. The images created are very striking and provided a completely new way to view and appreciate the intricate work which had gone into their fabrication, without the distraction of colour, texture or shadow. Fabrics An unusual nineteenth century table cover has multicoloured overlapping clamshell ‘puffs’ around a velvet centre (Figures 22.1a and b). It has two verses worked in a variety of beads. One is the first two verses of the original 1857 version of Bonar’s hymn ‘Thy Way, Not Mine, O Lord’ and the other is an anonymous poem starting: Lonely! No: not lonely While Jesus standeth by His presence always cheers me I feel that he is nigh An almost identical poem appears on an 1884 tombstone in Donaghcloney, Ulster, suggesting a common source. The different weaves used for the clamshells have produced distinctive images (Figure 22.1c). Selvedges have produced bright outlines on clamshells made from ribbons. The wave pattern created by the blue and white fabrics on this doll’s bed patchwork quilt completely disappears in the radiograph (Figures 22.2a and c). The hexagon edges appear white because the fabric is scrunched up by the neat, tight stitching. The brighter thread visible running vertically down the top and left hexagons is the thicker thread which marks the boundary between the lighter and darker shades in the blue fabric printed with black fruit motifs

276

Case studies

(a)

(b)

(c)

Figure 22.1 Beaded table cover, (a) photograph, central area, (b) photograph, detail showing clamshell ‘puffs’ (c) radiograph, detail showing clamshell ‘puffs’ and beads. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

X-radiography of patchwork and quilts 277

(a)

(b)

(c)

Figure 22.2 Hexagon patchwork doll’s bedspread, (a) photograph, overall, (b) photograph, detail, (c) radiograph, detail of four hexagons. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

278

Case studies

(Figure 22.2c). The double stripes of eight thicker threads in the plain weave white fabric, which are rather masked by the printed floral design, are brought to prominence in the radiograph (Figures 22.2b and c). Paper Radiography has the potential to provide information about paper templates in patchwork (O’Connor and Brooks, 2005). Different qualities of paper may be detected as a result of variations in their thickness and mineral inclusions. Research so far undertaken indicates that carbon-based inks are not going to be picked up by radiography. However, metal-based inks, such as iron-gall inks, should be visible and, in some circumstances, these might be better investigated with transmitted light (Thomson and Halliwell, 2005). Dyes In some cases, mordants, dyes or pigments may show up on radiographs. The quilt worked by Mary Burnett has an outer cover of two different fabrics, both with red grounds. The fabric on one side is printed with large flowers, including daisies. The fabric on the other side has a lily and pansy design with patterns of small yellow dots clustered around groups of circles of increasing sizes (Figures 22.3a and b). A large tear in the daisy printed fabric allows the next layer of this complex object to be seen (Figure 22.3c). The detail of the radiograph is from this area (Figure 22.3d). Images produced by different colourants from the fabrics of the outer cover are visible, including parts of the floral motifs from the daisy print and the small dots from the lily and pansy print. Fillings Another feature of Mary Burnett’s quilt is the dappling produced by the uneven distribution of the filling (Figure 22.3d). There are randomly scattered bright, often curled, hard-edged fragments within the filling. These are the remains of cotton seeds which may still be covered with short seed hairs, so producing a light halo around the fragments on the radiograph. Although individual cotton fibres cannot be distinguished on the radiograph, the presence of these seed coat fragments can be taken to confirm that the filling is cotton rather than, for instance, wool or polyester. Feathers and polyester have characteristic images on radiographs (see Chapter 8, p. 113). Figure 22.4, p. 114 shows curled down feathers and

short lengths of hollow rachis from straighter feathers in a 1950s eiderdown. Construction and stitching Information about construction can be obtained through radiography but these images do require careful analysis to obtain useful information. New data about the construction of individual sections becomes visible, as with the method used to make the clamshells on the beaded table cover (Figure 22.1c). Those made from fine fabrics are folded double. In contrast, those made from heavy fabrics have narrow turned hems. The double threads securing the clamshells to the backing fabrics can be seen. Tantalising glimpses of vertical and diagonal seams are visible through the rows of clamshells. Other stitching threads show how the beads are attached. Log cabin patchwork is made up of narrow strips pieced to make a square around a central small square; these blocks are then joined together to create alternating light and dark areas (Sheppard, 2005: 149). This log cabin quilt, initially thought to be hand stitched, contains a riot of colours and patterns (Figure 22.5a; see Seward, 1987: 62). The radiograph enables new understanding of the way the quilt was put together (Figure 22.5b). By following the lines of the stitching between the blocks and mapping displacement of corners, it is possible to determine the stages of construction. As is normal, four log cabin blocks were first joined together into a larger square which was then sewn to other squares of four blocks to form strips. Finally, the strips were joined together with seams running across the quilt’s width. All these seams were machine sewn except those between the individual blocks of the squares along the bottom row; in the radiograph, this row is at the top of the squares, the rest of which were beyond the edge of the radiograph. These seams were hand sewn and the blocks are larger than the others (Figure 22.5c). This seems to provide evidence that more than one person made this quilt. This large and elaborate bedcover (Figure 22.6) combines crazy patchwork, chevron panels, embroidery and appliqué with a heavy wadding and decorative lining (Seward, 1987: 50). At first sight, it appears to have been entirely hand sewn (Figures 22.6a and b). However, all the construction stitches and raw edges are visible in the radiograph showing that the crazy patchwork and chevron panels were actually stitched together by machine (Figure 22.6c). It also reveals the presence of intensive zigzag stitching inside the cover which was not otherwise apparent. Given the thickness of wadding and interlining

X-radiography of patchwork and quilts 279

(a)

(b)

(c)

(d)

Figure 22.3 Mary Burnett’s quilt, (a) photograph, daisy print face showing the corner of quilt which was radiographed, (b) lily and pansy print face; note the dots and circles which are most visible along the top and left-hand sides, (c) photograph, detail showing tear, (d) radiograph, detail in area of tear. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

280

Case studies

inside, this stitching was probably used to control and anchor the various layers. The line of featherstitch embroidery visible in the radiograph is worked on the quilt’s lining, not on the face.

Although the creasing is visible on the radiograph of a Durham quilt, the line of quilting stitches is more readily apparent on the radiographic image than on the quilt (Figures 22.7a and b). The difference

(b)

Figure 22.4 Radiograph of 1950s eiderdown showing feather filling (Private collection © Sonia O’Connor, University of Bradford.)

(c)

Figure 22.5 Log cabin coverlet, (a) photograph, detail, (b) radiograph, detail of the same area, (c) radiograph, detail of the junction between four blocks. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

X-radiography of patchwork and quilts 281

between hand and machine sewing is also very apparent. The machined seams have very regular stitch spacing. Two threads can be seen on either side of the seam, crossing to form a lock stitch. Radiography is particularly helpful in understanding complex pieces such as Mary Burnett’s quilt (Figures 22.3a and b). This much-used quilt has been cut into two. The striking red floral print

(a)

covers are worn and damaged. A range of white fabrics, some printed in blue checks, delicate florals and abstract patterns and a cotton wadding, can be glimpsed through tears in the top covers. Quilting seen in these fabrics does not relate to the quilting of the red covers and establishes that these were not just wadding but a recycled quilt. Radiography revealed details of construction, both quilting

(b)

(c)

Figure 22.6 Bedcover, (a) photograph, (b) photograph, detail, (c) radiograph of the same detail. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

282

Case studies

(a)

(b)

Figure 22.7 Durham quilt, (a) photograph, detail, (b) radiograph, detail. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

patterns and the distribution of the cotton wadding. The amount of information in the radiograph made it hard to read (Figure 22.8a). Overlaying coloured lines on the digitised image enabled the mapping of the seams and both quilting patterns (Figure 22.8b) (Brooks and O’Connor, 2005; O’Connor and Brooks, 2005). Looking at edges can be rewarding. The radiograph of the doll’s patchwork quilt shows that the construction of this edge is not made of part-hexagons, as might be expected, but is a continuous strip of fabric (Figure 22.2c). A line of running stitch is sewn around the entire quilt through the patchwork and the lining. The delicately stitched signature on Elizabeth Watson’s 1819 quilt stands out more clearly on the radiograph than on the quilt itself (Figures 22.9a, b and c). The quilting of additional foliage and outlining around appliquéd motifs is also more evident. Knots indicating the starting points for the threads can be seen along these stitch lines. The narrow seam allowances and clever manipulation of the appliqué fabrics testifies to Elizabeth’s considerable needlecraft skills. These are particularly evident in the image of the dogtooth appliqué. The section of

the circle, about a third of the circumference, captured on the radiograph is a single piece of straight fabric neatly pleated to form the curve. Making and makers Radiographs can help provide information about methods of making. The radiograph of the quilt initialled ‘YGF’ shows how the quilted design is superimposed over the piecing layout with relatively little correspondence (Figures 22.10a and b). However, the diamond pattern is not formed by intersecting straight lines but by zigzags of various lengths (Figure 22.10c). Following the longest stitch line to its turning point gives an indication of the maximum depth of the area worked at one time. This suggests the quilting of this large bedcover was made easier by rolling it onto a frame and that the length of the longest stitch line is the width of the working surface exposed when the quilt was being worked on the frame. It may be possible to identify stitching lines made by different hands (Brooks and O’Connor, 2005: 172–174). The radiograph of the small cord quilted oblong with embroidered floral motifs from the late seventeenth or early eighteenth century provides clear

X-radiography of patchwork and quilts 283

(a)

(b)

Figure 22.8 Mary Burnett’s quilt, (a) radiograph of the corner in Figure 22.3a, (b) radiograph overlaid with coloured lines indicating the different stitching. Red internal seams; green stitching visible on outer covers; yellow hidden quilting of original coverlet, reused as wadding. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

information about the quality of embroidery, stitching and the cord filling (Seward, 1987: 144–145). It shows that two techniques were used to produce the corded background although they both give the same visual effect (Figures 22.11a and b). The majority of the curving lines forming the background were created with a continuous cord, threaded between parallel lines of stitching. The image shows that this cord passes continuously from one row to another and sometimes moves on from one area to the next. However, one side of the quilt has a small area, about a fifth of the whole piece, where the floral motifs are incomplete. Here, the curved motifs are worked with short, discontinuous cords. A full radiographic survey confirmed that the change of technique in this area indicates the work of a second person completing the quilting.

Alterations, damage and repairs Damage to quilts can sometimes be located more quickly and recorded accurately using radiography. For example, the radiograph of the doll’s quilt shows four complete hexagons with a stitch line running

inside the seams which join the hexagons on the far right to the inner ones (Figure 22.2c). It is apparent that this stitching is missing in the upper section of the top hexagon. Surviving early textiles are often fragments salvaged from larger pieces, kept because of the appeal of their age and workmanship. The cord quilted oblong appeared to be one such piece. Its current ribbon border and reused silk backing are probably nineteenth century additions (Figure 22.11c). These have the effect of completely concealing the edges and back of the work, so making it impossible to determine whether the piece was entire or a fragment from a larger textile without unpicking the ribbon edging. Could it have originally been part of a garment or was it always intended to be a flat oblong? Radiography established that the oblong is not a fragment cut down from a larger piece and that the current shape is the original size (Figure 22.11d). The layout of the design shows that edges, now hidden under the ribbon, are not cut edges but are the original edges of the quilted piece. This piece was intended to be the size that it now is, and thus was probably a decorative domestic item. Whatever its function, it was clearly used. Repair stitches, showing

284

Case studies

(a)

(b)

(c)

Figure 22.9 Elizabeth Watson’s quilt, (a) photograph, detail of central appliqué motifs and stitched designs, (b) photograph, stitched signature, (c) radiograph, stitched signature. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

up with a bright fuzzy outline along sections of the cording, secure the exposed cord in areas of wear.

Benefits of radiography Overcoming the challenge of radiographing large, sometimes heavy and unwieldy quilts and coverlets is amply justified by the wealth of information that can be obtained. Data about fabrics, fillings, stitching, construction, alterations, additions and condition

can be systematically and accurately documented and recorded. There is huge potential for research using such evidence to understand more about stitching techniques and, possibly, to identify the work of different makers. Radiography can also be used to enhance public understanding and enjoyment of quilts. In the 2005 exhibition Through the Needle’s Eye at York Art Gallery, eight quilts were displayed with their radiographs. These were selected to show the very different types of information that radiography reveals about these

X-radiography of patchwork and quilts 285

(a)

(c)

complex and intriguing textiles (Sheppard, 2005). The radiographs added a new dimension to the display which clearly engaged and fascinated the public (Figure 22.12).

(b)

Figure 22.10 Quilt initialled ‘YGF’, (a) photograph, detail, (b) radiograph of the same area, (c) radiograph showing turning points of the zigzag quilting. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

Radiography adds to the body of knowledge about quilting and patchwork history and techniques, providing new information which benefits curators, conservators, collectors and visitors as well as creating

286

Case studies

(a)

(b)

(c)

(d)

Figure 22.11 Oblong panel, cord quilting and embroidery, (a) photograph of corner radiographed, (b) radiograph, detail showing two different cording techniques, (c) photograph, detail of corner showing ribbon, (d) radiograph, same detail. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

X-radiography of patchwork and quilts 287

National Museums Liverpool, and Alan Phenix, Senior Lecturer, Conservation of Easel Paintings then at, University of Northumbia.

Note 1.

Facilities used included those at the Conservation Unit, School of Arts and Social Sciences, University of Northumbria, Newcastle upon Tyne (Kodak Industrex AA400 ReadyPack film), and the Conservation Centre, National Museums Liverpool (Agfa Structurix D7 Daylight Wrapped film).

References

Figure 22.12 Visitor examining radiograph at the 2005 exhibition Through the Needle’s Eye, York Art Gallery. (© Sonia O’Connor, University of Bradford; reproduced by permission of York Museums Trust, York Castle Museum.)

images which are beautiful in their own right and which – hopefully – will be an inspiration to contemporary makers.

Acknowledgements The authors would like to thank York Museums Trust, particularly Caroline Worthington, Curator of Art, York Art Gallery. Thanks are also due to Nicola Christie, Head of Paintings Conservation,

Brooks, M. M. and O’Connor, S. A. (2005). New insights into textiles. The potential of X-radiography as an investigative technique. In Scientific Analysis of Ancient & Historic Textiles, Informing Preservation, Display and Interpretation. Post-prints of the AHRB Research Centre for Textile Conservation & Textile Studies, 13–15 July 2004 (R. Janaway and P. Wyeth, eds), pp. 168–176, Archetype Press. O’Connor, S. and Brooks, M. M. (2005). Making the invisible visible: the potential of X-radiography as an investigative technique for textile conservation decisionmaking. Pre-prints of the 14th Triennial Meeting of the ICOM Conservation Committee, The Hague 12–16 September 2005 (I. Verger, ed.), pp. 952–960, James & James (Science Publishers). Seward, L. (1987). The Complete Book of Patchwork, Quilting and Appliqué. Mitchell Beazley. Sheppard, J. (2005). Through the Needle’s Eye. The Patchwork and Quilt Collection at York Castle Museum. York Museums Trust. Thomson, K. N. and Halliwell, M. (2005). An initial exploration of the benefits of using transmitted visible light and infrared photographs to access information concealed within multilayered textiles. In Scientific Analysis of Ancient & Historic Textiles, Informing Preservation, Display and Interpretation. Post-prints of the AHRB Research Centre for Textile Conservation & Textile Studies, 13–15 July 2004 (R. Janaway and P. Wyeth, eds), pp. 177–184, Archetype Press.

23 Revealing the layers: The X-radiography of eighteenth century shoes at Hampshire County Council Museums and Archives Service Sarah Howard and Robert Holmes

Introduction June Swann pioneered the understanding of shoes and their structure through her extensive research and deep knowledge of shoe manufacturing, construction and the impact of wear (Swann, 1982). Visual inspection of shoes can be enhanced by radiography which offers an additional route to understanding these complex three-dimensional, mixed-media artefacts. This chapter will examine the radiography of ten pairs of eighteenth century shoes from the historic dress and textiles collection belonging to Hampshire County Council Museums and Archives Service (HCCMAS). The process of radiography and the facilities available are briefly discussed and the results of the radiography of the shoes examined. This demonstrates how this form of in-depth study can enhance knowledge of textile techniques and construction and assist in preventive conservation measures by providing thorough knowledge of the condition of objects. The use of X-ray imaging as a means of informing conservation work is not new at HCCMAS. Radiography has been used in the past on a teddy bear from the historic dress and textiles collection in order to reveal the ‘growler’ mechanism present inside. However, the majority of use has been for the examination of archaeological material, mainly metals and occasional organic remains such as bone or wood. Natural history specimens have also been examined, both fresh and mounted specimens giving good results. In recent years, the reinstatement of the HCCMAS X-ray facility, after a change in location, and the results of research carried out by the Arts & Humanities Research Council Research Centre for Textile Conservation and Textile Studies at the Textile Conservation Centre, University of Southampton, 288

led to the decision to experiment with X-ray imaging to see what results could be achieved with the textiles collection.

Radiography at HCCMAS The X-ray facility at HCCMAS consists of a Faxitron 43805N built by Hewlett Packard in 1978 and bought as new in 1979. This is the single cabinet type with facility for a fluorescent screen. The maximum exposure area is about 380 mm across the diagonal, which obviously restricts the size of object that can be X-rayed. The cabinet itself is 380 mm deep, 460 mm across and about 320 mm high. The machine is located in an instruments room and is available for use by conservation staff, although only a limited number are trained and allowed to operate it. A purpose-made bench was constructed to take the weight of the machine and give working space on either side. Considering its age, the machine has given extremely good service with very little maintenance beyond a regular ‘health check’ and radiation survey. The longevity of the tube may perhaps be explained by a strict warm-up procedure and careful use. The X-ray tube energy is continuously variable between 0 and 130 kVp. In normal operation the tube current (mA) is not adjusted but the exposure is controlled by a manually operated timer. This makes the operation of the unit very simple but the analogue controls and read-outs for the kV and time make it difficult to repeat short exposures accurately. An automatic exposure system is available but is rarely used as it has not proven particularly suited to the type of material in the collections. In general, the machine seems to give good results

The X-radiography of eighteenth century shoes 289

with relatively thin materials, short exposures and low to mid-range kVp.

Films and processing It is not possible to use unencapsulated film with this X-ray unit as it is housed in an area that cannot be blacked out. Pre-packed film is used exclusively as it is convenient to handle and, in most cases, the paper of the light-proof envelope does not noticeably interfere with the exposure at the kV levels used in this project. Initially, Kodak Industrex ‘C’ or ‘CX’ film ReadyPack was used. At present Agfa Structurix D4 Daylight Wrap (DW) Ready Pack film is preferred, mainly in size 180 mm 240 mm as larger sizes are very expensive and would only be useful very occasionally. Ready Pack films also have the advantage that, left unopened after exposure, they can be stored for a short period and developing can take place as a batch, thus saving darkroom set-up time. Plates are processed individually by hand in open dishes using the manufacturer’s recommended chemicals and protocols. In the past, examination of radiographic plates was restricted to light box, magnifier, or low power microscope. With the advent of digitisation, it has been possible to scan radiographic plates onto a computer, thus enabling manipulation of the resulting image in Adobe Photoshop to reveal even more detail.

Selection of shoes for radiography Experiments were conducted to perfect exposures for different textile object types. After initial tests, it was decided to concentrate on radiography of the collection of eighteenth century shoes as a means of discovering and confirming techniques of production. Although the shoes did not require interventive conservation work, additional information about their construction and materials would enhance future study as well as reducing the need for excessive handling in their examination. Their small size also meant that they would be easy to radiograph in the HCCMAS X-ray machine. Radiographs of ten pairs of shoes dating from 1720 to 1790 were taken. All the shoes are constructed with textile uppers and leather soles, typical of shoes of their time when fabric was often the main choice for those who could afford them (Swann, 1986). It was hoped to discover information about the construction of heels such as whether metal stilettos were present and what materials were used in their composition.

The minimal handling needed to produce the X-rays was advantageous with these fairly fragile items. Low kV levels proved best, typically 30 or 40 kV for 1 minute. The results of this radiography have been very interesting and it is planned to examine more of the historic dress and textiles collection in the future. The following information has been revealed through these initial experiments with the shoes.

Construction of heels Revelations about the materials used to construct the heels of these shoes was a major benefit of using radiography. These typically have a covering, often of leather, so it is not always possible to determine their core. The radiograph of a pair of shoes from c. 1780s with ribbed silk and metal thread fabric uppers and leather soles shows the grain of the wood used to construct the heel (HCCMAS ACM1951.13; Figures 23.1a and b). This appears as a series of concentric rings in a horizontal direction, where the grain of the wood will be at its strongest (Figure 23.1c1). Figure 23.2a shows one of a pair of shoes, c. 1700–1730, made with silk fabric uppers and leather soles (HCCMAS HCMS1968.50.1&2). These have matching clogs2 made of leather with silk fabric latchets3 (HCCMAS HCMS1968.50.3&4). The radiograph of the clog reveals the wooden understructure hidden beneath the leather (Figure 23.2b). It is possible to see that the wood has been cut across the grain and the tree rings are seen in cross-section as vertical lines. The use of wood cut in this direction may have provided greater flexibility for the clog when worn. This wooden structure extends along almost the entire length of the clog, mirroring its overall pointed shape and ending approximately 100 mm from the edge of the leather sole surrounding it. Evidence of wood in the heel cannot be seen in the radiograph, suggesting that it does not extend that far. The image of the wood used for these clogs has the characteristics of a coarse grain softwood, for example alder. It is hoped that further research will be able to identify the types of wood used. It is interesting to note that the radiographs showed that none of these shoes had heels made from a stack of leather. Nor was there evidence of any pegs or stilettos inside the heel as some written sources suggest were present in eighteenth century shoes (Grew and de Neergaard, 1998; Swann, 1986). The heels in every pair of shoes radiographed are made of wood.

290

Case studies

(a)

(c)

(b)

Figure 23.1 A pair of fabric and leather shoes from c. 1780s (Hampshire County Council Museums and Archive Service; Accession No. ACM1951.13), (a) photograph of both shoes and (b) radiograph of the left shoe, taken from above, (c) detail of radiograph digitally adjusted to highlight the bracing stitches. Note also nail holes from the last, radiating rings of the wood grain and the crack in the wood. (Radiography by Robert Holmes; © Hampshire County Council Museums and Archive Service; image adjustment by Sonia O’Connor, University of Bradford.)

Stitching Although stitching is evident when examining the shoes by eye, it is interesting to see the extent of the stitching revealed in the radiographic images. The images of the shoe in Figures 23.2d1 and e show the different types of stitching present. The

stitching at the sides of all the shoes radiographed appears as a decorative chain on the image because the thread, which is intertwined when the sole, insole, rand4 and upper are stitched together, is highlighted in its entirety. These stitches are the edge/flesh and grain/flesh stitches used to attach the sole to the upper and the upper and rand. The images reveal the neatly

The X-radiography of eighteenth century shoes 291

(a)

(b)

(c)

(d)

(e)

Figure 23.2 One of a pair of shoes with clogs, c. 1700–1730, made from silk fabric uppers and leather soles (Hampshire County Council Museums and Archive Service; Accession No. HCMS1968.50.1&2) with matching clog (Hampshire County Council Museums and Archive Service collection; Accession No. HCMS1968.50.3&4), (a) photograph of shoe and clog, (b) radiographic views of clog, from above and from the side, (c) radiograph of shoe from above, (d) detail of radiograph digitally adjusted to highlight constructional detail, (e) radiograph of shoe taken from the side. (Radiography by Robert Holmes; © Hampshire County Council Museums and Archive Service; image adjustment by Sonia O’Connor, University of Bradford.)

292

Case studies

(a) (a)

(b)

(c)

Figure 23.3 One of a pair of fabric and leather shoes, c. 1700–1740, made with silk uppers and leather soles (Hampshire County Council Museums and Archive Service; Accession No. C1980. 49.1&2), (a) photograph, (b) radiograph of shoe showing the bracing stitches, (c) detail of radiograph digitally adjusted to highlight the layers in the sole and zig-zag bracing stitching. (Radiography by Robert Holmes; © Hampshire County Council Museums and Archive Service; image adjustment by Sonia O’Connor.)

(b)

(c)

Figure 23.4 A pair of clogs, 1730–1760, made with leather soles and latchets (Hampshire County Council Museums and Archive Service; Accession No. CRH1973.8.2), (a) photograph, (b) radiograph of clog, one from above and one from the side, (c) detail of radiograph digitally adjusted to highlight the pin. Note also the structure of the metal braid and other textiles. (Radiography by Robert Holmes; © Hampshire County Council Museums and Archive Service; image adjustment by Sonia O’Connor, University of Bradford.)

The X-radiography of eighteenth century shoes 293

stitched, finely crafted shoes of this period, demonstrating the shoemaker’s skill acquired over many years. Another type of stitching is also revealed by the Xrays which cannot be seen with the eye. These are the criss-cross of bracing stitches which hold together the upper and insole from one side of the shoe to the other prior to the sole being attached (Vass and Molnar, 1999). Figure 23.3a shows one of a pair of shoes, c. 1700–1740, made from silk fabric uppers and leather soles (HCCMAS C1980.49.1&2). In the radiograph of this shoe, long straight zigzag lines are visible across the sole (Figures 23.3b and c1). These are visible in Figure 23.1c1 and were also observed across the top piece (the sole) of the heel in Figure 23.3b.

General construction It is interesting to note distinct, hidden features of construction. The pair of 1730–1760 clogs in Figure 23.4a is made with leather soles and latchets (HCCMAS CRH1973.8.2). The radiograph reveals their wooden understructure and its grain (Figure 23.4b) as well as a pin hidden at the side of the latchet which cannot be seen by eye. The pin’s function is uncertain but it probably holds together elements inside the latchet (Figure 23.4c1). It is contemporaneous in date to the shoe because the round head of the pin is consistent with pins made in the early eighteenth century. In Figures 23.2c and d1 it is possible to detect the outline of an additional material along the waist of the shoe where the arch of the foot would sit. Initial research suggests this could be a leather shank placed over the insole, extended to the waist. This would have been applied before the sole was attached, under the bracing stitching, in order to provide strength to the shoe and to prevent undue bending at this point.

Conclusion This radiography has proved highly informative. Hidden information about the construction of these

shoes has been detected, enabling a detailed study of shoemaking techniques of the time. Although the work was not intended to inform any immediate interventive conservation work, a thorough knowledge of any object to be conserved is important and these radiographs will help should such a treatment be necessary in the future. The information gained does help, however, with the general preservation of the shoes because obtaining this knowledge in any other way would require excessive handling and manipulation which could place the shoes under great pressure. This new knowledge will be included with the documentation accompanying each shoe and will be available to anyone wishing to study the shoes in depth. Further research, using the digitised images on computer, is planned to shed light on some of the earliest items of dress in the HCCMAS historic dress and textiles collection.

Notes 1. 2.

3. 4.

Digital manipulation was carried out by Sonia O’Connor, University of Bradford. Clogs are overshoes made either from a platform of wood or a stack of leather soles. The delicate shoes whose uppers were made from fabric were not hardwearing and, for outdoor use, such overshoes, often made from matching material, were worn. A latchet is the strap extending across the front of the clog. A rand is a narrow strip of leather used to make the join between the sole and upper more waterproof.

References Grew, F. and de Neergaard, M. (1988). Shoes and Pattens. Her Majesty’s Stationery Office. Swann, J. (1982). Shoes. Batsford. Swann, J. (1986). Shoemaking. Shire Publications. Vass, L. and Molnár, M. (1999). Handmade Shoes for Men. Könemann.

24 The contribution of X-radiography to the conservation and study of textile/leather composite archaeological footwear recovered from the Norwegian Arctic Elizabeth E. Peacock

Introduction In the summers of 1955 and 1960 inter-Nordic archaeological/ethnographical research expeditions were undertaken on the island of West Spitsbergen, the largest island in the Svalbard archipelago in the European Arctic (Figure 24.1). The purpose of these expeditions was to investigate whether Stone Age and medieval settlements existed on Svalbard before 1596 when the islands were officially discovered by the Dutch explorer Barents. A further objective was to study the sites of whaling and hunting stations from the seventeenth and eighteenth centuries. On both occasions most of the expedition time was spent excavating and recording the Russian Pomor1 hunting station at Russekeila (Figure 24.2). Artefacts recovered from these investigations were sent to the Tromsø Museum in Tromsø, Norway. In the 1980s, the textiles from Russekeila, including textile/leather composite cold weather footwear, were sent for conservation to the Vitenskapsmuseum, at the Norwegian University of Science and Technology (NTNU), Trondheim. The application of radiographic techniques in the initial examination of this footwear proved vital, not only in assessing and documenting its condition and technology but also in developing an appropriate conservation strategy. Almost 50 years later, rescue excavations were carried out at Kapp Wijk, West Spitsbergen, another hunting station of Russian Pomor origin. All the recovered artefacts, including severely deteriorated, wet textile/leather composite footwear were also 294

conserved at the Vitenskapsmuseum. The Russekeila radiograph archive proved invaluable in the investigation of the technology of this culturally similar but poorly preserved footwear, emphasising the importance of such an archive and the analytical potential of radiography in archaeological textile studies.

History of Russian Pomor hunting activities on Svalbard The finds from Russekeila and Kapp Wijk represent the indigenous folk culture of eighteenth century Northwest Russia. The discovery of Svalbard in 1596 by the Dutch explorer Willem Barents initiated a long period of extensive hunting and exploitation in the Svalbard area. Whales and walrus were heavily culled by large hunting expeditions, particularly from England and the Netherlands. Russian Pomors from areas along the White Sea hunted extensively on the islands, primarily between 1700 and 1850.2 Unlike the West European whaling expeditions that hunted on Svalbard in the summer months, the Russian Pomors remained in the area throughout the winter. They were the first to winter on Svalbard on a large and organised scale. Their expeditions originated in the White Sea area of Northwest Russia and consisted of hunting groups of twenty to thirty men. They mainly hunted walrus, but also took seal, various bird species, arctic fox, Svalbard reindeer and polar bear. Their buildings, often prefabricated and transported from Russia, were constructed as log cabins with plank flooring and clay-covered plank

The contribution of X-radiography 295

Figure 24.1 Map of the Arctic showing the location of the Svalbard archipelago. (© G. H. Turner-Walker.)

roofs and functioned as both living and working quarters (Simonsen, 1957; Christiansson et al., 1967). They gradually withdrew from the Svalbard hunting grounds in the latter half of the nineteenth century (Blake, 1961).

The burial environment on West Spitsbergen, Svalbard The excellent – and well-publicised – preservation of artefacts and mummies discovered in arctic regions or conditions has given the misleading impression that all such finds from similar regions and burial conditions are comparably well preserved (Chamberlain and Pearson, 2001; Hart Hansen, 1998). Artefacts and human remains which become incorporated into the permafrost layer are well placed for potential preservation. Although the Norwegian Arctic provides for near year-round frozen conditions, the short summer thaw can be accompanied by aggressive botanical growth. This can be especially destructive, both physically and chemically, to organic artefacts in the often shallow and exposed cultural levels. On Svalbard the permafrost is 10–40 m thick at the coast, but the cultural levels at the archaeological sites are relatively shallow and may lie within the active soil

Figure 24.2 Map of Svalbard with the locations of two eighteenth century Russian Pomor hunting stations based on Isfjorden on West Spitsbergen: Russekeila and Kapp Wijk. (© E.E. Peacock.)

layer which goes through a period of successive freezing and thawing events each summer season. These conditions result in a number of factors that cause degradation to artefacts. The burial environment is wet but not anoxic. Artefacts, if not still water saturated from the previous season, become so. Watersoluble tannins, oils and fats have a tendency to leach out of leather leaving the water-saturated collagen fibres susceptible to denaturation by watercatalysed hydrolysis. In the long term, daily freeze/ thaw cycling during the pre-summer through to the post-summer seasons results in ice crystal damage which reduces tensile strength by puncturing the fibres in both textile and skin materials. The close proximity of the sites to the sea, combined with sea flooding during winter storms, leads to artefacts becoming contaminated with soluble salts. These salts can bring about increased swelling of organic material when wet and physical breakdown upon drying. Micro- and macro-organisms and plants are very active during the short summer months. Micro-rootlets grow into soft organic materials such as textiles and leather spreading throughout the three-dimensional fibre and yarn networks. This type of damage was reported by Logan (1983) from the excavations at Red Bay, Labrador, Canada.

296

Case studies

The Russekeila site The Russekeila site (78°N, 14°E) is situated just west of Barentsburg on the south shore of Isfjorden, the largest fjord on the west coast of West Spitsbergen. At the time of the joint Norwegian/Swedish excavation, the ruins of Russekeila were located near the edge of a flat, vegetation-covered raised terrace seven metres above sea level and 16 metres inland from the hightide line. The edge of the terrace sloped steeply down towards the beach and the part of the site facing this was covered with a thick layer of gravel and cobbles from winter storm activity. The cultural remains consisted of the exposed decayed wooden ruins of a multi-roomed building and several huts covering an area of 200 m2 from an eighteenth century Russian Pomor hunting station, a kitchen midden (refuse heap) located on the side of the site sloping down to the sea, and a mass grave about 200 metres south-east of the ruins. The intact cultural layers inside the ruins were relatively shallow at 10–30 cm thick, while the midden ranged from several centimetres to 1 metre thick.

The artefacts and their recovery Approximately four thousand artefacts representing a wide range of material groups were recovered during the excavations. The artefacts recovered in 1955 were reported to be in an excellent state of preservation and that most of them simply had to be freed from the frozen ground (Simonsen, 1957). The 1960 excavations of the kitchen midden encountered permafrost at a depth of 100–200 mm (Müller-Willie, n.d.) but the report of the excavation of the mass grave site (Christiansson et al., 1967) does not comment on the condition of the ground. It seems, however, that much of the material at Russekeila was preserved in permafrost conditions.

The footwear recovered at Russekeila A large amount of footwear and associated leather and textile offcuts was recovered during the excavations at Russekeila (Christiansson, 1956). The textile/leather composite footwear which is the subject of this discussion was among the artefacts that made up the Svalbard Textile Conservation Project (1982 to 1986), several aspects of which have been described elsewhere (Peacock 1990a, 1990b and 1987). This material was predominantly associated with cold weather

Figure 24.3 Textile/leather composite archaeological cold weather work shoe (TS 6155yy) recovered from excavations of the Russian Pomor hunting station at Russekeila, West Spitsbergen, Svalbard, in 1960. (© E.E. Peacock.)

work footwear although several shoe styles were represented among the textile/leather composite finds. Leather makes up the supporting framework of toe, heel and sole with the remaining upper built up of several layers of textile fabrics (Figure 24.3). There is no record of on-site treatment or packing of the artefacts. When these finds were collected from their depository for conservation twenty years later, they were either packed in brown paper bags or lying open in cardboard boxes in wooden storage drawers. The textile/leather composite footwear was heavily contaminated with sand and gravel and with soil, organic and particulate matter. Moreover, it was distorted but not collapsed, inflexible and fully dried out. This was especially the case with the composite shoes, which had a leather sole, heel stiffener, cap and vamp. Based upon the nature of the site, it is assumed that this material was water-saturated when excavated and had subsequently dried out. Despite this, their condition was relatively good, considering the shoes were well worn from both a hard service life and the harsh environmental conditions of the Norwegian Arctic. They were free of hobnail soles and other metal elements, such as fastenings; consequently, there were no corroding parts or corrosion products to address or stabilise. Much of the connective stitching was still in place. Although humble in appearance, these shoes were in a stable condition.

Radiography of the footwear Radiography is an established technique in the archaeological conservator’s toolbox where it forms part of the preliminary examination and documentation of archaeological material. Its primary use has been to assess the condition of the artefact before

The contribution of X-radiography 297 Table 24.1 Specifications for radiography of textile/leather composite shoe (TS 6155yy) X-ray View number

Film

Sheet size (mm)

Focus

Focus to kV mA Exposure Comments film distance time (sec) (mm)

696a

Overview Kodak 300 400 (1/2) Normal 560 Industrex C

70

5

30

696b

Overview Kodak 300 400 (1/2) Normal 560 Industrex C

70

5

90

697

Overview Agfa D7

100 235

Normal 440

25

5

30

698a 698b 699

Side view Agfa D4 Side view Agfa D4 Overview Agfa D4

100 290 100 315 100 235

Fine Fine Fine

440 440 320

25 25 25

2 2 2

180 180 35

833

Side view Agfa D4

100 385

Fine

560

25

2

720

834

Overview Agfa D4

100 385

Fine

560

25

2

720

conservation and determine what lies beneath the obscuring corrosion of metal artefacts, especially iron objects, or within lifted soil blocks. In addition, information such as manufacturing techniques, decoration and artefact technology is often revealed. Until relatively recently, artefacts composed of organic materials were rarely the subject of radiographic examination, although skeletal remains are occasionally radiographed to investigate traumas or pathologies. This was partly because organic materials are relatively transparent to X-rays and are unlikely to produce an image at the exposures normally applied to metal artefacts or soil blocks. Since the Svalbard artefacts were being conserved in an archaeological conservation laboratory with in-house film-based X-ray facilities and where radiography of archaeological material is routine, access to radiography was convenient. Previous experience of applying radiography to investigate the contents of excavated soil blocks led to the decision to radiograph the distorted textile/leather footwear to investigate the extent of soil particulate matter and gravel contained within. The radiography was carried out using a small selfcontained low-energy Minishot Dual 150 industrial radiography cabinet system.3 Both Kodak Industrex® and Agfa Structurix® industrial radiographic films in daylight-packaged sheets were used.4 Initially the

Grey, featureless, too bright and too little contrast Better contrast, good overview but no distinct details Film length too short, exposed area diameter too small, good result Good result Good result Film length too short, exposed area diameter too small, too little contrast in result Excellent result, good detail Excellent result, good detail

images obtained were grey and lacking in contrast and detail but, through the systematic use of different films and exposure conditions, it was possible to arrive at a combination which resulted in images of the shoe shown in Figure 24.3 with clear detail and definition (Table 24.1). Optimal exposures were obtained using a fine-grained high-contrast radiograph film (Agfa D4), low tube energies (25 kV), long exposures times (720 seconds) and long X-ray focus-to-film distance (560 mm). Radiographs were taken both under the shoe and with the shoe positioned on its side to reveal the maximum detail and technical information. Two pieces of radiographic film were aligned beneath the shoe and exposed simultaneously to capture a complete overview or side view. Consequently, each pair of films had the same density range. Furthermore, they were developed together in an attempt to avoid differences in the final radiographs due to the manual development regime. One drawback of using several sheets of film is that because the sheet of film is enclosed in a light-tight envelope, it is the two envelopes – and not the two sheets of film – which are aligned below the object. This inevitably resulted in a slight offset when lining up the image on the developed radiographs. Although the radiographic image is two dimensional and the footwear is three dimensional, the resulting superimposed features were not

298

Case studies

difficult to interpret and, in the most successful images, the detail and clarity were excellent.

Results and implications Radiography did highlight the extent of sand and gravel inclusions within the shoes. From the radiographs, it could be seen that the interior of the shoe was not heavily soiled with particulate matter. There was no gravel but a small amount of fine sand evenly distributed and caught between layers of textile as well as at the interface between the outer leather sole, heel and vamp, and the inner textile layers. In addition, it revealed that the textile/leather construction was much more multi-layered than the shoe itself appeared. Stitching holes in the leather sole were revealed, as was the twill weave of one of the textile layers. It was not possible to distinguish between internal layers of leather and textile felt, as both these materials consist of felted fibres with no other distinguishing structural features (Figures 24.4a and b). The clear image of bones from a left foot seen inside the shoe (TS 6155yy) in Figures 24.4a and b was completely unexpected. The bones were in their correct anatomical alignment, taking into account the three-dimensional positioning of a foot in a shoe and the distorted nature of the shoe. They consisted of the fifth metatarsal, all the phalanges and one of the small sesamoid bones5 opposite the metatarsophlangeal joint of the first toe. This latter small bone, which could have been mistaken for a waterworn pebble, although its image density was the same as that of the associated bones, proved especially interesting. Its presence was consistent with the conclusion drawn from the study of the skeletal remains (Christiansson et al., 1967) that many of the features observed were suggestive of a physically hard and active lifestyle. The presence of human remains within the shoe introduced a complex dimension to the developing conservation strategy for the shoe, and the footwear in general (Peacock, in press). Initially, linking the artefacts with the specific area of the site where they had been recovered had not been considered. The discovery of the human remains in the shoe changed this. It was not clear from which area of the site the shoe had been excavated (i.e. huts, kitchen midden, or the mass grave site). Neither Simonsen (1957) nor Christiansson (1956) mention human remains in their descriptions of the 1955 excavations of the hut ruins. Neither does Müller-Willie (n.d.) in his report

(a)

(b)

Figure 24.4 Radiograph of the shoe shown in Figure 24.3 showing, (a) side view (XR 833), (b) overview (XR 834); Agfa D4 industrial X-ray film image, captured on Ilford FP4 BW film with a Hasselblad medium format camera and an Epson 2450 flatbed scanner. (Radiographs: E. E. Peacock, © Vitenskapsmuseum; Photograph: © E. E. Peacock.)

on the 1960 excavation of the kitchen midden. Christiansson does note that two graves at the mass grave site were investigated in 1955 and Christiansson et al. (1967) report that twenty well-preserved human skeletons were recovered in 1960. They add that all

The contribution of X-radiography 299

the buried individuals were covered with coarse sackcloth and that many of them had been buried in their clothing. Christiansson et al. also mention that scattered and broken human bones in the ruins of the huts themselves indicated a further two to three persons had died there. It was concluded that the shoe in question (TS 6155yy) had been recovered either from one of the skeletons in the mass grave or from within the hut ruins. Detailed drawings of the positioning of the skeletons in the mass grave as well as photographs taken at the time of excavation, both now housed in the archives of Tromsø Museum, were carefully studied in an attempt to localise the origin of the shoe. One photograph revealed that one skeleton (grave IIIf ) had the shoes in question on his feet. This then linked the shoe to an unnamed but purposely buried individual. Further investigation to localise the distribution of the remaining textile artefacts led to the study of unpublished material (e.g. reports, photographs and journals) in the expedition’s archive at Tromsø Museum and Tromsø Museum’s collection catalogue. This brought to light the fact that many of the textiles had, indeed, been removed from the buried individuals, but now they are completely disassociated from human remains.

Conservation strategy and implementation The conservation of this material was undertaken in the mid-1980s when it was common practice to subject textile objects (historic and archaeological) to a conservation regime that included wet cleaning. In addition, it was not uncommon for composite objects to be separated into their component materials with each material being treated separately and then reassembled into the object. Certainly, this can be the simplest means of treating composite objects. Consequently, to have separated the textile and leather layers of the footwear, wet cleaned the textiles, lubricated the hard, misshapen leather and then reassembled would not have been an unusual treatment for that era. All the more so had the footwear been regarded as historic, or even ethnographic, because both these fields of conservation then had a more restoration-based approach to conservation. The conservation strategy for this material was for a less invasive approach despite the fact that only one artefact contained human bones. The intact footwear was gently vacuumed both inside and out to remove

adhering soil, vegetable and particulate matter. Where possible, micro-rootlets which had tunnelled their way through layers of textile were mechanically removed with tweezers. Since this corpus of artefacts is a research collection, the intact shoes were housed in clear polystyrene (‘crystal’) boxes for future study. Currently, the collection is housed at Tromsø Museum, but eventually it will be moved to Longyearbyen, Svalbard, to the recently completed Svalbard Museum. Only one textile/leather composite shoe proved to contain human remains. It may well be that there are human remains in some of the leather-only footwear recovered at Russekeila, but this material was not included in the Svalbard Textile Conservation Project. It was beyond the scope of the project to determine whether other of textile/leather shoes could be localised on the drawings or photographs of the mass grave site.

Russian Pomor textile/leather composite archaeological footwear revisited In the summers of 2001 and 2002 rescue archaeological excavations were undertaken by the Environmental Department of the Governor of Svalbard at another eighteenth century Russian Pomor hunting station. The site at Kapp Wijk (78.5°N, 16.5°E) is located on the same fjord, east of Russekeila and north of Longyearbyen (Figure 24.1). Winter storms and aggressive coastal erosion had been breaking down the ruins of this station and this led to the decision to remove the archaeology (i.e. preservation by excavation record). These excavations also recovered a large number of artefacts representing a wide range of material groups, which were sent to Vitenskapsmuseum for conservation (Peacock, 2005). By the summer of 2004, less than two years after the excavations were carried out, the site was flooded and washed into the sea.6 Footwear makes up much of the wet textile and leather finds, with numerous textile/leather composite shoes – none of which is intact (Figures 24.5a and b). Unlike the Russekeila artefacts, which were for the most part preserved in the permafrost, the cultural layers at Kapp Wijk were entirely within the active soil layer. Consequently, the organic materials, especially the textiles and leather, are severely degraded through water-catalysed hydrolysis, salt contamination, ice crystal damage from repeated freeze/thaw cycling, and rootlet infiltration during the short arctic summers.

300

Case studies

(a)

(b)

Figure 24.5 Textile/leather composite archaeological shoe (KW 1270) recovered from rescue excavations of the Russian Pomor hunting station at Kapp Wijk, West Spitsbergen, Svalbard, in 2002. Shoe, (a) prior to conservation, (b) following freeze-drying. (Ina Neese, © Vitenskapsmuseum and Per E. Fredriksen, © Vitenskapsmuseum.)

Few seams are intact, but the sewing thread is preserved in sewing holes along the seam edges. Many of the shoes are in a disassembled, cut-up state in which the textile upper is missing and sections of the leather cap, vamp, heel stiffener and sole have been cut away indicating reuse of materials. The initial appearance of the Kapp Wijk composite textile/leather shoes indicated that they resembled the footwear from the Russekeila site except that some of the heel soles are fitted with iron hobnails which were extensively corroded and breaking up. The Russekeila radiographs were consulted to assist in interpreting the Kapp Wijk shoes, which were in a much-deteriorated and collapsed wet state. This provided the opportunity to digitise both the radiographs and the black and white film images of the radiographs, which were taken in a pre-digital era. Once digitised it was possible to make more detail

visible in the earlier film records through localised contrast adjustment. With traditional viewing of radiograph pairs using a light box one must repeatedly line up and secure the positioning of the radiographs and studying the subject matter becomes a cumbersome affair. This was eliminated as the digitised images were mosaiced together to form a complete image of each shoe. Viewing using a computer screen was also more visually comfortable for studying the radiographs, for zooming in on areas of interest and for scanning the subject matter in much more detail. The radiographs were captured with an Epson 2450 flatbed scanner and the digitised images manipulated using Jasc® Paint Shop Pro 8. Experience with the Russekeila material provided an informed basis from which to analyse and develop a conservation treatment for the footwear from Kapp Wijk. The Russekeila footwear and its radiographs have shed light on the Kapp Wijk shoes in terms of both preservation and construction technology, and vice versa. For example, the layers of textile and leather in the disassociated Kapp Wijk shoes can be more easily investigated than those of the distorted textile uppers in the inflexible, driedout Russekeila footwear. In contrast, the intact, robust Russekeila shoes provided a more complete picture of how the compressed, water-degraded Kapp Wijk footwear might have originally appeared.

Conclusion The application of radiography in the preliminary investigation and documentation of archaeological materials, especially ferrous metals and soil blocks, is routine but its use with soft organic archaeological materials such as textiles and leather has not generally been appreciated. The application of filmbased radiography in the initial examination of the Russekeila textile/leather composite footwear proved vital in assessing and documenting its condition and technology, but also in developing an appropriate and ethical conservation strategy. Twenty years later, the Russekeila radiographs provided an invaluable comparative database in the interpretation of the culturally similar, but less wellpreserved, footwear recovered from Kapp Wijk. This highlights the importance of such an archive and the analytical potential of radiography in archaeological textile studies. Moreover digitising this archive for use with the Kapp Wijk footwear illustrated its long-term and broader potential.

The contribution of X-radiography 301

The digitised archive is first and foremost more user-friendly than the film-based archive. It is much more convenient to be able to pull up multiple images on the computer screen for examination and detailed study, and to share images with distant colleagues. The still wet, water-degraded Kapp Wijk textile/leather composite shoes were in an advanced state of deterioration and could not be manipulated. Reference to the Russekeila photograph and, especially, radiograph archive provided insight into their probable construction and a reminder of their complexity.

Notes 1. 2.

3.

4.

5.

6.

The word Pomor (also occurring as Pomori) is derived from the Russian word pomorje meaning ‘coastland’. It is presumed that the Pomors reached Svalbard around the mid-seventeenth century; however, the question of the timing of the first Pomor discovery of Svalbard still remains to be solved (Hultgreen, 2002: 125–145). The Andrex Minishot Dual 150 X-ray unit has a beryllium window and a maximum output of 150 kV and 5 mA. This cabinet unit has the advantage of integral protective lead shielding and requires no additional radiation protection for the operator. The unit can be used for viewing (i.e. fluoroscopic inspection) as well as for normal film radiography. When used in the inspection mode, the image of the object is produced on a fluorescent screen, magnified through a lens and viewed through a mirror and lead glass window (Andrex Radiation Products AS, n.d.). Both films have an emulsion on both sides of the support film. Kodak Industrex® industrial X-ray film and Agfa Structurix® industrial X-ray film are produced in a wide variety of types. Agfa D7 is a fine grain film with high contrast and high speed. Agfa D4 is an extra fine grain film with average speed and very high contrast (Agfa-Gevaert, n.d.). Gray (1985: 187) writes that sesamoid bones ‘…are small rounded masses,… which are developed in those tendons, which exert a great amount of pressure upon the parts over which they glide. It is said they are more commonly found in the male than in the female, and in persons of an active muscular habit than in those who are weak and debilitated’. Personal communication from K. Prestvoll (2004), Cultural Heritage Advisor, the Governor Office of Svalbard.

References Agfa-Gevaert (n.d.). Technical Data Sheet NETTC USA 01 2002/09. Agfa-Gevaert N.V. Andrex Radiation Products AS (n.d.). Technical Data Sheet 115.2. Andrex Radiation Products AS. Blake, W. (1961). Russian settlement and land rise in Nordaustlandet, Spitsbergen. Arctic, 14, 101–111. Chamberlain, A. T. and Pearson, M. P. (2001). Earthly Remains. British Museum Press. Christiansson, H. (1956). Den kulturhistoriska expeditionen till Spetsbergen 1955. Fornvännen, 51, 286–289. Christiansson, H., Hedegård, B., Lewin, T., Norrby, A. and Sarnäs, K. V. (1967). Skeletal Remains at the Russian Settlement at Russekeila in West Spitsbergen. Göteborgs University. Gray, H. (1985). Gray’s Anatomy (T. P. Pick and R. Howden, eds), Chancellor Press. Hart Hansen, J. P. (1998). Bodies from cold regions. In Mummies, Diseases and Ancient Cultures (2nd ed.) (A. Cockburn, E. Cockburn and T.A. Reyman, eds), pp. 336–350, Cambridge University Press. Hultgreen, T. (2002). When did the Pomors come to Svalbard? Acta Borealia, 19(2), 125–145. Logan, J. A. (1983). Red Bay 1982 – textile discovery. Textile Conservation Newsletter, October, 3–5. Müller-Willie (n.d.). Rysk fångststation vid Russekeila, Vestspetsbergen, Svalbard. Rapport om möddingens utgrävning 1960. (Internal report.) Tromsø Museum. Peacock, E. E. (in press). Archaeological textiles intimately associated with human remains – where is the dilemma? In 9th North European Symposium for Archaeological Textiles (NESAT) Conference (A. Rast-Eicher, ed.). Peacock, E. E. (1987). Anthropological textiles: a mounting solution. In ICOM Committee for Conservation, 8th Triennial Meeting, Sydney, Australia, 6–11 September 1987 Preprints (K. Grimstad, ed.), pp. 413–416, Getty Conservation Institute. Peacock, E. E. (1990a). Freeze-drying archaeological textiles: the need for basic research. In Archaeological Textiles (S. O’Connor and M. M. Brooks, eds), pp. 22–30, United Kingdom Institute for Conservation. Peacock, E. E. (1990b). The Svalbard textile conservation project. In Textiles in Northern Archaeology (P. Walton and J-P. Wild, eds), pp. 195–204, Archetype Publications. Peacock, E. E. (2005). Conservation of severely deteriorated wet archaeological leather recovered from the Norwegian Arctic. Preliminary results. In Proceedings of the 9th ICOM Group on Wet Organic Archaeological Materials Conference (P. Hoffmann, K. Strætkvern, J. A. Spriggs and D. Gregory, eds), pp. 565–578, Deutsches Schiffahrtsmuseum. Simonsen, P. (1957). Fra den første arkeologiske Svalbardekspedisjonens arbeid. Polarboken, 76–84.

25 Controlled lifting and X-radiography of gold threads from ancient archaeological textiles Elizabeth Barham

Introduction Radiography, combined with good lifting techniques on site, can uniquely reveal evidence of ancient, archaeological textiles incorporating metal threads when the organic elements are poorly preserved. The following case studies, taken from two Museum of London Specialist Services (MoLSS) conservation projects, demonstrate the benefits of the noninterventive nature of radiography. They also show that collaboration and liaison between archaeologists and archaeological conservators during excavation can be of great benefit in the recovery and study of these types of finds.

young woman (Figure 25.2a). The organic elements of the garment, which had once held the gold thread together, had not survived burial. These were probably the remains of a garment worn by the woman

The Spitalfields Roman sarcophagus textile finds The first example occurred during the excavation and conservation of the outstanding find of an undisturbed Roman burial of the fourth century AD in a lead coffin and stone sarcophagus, discovered at Spitalfields, London, in 1999 (Figure 25.1). The sarcophagus and its coffin were unearthed during a Museum of London Archaeology Service (MoLAS) excavation, still with their lids and contents in place. The assemblage was lifted intact from site, and brought for more detailed excavation and investigative conservation at the Museum of London (Barham and Lang, 2001). Areas of a very fine loose gold wrapped thread consisting of tight spirals of narrow gold strips less than 0.2 mm wide and 0.002 mm thick, which originally would have been wound around an organic fibre core, were found in approximately 20 mm of wet silty soil at the bottom of the lead coffin, among the skeleton of a 302

Figure 25.1 Spitalfields Roman sarcophagus with the skeleton of a young woman in situ. (© Andy Chopping at MoLAS.)

Controlled lifting and X-radiography of gold threads from ancient archaeological textiles 303

(a)

(b)

(c)

Figure 25.2 An area of lifted gold thread from the Spitalfields coffin, (a) photograph, (b) detail of the gold threads, (c) detail from radiograph showing gold thread. (© Museum of London.)

304

Case studies

buried in the coffin, possibly a dalmatica enhanced with gold thread decoration. No technique could be clearly identified from these fragments. Parallels with better preserved archaeological examples from the same period elsewhere of similar combinations of gold thread, wool and damask silk suggest that the gold thread could have formed a gold tapestry decoration.1 The organic elements of the garment, which had once held the gold thread together, had not survived burial. The silt was systematically excavated. After recording its locations, the areas of gold thread, lying on the base of the coffin and mostly obscured by the silt, were lifted by conservators. The silt was the consistency of a viscous liquid, and thin, flexible polyester (Melinex) sheets cut to size were used as a support when lifting the threads. These were floated underneath the areas of gold, without disturbing the individual threads, by pipetting small amounts of deionised water underneath the Melinex. The areas were then X-rayed while still fresh and damp using an Andrex X-ray cabinet with Kodak Industrex MX125 film at 70 kV, for 36 seconds on a fine focus setting. The kV and exposure time were kept light and fine focus used as the layers of thread were exposed and very thin. The 180 mm 240 mm film was enclosed in a rigid plastic and aluminium cassette. The aim was to try to detect any surviving patterning in the gold. One X-ray plate was taken for each area of gold thread (Figure 25.2b). The Xray images revealed the individual threads clearly (Figure 25.2c) but, unfortunately, the gold thread had been washed around in the coffin during burial and any decorative scheme to which it had contributed had not survived. However, the radiographs did provide a record of the orientation and condition of the gold thread on discovery, in a situation where cleaning away the silt risked breakage and damage to the scant remains of the textile evidence from the coffin. Later, the radiographs provided a guide to the conservation work and further investigation of the threads, which included light microscopy, scanning electron microscopy and electron probe microanalysis.

The Prittlewell Anglo-Saxon chamber-grave textile finds A similar approach was taken during the excavation of an Anglo-Saxon chamber-grave at Prittlewell, Essex, in late 2003. This sensational discovery provided new glimpses into the burial rituals of Anglo-Saxons of the highest social status in the late sixth to early

seventh century AD as well as new information about artefacts in their graves. The chamber was excavated by MoLAS during an archaeological evaluation prior to a road-widening scheme. Lying beneath a grass verge between a road and a railway line, the chamber had remained undisturbed by later human activity up to the point of discovery. On excavation, most of the surviving artefacts were found still lying in their burial positions. Preservation of the metalwork on the site was remarkably good, but very little organic material had survived in the sandy soil over gravel. No skeletal material remained except for some tiny dental fragments. However, a number of metal objects, including a gold buckle, two small gold crosses and two small gold coins, which had been laid in the coffin with the body, were found in situ during the excavation of the area where the coffin would once have been. In the area where the upper torso would have lain, the archaeologist excavating the damp sand uncovered a patch of ground in which tiny gold threads could be seen (Figure 25.3a). After consultation with the site conservator, the supervising archaeologist lifted the threads in a small 100 mm 140 mm 30 mm block of the surrounding damp sand, without disturbing the individual threads. The orientation of the block in the grave was recorded. It was isolated from surrounding sand and then gently undercut with a piece of thin, flat and firm plastic that had been cut to size. The block was then sandwiched with a similar piece of plastic, bound together firmly and densely with masking tape to avoid movement during transport and labelled with its orientation and identification numbering. The small ‘cassette’ of sand was placed in a small rigid polyester box of a similar size to give it physical protection and removed to the conservation laboratory. There it was X-rayed using an Andrex radiography cabinet on 180 mm 240 mm Kodak Industrex MX125 film at 100 kV for 120 seconds. The kV and exposure time were slightly higher in this case because the gold thread was embedded in a dense layer of sand and tape bindings. The film was enclosed in a rigid plastic and aluminium cassette with a lead screen on either side of the film to produce image intensification (see Chapter 3, pp. 41–3). The resulting radiographic image (Figure 25.3b) revealed, for the first time, that the gold element of two lengths of tablet weaving had survived. The remains of a repeating pattern were still discernible, despite the fact that the organic thread that originally held the gold threads together had rotted away. The radiograph gave a very good level of detail; sufficient,

Controlled lifting and X-radiography of gold threads from ancient archaeological textiles 305

(a)

(c)

for example, to show that each thread was a continuous strip less than 1 mm wide. It also showed clearly where the strip had been doubled over at the ends to change direction (Figure 25.3c). If the threads had been excavated individually and bagged on site,

(b)

Figure 25.3 Gold thread from the Prittlewell chamber grave, (a) in situ as lifted with the gold coin lying alongside; this was later removed (© Andy Chopping at MoLAS), (b) radiograph of the same area of gold thread in the same orientation, (c) radiographic detail of the thread. (© Museum of London.)

any evidence of the pattern, and evidence of artefacts of which the gold thread had formed a part, would have been lost. It is thought that these two lengths of tablet weaving probably formed the decorative gold highlights to the neck opening at the front of the

306

Case studies

buried man’s clothing. The fact that they are lying crossed over each other may also be evidence of the type of neckline that the garment once had.2 This radiograph can be used to provide reasonably accurate measurements of the original pieces of tablet weaving for the purposes of formal assessment. It has also made it possible for public audiences to view an image of the thread, even though the thread assemblage itself cannot at present be separated from its sand matrix without destruction of the patterning. Showing the radiograph during lectures about the site has helped the public to appreciate the benefits of the meticulous excavation work of the archaeologists and the investigative conservation process.

Conclusions Although the results of the Spitalfields radiography were less dramatic than those of the Prittlewell excavation, in both cases radiography proved extremely informative.3 It provided crucial information quickly which immediately benefited the archaeological projects. Radiographic images of the gold threads could be viewed at a very early stage by interested textile specialists in other parts of the country, and beyond, as well as by members of the project team while the material was still fresh. Radiography made this possible at a point in the projects when there was a demand for immediate information about all aspects of these burials from national and international media. It was also invaluable for preliminary assessment purposes and conservation decision making before any more

painstaking investigation of the obscured threads could begin.

Acknowledgements The author would like to thank John-Peter Wild and Penelope Walton Rogers for their observations, the Museum of London Archaeology Service project teams and the Conservation Department at the Museum of London. MoLAS Conservation also thanks English Heritage and Southend-on-Sea Borough Council for their financial support of the Prittlewell archaeological project at Prittlewell and the Spitalfields Development Group for their financial support of the recent MoLAS excavations at Spitalfields.

Notes 1. 2. 3.

Personal communication, John-Peter Wild. Personal communication, Penelope Walton Rogers. These textiles will be discussed further in forthcoming publications in the MoLAS monograph series.

Reference Barham, E. and Lang, R. (2001). Hitting the ground running – excavation and conservation of a Roman burial in the media spotlight. In Human Remains, Conservation, Retrieval and Analysis. BAR Series 934 (E. Williams, ed.), pp. 45–54, Archaeopress.

26 X-radiography of ethnographic objects at the Horniman Museum Louise Bacon

Introduction

may mean that radiography cannot be used due to the ritual or sacred nature of the artefact.

Ethnographic objects are the material evidence of living peoples (Fowler and Fowler, 1996: 129). Today, new collections are supported by recorded fieldwork or documentary evidence. However, in the past, objects were acquired often with little knowledge about either the materials from which they were made or their method of manufacture. Radiography provides an essential tool to shed light on the technique of construction and to allow for informed interpretation of an object. It is one of the main non-destructive methods of examining and understanding ethnographic objects without using intrusive interventive conservation methods, avoiding loss of original information. This chapter will explore the methods used at the Horniman Museum, London and, through the use of case studies, highlight the benefits of radiography for ethnographic objects which include textile elements. It will also highlight some ethical considerations which

Radiography equipment and methods used at the Horniman Museum Three types of radiography equipment have been used by the Horniman Museum: a mobile unit, a cabinet unit and a freestanding unit; see Table 26.1. Standard Kodak AX and MX industrial film are used. In the past, xeroradiographs were produced using 125 transfer paper over a charged selenium plate (see Chapter 3, p. 50). The wide exposure latitude and edge-enhancing effect of this recording medium produced particularly helpful results when examining organic materials, such as wood. However, this technology, once commonly used by hospitals, is no longer available (O’Connor et al., 2002).

Table 26.1 Types of radiography equipment used at the Horniman Museum Type of unit Free-standing Pantak water cooled industrial unit (English Heritage) Cabinet, Andrex Minishot Dual 150 Industrial inspection unit (Museum of London) Sovereign portable unit

kV

mA

Exposure time

Filters

FFD

Max. 250

Max. 15

None, usually done on a lead covered table

1–1.5 m

Max. 150

Fixed at 2.5 or 5

Multi-range but never usually more than 10 minutes to avoid it overheating 15 seconds to 3 minutes (can be increased by repeating)

Beryllium disc and aluminium filter

520 mm

Max. 90

10–25

Up to 5 seconds (can be increased by repeating)

None

Max. 2 m

307

308

Case studies

Often, the main problem with X-raying ethnographic objects is their size and their complex threedimensional structure. This means that the resulting radiographs may be difficult to interpret where the X-rays have passed through, and recorded, many different layers. Consequently, it is very important to note which way up an object was when the radiograph was taken. It may take several images at different exposures and changes of the position of the object to understand what is going on.

Textile elements in ethnographic artefacts The composite and ephemeral nature of many ethnographic objects is a well-recognised problem. They can contain a great variety of materials such as metal components, vegetable and animal parts, organic fibres, including cotton or wool textiles, and a host of other inorganic materials such as glass, ceramics and plastics. The way that textiles are used as part of the construction of ethnographic artefacts means that radiography is a particularly useful tool. Woven textiles are commonly used in a functional and aesthetic manner, for example as coverings to hide unsightly construction methods or as part of a desired end result such as with the marionette discussed below. Consequently in ethnographic conservation, textile elements – although of interest in their own right – can also function to obscure information. Radiography may be used to pass through such obscuring outer textile layers to obtain information about what lies below. This has the benefit of avoiding the potential risk of damage through lifting the textile covering. Appropriate control of humidity, temperature and light is often the best – and sometimes the only – way to slow down processes of deterioration in such mixed-media artefacts. In order to establish what form of environmental control or even what conservation treatment would be most appropriate for such artefacts, an in-depth knowledge of the materials employed in the manufacture of the objects is essential. Hidden elements, more often than not obscured by woven materials in the form of a textile, are therefore of great concern and interest to the conservator. Taking an object apart to investigate what may lie beneath is an option, but is not to be undertaken lightly. By far the least destructive approach to understanding such artefacts is radiography (Lang and Middleton, 1997).

The conservation benefits of radiography for ethnographic artefacts with textile components: two case studies Radiography of a nkisi This nkisi1 (Horniman Mus. No. 13.33) is a prime example of a multi-media object (Figure 26.1a). The figure, from the Congo in Central Africa, was accessioned into the Horniman collection in 1913 but was probably made sometime during the nineteenth century. The most common feature of a nkisi is the pack attached either to the abdomen or to the back. In the Horniman example this contains beads, a small bone or tooth and granular material that is possibly grave soil or gunpowder. The Horniman nkisi has many attachments which are made up of at least fifteen different types of materials, including animal hooves, feathers, porcupine and mongoose hair, strips of cloth, copper alloy bells and grass seed. Small mounds of a resinous material have also been applied to the body. A gun cartridge, only identified through radiography, was buried in one of these.2 It was X-rayed using a Pantak Unit (60 kV, 3 mA, 12 seconds) to produce a xeroradiograph (XR218, Figure 26.1b). The wide exposure latitude of the xeroradiograph has allowed information on a wide range of materials to be captured in one exposure. This included the grain of the wood in the head as well as detail of the bone, hoof, skin, textiles, feather rachides, basketry and metal components (Hobbs, 1999). The arrangement of the textiles over the lower body can also be clearly seen. Although radiography was not primarily used in this instance to penetrate obscuring layers of textile, part of the woven raffia skirt did obscure the lower part of the wooden body. However, the external appearance of the figure had already raised the issue of multiple materials of different origins in close proximity to each other. It is known that it was customary to hide items within such figures (Shelton, 1995) and, indeed, radiography revealed a hidden metal component in the form of the cartridge case. The radiography also heightened awareness of the environmental needs of the mixed media figure as well as establishing its robustness for display. Radiography of a string marionette required for display Conservators are often requested to prepare an object so that it can be displayed in the manner in which it was used. This was the case for a string

X-radiography of ethnographic objects at the Horniman Museum 309

(a)

(b)

Figure 26.1 Nkisi power figure (Mus. No. 13.33), (a) photograph, (b) xeroradiograph showing the cartridge case in the chest of the figure and the arrangement of the textiles over the lower body. Composite image created by Sonia O’Connor, University of Bradford, from two xeroxradiographs. (© The Horniman Museum, London.)

marionette of a harlequin or tumbling clown from Naples, Italy, probably dating to the 1950s (Horniman Mus. No.20.8.63/12, Figure 26.2a). The puppet was X-rayed using a Pantak Unit (60 kV, 5 mA, 60 seconds) to produce a radiograph (XR223) using Kodak AX Industrial film (Figure 26.2b). In this instance, the goal of the radiography was three-fold. First, to see the condition of the component parts for any conservation treatment that might be required. Second, to identify what materials were underneath the clown’s clothing so that the appropriate environmental conditions could be provided for the object on display. Third, to acquire knowledge of the construction of the puppet under its clown’s clothing, primarily to establish whether it could be seated or not and whether it could take any weight at all. The radiograph reveals the inorganic and organic components of the puppet’s segmented structure

below its garments. There is a two-piece metal wire armature attaching the head to the shoulders. The shoulders, arms and torso are made from billets of wood, which appear to be strung together so that these elements articulate. The surface gesso and paint on the wooden face and hands is clearly visible; this does not extend inside the garment. The upper part of the legs is composed of overlapping wooden discs, which hang vertically and are clearly strung together through their centres. The lower legs are again made of several billets of wood in the same manner as the arms. The feet are weighted in the heel with lumps of metal. It is also possible to see where the strings from the control rod are attached to the wood. The string that would be used to raise the knee is visible, looped through the lower edge of the lowest disc in the proper left leg. Little information on the structure of textile components was visible at this exposure level, which was

(a)

(b)

(c)

(d)

Figure 26.2 Marionette of a harlequin or tumbling clown (Mus. No. 20.8.63/12), (a) photograph, (b) radiograph showing different materials, jointed limbs, nails and metal braid, (c) detail showing wooden billet from the proper left arm overlain with metal braids from the costume, (d) detail showing smiling face motif. The arrow indicates the tape running between successive billets which is attached by an iron nail. Composite image from two radiographs; due to movement between the two exposures these have not been completely joined. (© The Horniman Museum, London.)

X-radiography of ethnographic objects at the Horniman Museum 311

primarily concerned with the recording of the internal structure of the puppet. Nevertheless, as well as highlighting the structure of the piece, the radiograph makes it possible to see beneath the textiles without entirely losing their outline. Creases in the textiles are evident, as is the metallic braid, which runs vertically down the costume and around the wrists and ankles, and the beaded smiling face motif over the abdomen. Many nails are visible, attaching narrow strips of organic material to the wooden components of the body and limbs. These strips maintain the wooden components in their relative positions while providing appropriate articulation. Enlarging and enhancing the radiograph shows that, on limbs such as the proper left arm, this is a woven tape (Figure 26.2c). The smiling face motif on the abdomen is interesting, not for what is visible at this kilovoltage but for what is visible (Figure 26.2d). Some components, like the non-woven material, possibly a metallised paper, used for the mouth, are very conspicuous on the radiograph. However, other elements, like the large sequins used for the eyes, and the smaller ones, which cover the whole of this motif, are virtually invisible, even after considerable image processing. It is only the spacing of the beads with which they are sewn down that gives a clue to their spacing on the radiographic image. They are made from a material, which is radio-lucent at this exposure. The distribution of the beads follows the outline of the lips but the radiograph also reveals the full size, shape and condition of the paper that cannot be seen on the photograph. Stitch holes and tears are clearly visible around the mouth. However, the threads used in the construction of this motif are not visible at this kilovoltage. The radiograph ably records the condition of these internal components showing the internal stringing to be intact. It also shows damage to the metal braid, particularly at the point between the hands just above the knee joint. When the puppet was in use this could have been due to the constant bending of the costume and abrasion from the hands, and the strings, which go through the textile to the knees at this point. Alternatively, it could be damage due to storage if the puppet had been folded in the past. Although full of nails, none appeared to be actively corroding although the slight fuzziness around the shanks of the nails shows that they have corroded in the past and this corrosion has seeped into the wood and in some places into the textile (Figures 26.2c and d). However, intrusive interventive conservation work was not required. The radiograph confirmed that the puppet was robust

enough to be displayed seated in the showcase in controlled environmental conditions.

The ethics of radiography Kingery (1996: 195) noted that ‘Artifacts are purposeful creations.’ Their importance relates to their use and in the wider context their importance to the community and how they view them. Radiography can reveal items that have been deposited in objects, which were neither meant to be seen nor intruded upon. Radiography may be technically possible but conservators should always ask whether it is ethically acceptable when dealing with certain types of ethnographic or sacred artefacts. A collection of Buddhist figures from Tibet, made of painted clay, illustrates this point. For example, this figure (Museum No. 12.11.51/7.28, Figure 26.3) has arms formed from twists of strips of textile as an armature that is then covered with clay. However, the base of many of these figures is hollow. When in use, prayers written on paper would have been rolled up, inserted into the base, and the base covered over with clay. Advice was sought from the Head of the Tibetan Community in London who supported the implementation of the research programme (Sertic and Bacon, in press), but advised that the museum was not to X-ray nor intrude upon the religious nature of the few intact figures that were in the collection; see Reedy (1991) and Hall (2004) for a discussion of comparable issues relating to consecrated Tibetan bronzes. In order to respect this request, none of the figures where the bases were intact were X-rayed. Similar discussions took place prior to radiography with Anthony Shelton, then Keeper of Anthropology at the Horniman Museum, on the ethics of examining the nkisi power figure from the Congo. In this instance there were no overriding considerations.

Conclusion The nature of the artefacts in ethnographic collections, their multi-media make-up and their oftencomplex construction, sometimes hidden by textile coverings, makes radiography an essential tool in the understanding of the objects in our care. However, with objects of a ritual nature every due precaution must be taken to respect the cultural context of the object and this may mean that radiography is not always ethically appropriate.

312

Case studies

(a)

(b)

Figure 26.3 A Buddhist figure from Tibet (Museum No. 12.11.51/7.28), (a) photograph, (b) xeroradiograph showing the twists of textile used to make up the arms and legs. (© The Horniman Museum, London.)

Acknowledgements The author is most grateful to the following: Dr Barry Knight (formerly of English Heritage), Head of Research at the British Library for information on the English Heritage equipment; Helen Ganiaris, Senior Conservator Archaeology, Department of Conservation and Collection Care, for information on the cabinet machine at the Museum of London; Tony Riber of Mobile Radiographic Services for the mobile unit.

Notes 1.

Nkisi are considered as material manifestations of spiritual entities which do not represent a spiritual personality but provide a local habitat for it. As a material object, an nkisi is invoked to perform a certain duty which is attributed to it during its creation (Guzman, n.d.). Nails or sharp objects are driven into the nkisi to activate it and also to seal deals. It was believed that if someone lied and a sharp object was driven into the nkisi, that person would either die or fall sick. The nkisi, both in its appearance and

2.

use, was a powerful object (Dr Hassan Arero, Keeper of Anthropology, Horniman Museum, personal communication, 2006). For an in-depth overview on the function of the various elements of a fetish figure with particular reference to the Horniman object, see Hobbs (1999).

References Fowler, C. S. and Fowler, D. D. (1996). Formation processes of ethnographic collections. Examples from the Great Basin of Western North America. In Learning from Things: Method and Theory of Material Culture Studies (W. D. Kingery, ed.), pp. 129–144, Smithsonian Institution Press. Guzman, G. (n.d.). Fetish/Power/Art. Changing Perspectives on African Art. University of Virginia. http://lists. village.edu/uvamesl/bestpractices/grd3m/fetish.html# Community (accessed 5 January 2006). Hall, A. (2004). A case study on the ethical considerations for an intervention upon a Tibetan religious sculpture. The Conservator, 28, 66–73. Hobbs, V. (1999). The function of the ‘fetish’ figure. V&A Conservation Journal, 31, 17–19.

X-radiography of ethnographic objects at the Horniman Museum 313 Kingery, W. D. (1996). Materials science and material culture. In Learning from Things: Method and Theory of Material Culture Studies (W. D. Kingery, ed.), pp. 181–203, Smithsonian Institution Press. Lang, J. and Middleton, A. (1997). Radiography of Cultural Material. Butterworth-Heinemann. O’Connor, S., Maher, J. and Janaway, R. (2002). Towards a replacement for xeroradiography. The Conservator, 26, 100–114. Reedy, C. L. (1991). The opening of consecrated Tibetan bronzes with interior contents: scholarly, conservation,

and ethical considerations. Journal of the American Institute for Conservation, 30, 1, 13–34. Sertic, L and Bacon L. (in press). Tibetan Buddhist clay figures – a technical examination. Third Forbes Symposium on Scientific Research in the Field of Asian Art (September 29–October 1, 2005). Shelton, A. (ed.) (1995). Fetishism; Visualising Power and Desire. The South Bank Centre, London, and the Royal Pavilion, Art Gallery and Museums, Brighton.

27 The use of X-radiography in the conservation treatment and reinterpretation of an incomplete musette Sylvie François

Introduction This case study explores the use of radiography in the treatment of an unusual musical instrument. The enclosed three-dimensional nature of the musette meant that its interior was not accessible and radiography was selected as a means of both understanding more about the instrument and to inform conservation decision making. It was treated by the author at the Textile Conservation Centre (TCC) in 1999 as part of a Final Year Student Treatment Project for the post-graduate Diploma in Textile Conservation (TCC 2494.2; François, 1999).

Musette Bagpipes are first mentioned in the classical period. Aristophanes seems to have alluded to the idea of sounding a reed pipe through an inflated skin in his plays Lysistrata (411 BC) and The Acharnians (426 BC) (Sadie, 1984: 1, 99). Musettes are believed to be a seventeenth to eighteenth century development, linked to the French court (Figure 27.1). The sound is produced by pumping air into a leather bag using an external bellows and then squeezing the bag to force the air over reeds, which vibrate, and out through the drones and chanters. The drones of the musette are contained in a small cylindrical box, the shuttle, and can be tuned by slides to produce a fixed note. The chanters are also cylinders but the pitch of the note they produce may be changed by the use of finger-holes, in a similar manner to playing an oboe. 314

Outer cover (robe)

Drone cylinder Stock Stocks

Bellows

Blow-pipe with leather tube

Chanters

Figure 27.1 Musette de cour. (Line drawing by Sylvie François.)

The Horniman Museum musette This musette belongs to the Horniman Museum, London (Figure 27.2). Its provenance or history is unknown as it was ‘rediscovered’ in the course of a backlog cataloguing exercise.1 It was damaged and incomplete, having lost many of its music-making components: the chanter, drone cylinder and bellows. The bag is made up of three different layers: an outer cover, an inner cover and the leather bag. The outer cover, a green silk velvet richly adorned with silver bobbin lace, trim and fringe, was in a fragile state with many areas of loss and previous repairs. The second cover is a cotton plain weave fabric. Through damaged areas, three stocks were visible. These connecting parts were made of wood in the case of the chanter,

The use of X-radiography in the conservation treatment 315

Figure 27.2 The musette, before conservation. (© Textile Conservation Centre; reproduced by permission of the Horniman Museum, London.)

wood and possibly ivory for the drone cylinder, and wood with a tube of leather for the bellows. The musette was in poor structural condition, the leather bag in particular being distorted and rigid.

Treatment proposal and the role of radiography The original treatment brief was agreed between the museum and the author with the goal of preparing the musette for educational display in the museum’s musical instrument gallery. The musette would be presented as a study piece to present conservation issues since a more complete musette was already exhibited in the gallery. At this stage the musette was considered to be a complex single object; its components were seen as inseparable. It presented a challenging treatment problem as it combined textiles, leather, metal threads, wood, ivory and possibly other, as yet hidden, materials. The original treatment strategy therefore aimed to stabilise the overall object. A full support for the velvet cover was proposed using a combined technique of adhesive and stitching in situ. The velvet cover would first undergo surface cleaning and local

humidification. The cotton cover would be surface cleaned and receive patch supports for localised tears. All three stocks would be surfaced cleaned. At this point, suggested interventions were limited to the visible layers – the covers. The treatment proposal included a preventive approach to limit further damage to the object by creating a fitted mount for the musette. Following detailed visual examination, radiography of the object was proposed in order to provide a better understanding of the hidden elements. The objectives were, therefore, to provide information on the physical structure, dimensions and condition of the leather bag on which the life of this musical instrument depends and the location, materials and condition of any remaining musical components.

Radiographic procedures The musette was examined using a free-standing Pantak water-cooled industrial unit at the English Heritage Ancient Monuments Laboratory, London. Two different types of radiographic images were taken: xeroradiographs2 (Lang and Middleton, 1997: 21–24) and a film radiograph; see Table 27.1.

316

Case studies

An exposure of 60 kV was selected for the film radiographs in order to provide a wide exposure latitude as it was not known what materials would be encountered.

●

●

●

Interpreting the radiographs The two image forming techniques have produced complementary information enabling better understanding of the material, dimensions and construction of the musette as a whole. The film radiograph (Figure 27.3) showed details of: ●

●

two tubular structures (drone cylinder and bellows stocks), possibly in wood or ivory a helical structure in an organic material within the connecting tube attached to the bellows stock

the edges of two reeds wrapped with twists of wire and attached to short metal tubes the structure of the leather bag, including the seam and numerous folds the structure and design of the lace worked in silver metal thread.

With the high kV selected, the image contrast is very low. It is therefore difficult to see all the information captured, particularly the reeds, without digitising the image and performing digital image manipulation. Carrying out this manipulation enabled features that were only visible in the xeroradiograph at the time of treatment, such as hooks and eyes at the opening of the outer cover, to be highlighted3 (Figure 27.4). The xeroradiograph enhanced the edges of the components within the musette (Figure 27.5). The reeds could be seen. The contours of the leather bag

Table 27.1 Radiography details Type

Film size

Exposure time

kV

mA

Focus to film distance

Xerox–125 transfer paper over a charged selenium plate Film Kodak AX

180 mm 120 mm

12 seconds

60

3

1 m–1.5 m

410 mm 340 mm

60 seconds

60*

5

60–90 cm

Figure 27.3 Radiograph of the musette with no digital enhancement. (© The Horniman Museum, London.)

Figure 27.4 Digitally enhanced detail of the radiograph showing the hooks and eyes. (© The Horniman Museum, London.)

The use of X-radiography in the conservation treatment 317

were highlighted, providing valuable information on its method of construction: two pieces of leather stitched together along all sides using a leather strip. However, the improved visibility of the edges of features has been gained at the expense of fine details such as the silver lace and wire twists. Information gained from the radiography changed thinking about the musette as well as informing

Figure 27.5 Detail of xeroradiograph of the musette showing the reeds. (© The Horniman Museum, London.)

treatment approaches. There were indeed two remaining music-making components within the musette: the reeds. However, no treatment programme could be prescribed for the reeds as they are located within the inner bag. What had been considered as a single object was revealed as three separate sections. The inner leather bag was clearly a cut and seamed bag. This evidence was crucial in identifying this set of bagpipes as a member of the musette family. Complete animal stomachs were not used as air bags in these more sophisticated musical instruments.4 However, the vital bonus of this radiographic examination lay elsewhere than in enabling understanding the nature of the leather bag alone. Intriguing shapes in the silver lace pattern around the neck of the musette were examined closely and revealed to be metallic hooks and eyes, previously concealed beneath earlier repairs (Figure 27.4). This discovery led to further research into one aspect of the musette – the rich outer cover, known as the robe (Hollinger, 1982: 173). A literature search, supported by comparison with other musettes, prompted the suggestion that the robe could be changed to match the player’s garment in keeping with the association of this small instrument with the aristocratic society of the era.5 The whole conception of the object was therefore changed. It was no longer viewed as one

Figure 27.6 Musette after treatment. (© Textile Conservation Centre; reproduced by permission of the Horniman Museum, London.)

318

Case studies

entity of three layers, but was more properly understood as consisting of two integrated elements (the leather bag with the cotton inner cover) with a removable outside cover, the velvet robe.

thorough reading and comments on the text; English Heritage Ancient Monuments Laboratory, London.

Notes

Impact of radiography on the treatment and interpretation of the musette

1.

These discoveries were presented to the curatorial staff at the Horniman Museum. The treatment brief was modified accordingly and a revised conservation plan developed. The presence of wooden, ivory (?) and metal components meant that the risks of a humidification treatment to reshape the leather bag were too great. This approach was abandoned. The treatment of all parts could, however, be facilitated by the removal of the velvet robe for separate treatment. The initial proposal was for a combined adhesive and stitched support; as the cover could be removed, it was possible to provide a full stitched support. The previous repair stitches closing the neck were undone and the top cover was removed and stabilised. The successful treatment of the top cover revealed the original kidney shape of the musette (Figure 27.6). This in turn influenced the final display of the musette. The reshaped and padded top cover could be exhibited next to the cotton covered leather bag with its stocks. The conserved musette was included as a study piece within the Music Room of the Horniman Museum to show the specific construction of musettes. Radiography had played a critical role in understanding the object’s cultural context and its material nature as well as informing the public about its conservation treatment and influencing its interpretation.

2.

Acknowledgements Thanks are due to Nell Hoare, Director, and Dinah Eastop, Senior Lecturer and Project Supervisor, Textile Conservation Centre; Louise Bacon, Head, Collections Conservation and Care, Horniman Museum, London; Frances Palmer, formerly Keeper, Musical Instruments, Horniman Museum, London, now Curator, York Gate Collection, Royal Academy of Music, London; Margaret Birley, Keeper, Musical Instruments, Horniman Museum, London, for her

3.

4. 5.

The instrument was in the collection of the Horniman Museum by 1977 when it was registered as 1977.80, having previously been allocated a temporary museum number M1010. Xeroradiography was a medical imaging technique available at the time this study was done. It was seen as having advantages over film radiography because the images had a greater exposure latitude allowing very different materials to be recorded in a single exposure and also because it produced bright outlines around features of similar radio-opacity. However, the downside is lack of resolution of detail compared with film images and a tendency for information to be lost where the bright outlines are formed (O’Connor et al., 2002: 100–102). This technique is no longer available. Digital image processing and interpretation by Sonia O’Connor, AHRC Research Centre for Textile Conservation and Textile Studies Research Fellow in Conservation, at the University of Bradford. Personal communication, Jean-Christopher, Maillard Université de Toulouse-le-Mirail, France (1998). Personal communication, Dinah Eastop, Textile Conservation Centre (1998). This suggestion is supported by published sources on the musette de cour such as Richard D. Leppert’s Arcadia at Versailles (1978: 45) which quotes a sale notice dated 3 December 1778 for two musettes de cour. The advertisement describes in some detail the two sumptuous outer covers that were offered with both instruments.

References François, S. (1999). Dressed to Impress: The Treatment of an Incomplete French Musette (Horniman Museum 1977.80; TCC 2494.2). (Unpublished Diploma Report. Textile Conservation Centre.) Hollinger, R. (1982). Les Musiques à Bourdons: Vielles à Roue et Cornemuses. La Flûte de Pan. Lang, J. and Middleton, A., eds (1997). Radiography of Cultural Material. Butterworth-Heinemann. Leppert, R. D. (1978). Arcadia at Versailles. Swets & Zeitlinger. O’Connor, S., Maher, J. and Janaway, R. (2002). Towards a replacement for xeroradiography. The Conservator, 26, 100–114. Sadie, S., ed. (1984). New Grove Dictionary of Musical Instruments. Macmillan.

28 X-radiographic examination of a historic mannequin on display in Edinburgh Castle, Scotland David Starley and Fiona Cahill

Introduction Mannequins are not usually the focus of conservation but are more often seen as supports for more significant artefacts. However, this case study presents the initial investigative phase of a project that incorporates the conservation of four historic mannequins on long-term loan from the Royal Armouries, Leeds to Edinburgh Castle in Scotland. The conservation treatment was initially targeted at the armour mounted on the mannequins. Once this had been removed, it was realised that the mannequins themselves were likely to be of historic importance. With few records of their origins, limited expertise in this type of object and little published on the subject, one mannequin (XVII.127) was radiographed early in the project to enable greater understanding of its history, construction and condition (Figure 28.1).

History of the mannequins The date of manufacture and early history of these mannequins is unknown. However, surviving records describe a number of items being moved from overfull stores at the Tower of London to Edinburgh in 1860, and again in 1890, after the Great Hall in Edinburgh Castle had been refurbished. In total, forty-one armours were sent to Edinburgh during this period. It is assumed that these included mounts such as the mannequin under consideration. The armours remained on open display in Edinburgh until 2005 when they were taken to the Royal Armouries Laboratory, Leeds, for conservation. No records of any conservation exist prior to this.

Figure 28.1 Photograph of the mannequin wearing armour. (Reproduced by permission of the Trustees of the Royal Armouries, Leeds.) 319

320

Case studies

Description The mannequin is of human size, being 1720 mm tall, 500 mm wide (including the arms) and 230 mm deep at the hips. The mass was 20 kg. Most of the surface is covered with a medium-weight, plain weave linen textile (Figure 28.2a). In the opinion of the Armouries’ wardrobe mistress, the selvedges visible below the left arm indicated that the cloth was woven on a power loom.1 Although now discoloured, this textile has been painted to give the appearance of skin. For reasons that are not clear, a band of unpainted textile has been exposed at the neck and shoulders. Here, a turned-over edge had been opened out at an unknown date. This showed the textile to be unbleached in the loom state. This outer textile ‘skin’ had been constructed using simple flat seams, sewn on the wrong side of the fabric. These run down the centre of the back and front and diagonally at the groin where the legs joined the torso. The seams on either side of the torso run from armpit to ankle. However, there is a considerable contrast between the neat stitching of the seams from the waist downwards and the crude oversewing above this, suggesting that the upper part of the body had been opened up and reclosed after the initial construction. The radiograph shows that there is a considerable depth of turning on the proper right-hand seam where the waist has been narrowed (Figure 28.2b). The mannequin’s head, upper torso and arms are solid wood as are the feet, where they are visible through the leather ‘boots’. The ‘boots’ are not functional footwear but are only a single thickness of leather, shaped and sewn around the lower legs and feet. The arms are jointed at the elbow, with a single screw acting as a pivot. Most of the rest of the body has been stuffed with a resilient material. This was assumed, prior to radiography, to be entirely straw. The ends of two bars were exposed in the neck cavity. These pass through the shoulder block to the socket where the arms were attached, providing a flexible, near universal, shoulder joint. The top of a vertical iron bar can also be seen. Later examination of images of other mannequins from Edinburgh showed that the head that was radiographed had previously been associated with the body of a different mannequin. This particular head has a painted face. However, heavy trimming in the past, presumably to make it fit a particular helmet, has removed all but the edge of the hair as well as further material from beneath the chin; the other heads are intact. Perhaps significantly, the moustache had merely been painted on and is not part of

the original carving. Aside from the head, there was some debate as to whether the figure was originally masculine, particularly given the narrow waist. However, the build of the legs, the lower portions of which would have been visible when on display, suggest the mannequin had always intended to represent a male.

Background to the projects A routine inspection of a Royal Armouries’ loan to Edinburgh Castle highlighted the need for conservation treatment of the armours. It was only when the armours arrived at Royal Armouries, Leeds, that the historical interest of the mannequins was realised. The mannequins showed abundant evidence of their long period of display. In addition to being very dusty and dirty, there was physical damage, particularly tears and holes in the outer fabric. The boots had deteriorated badly from red rot as a result of atmospheric exposure. There was no exterior evidence of insect activity. Some alterations were evident through the presence of old holes in both the fabric and wood of the shoulders and the realignment of the top seam mentioned above.

Purposes of radiography The programme of radiography was primarily intended to enable the development of the conservation strategy by providing information about structural stability, deterioration, biological attack and so on, thereby providing as great an understanding of the construction and condition of the mannequins as is possible without invasive investigation. At the same time, it was hoped that the radiography might help date the mannequin by providing evidence of manufacturing techniques, materials and fastenings; for example, through the types of nails used.

Radiographic procedure and equipment The Royal Armouries, Leeds, is fortunate to have a Pantak 320 kV instrument on the premises (Figure 28.3). This is housed in a shielded room, effectively removing limitations on the size of object examined. The film size used was 380 mm 430 mm for the body and 180 mm 430 mm for the limbs. The mannequin was radiographed face up and repositioned between individual exposures. With a focus

X-radiographic examination of a historic mannequin on display in Edinburgh Castle, Scotland 321

(a)

(b)

Figure 28.2 Mannequin, (a) photograph without armour, (b) composite of the radiographs of the mannequin. (Reproduced by permission of the Trustees of the Royal Armouries, Leeds.)

322

Case studies

Figure 28.3 Radiographing the mannequin at the Royal Armouries, Leeds. (Reproduced by permission of the Trustees of the Royal Armouries, Leeds.)

to film distance of 1.1 m, parallax distortion was not a serious problem though sharp detail was only achieved for that part of the object closest to the film. Due to the relatively low densities encountered, the film was placed in vinyl cassettes without intensifiers. As an economy measure, some outdated film was used resulting in some plates that apparently continued to darken after fixing. The resulting radiographic plates were digitally scanned and a composite image created (Figure 28.2b).

Interpretation of the radiographs The radiographs provided considerable information. This would not otherwise have been obtainable without disassembly, which would have damaged the integrity of the historic mannequin. Possibly the most striking finding was the crudeness of the construction of the supporting framework with roughly shaped timber joined by heavy-duty hand-forged nails at the base of the spine (Figure 28.2b). The

Figure 28.4 Radiograph detail showing where the textile is tacked to the shoulder below the armpit. (Reproduced by permission of the Trustees of the Royal Armouries, Leeds.)

radiograph shows a heavy ferrous spike running vertically which does not appear to pass through the horizontal timber at hip level. There is no evidence in the radiographs of stuffing compaction around it to suggest that it was a later addition. The coarse wood grain, visible on the image, indicates a rapidly grown softwood rather than a more prestigious timber. There are no movable joints apart from those visible at the shoulders and elbows. There does not appear to be more than a single thickness of textile covering where this has been tacked to the wooden shoulder below the armpit (Figure 28.4). The features of the painted areas of the face were clearly visible (Figure 28.2b). This is thought to be largely due to the use of lead compounds within the pigment but the ground layer may contribute some increase in density. Dense paint or ground layer residues in the ear cavities, clearly visible in the radiographs, provided evidence that those parts of the head which were later shaved back were also originally painted. The apparent beard visible on the radiograph is an artefact caused by the attenuating effect

X-radiographic examination of a historic mannequin on display in Edinburgh Castle, Scotland 323

(a)

(b)

Figure 28.6 Radiograph detail showing, (a) sheared cut nails in the shoulder joint, (b) small tacks and the screw in the foot. (Reproduced by permission of the Trustees of the Royal Armouries, Leeds.)

Figure 28.5 Radiograph detail showing the stuffing and the weave of the textile ‘skin’. (Reproduced by permission of the Trustees of the Royal Armouries, Leeds.)

of the particularly radio-opaque paint on the back of the neck and below the ears. It had been assumed that the stuffing was some form of straw. However, the radiograph clearly showed a denser, more sinuous material mixed in with what may be straw or horsehair. The image also includes some impression of the outer textile covering (Figure 28.5). The plain weave of the fabric can be seen, probably enhanced by the presence of the paint.

for certain purposes, continued in use to the present day. Slightly better dating may be available for the wood screws that attached the soles of the ‘boots’ to the feet. These are of the pointed type, produced in the US from the 1830s but not in the UK until 1854. However, their relatively early date is suggested by a feature known as an untapered core which was phased out following another American innovation, patented in 1876 (Dickinson, 1941/2). Some consideration was given as to whether these were original features as they were accessible from outside the ‘skin’. However, there seems no reason to doubt this. The construction is shown by the radiograph detail (Figure 28.6b). The small tacks attach the lower edge of the leather of the ‘boot’ to the underside of the coarse-grained wooden foot while the screw holds on, and has always held on, the fine-grained wooden sole below these. The dating of the nails helps to corroborate the suggestion that the textile was woven using a power loom.

Evidence for dating Given the very limited early documentation of the mannequin, it was hoped that radiography might reveal datable features. The main evidence was in the form of nails. The larger nails holding the frame in place were handforged, suggesting a date prior to 1870 which Geselowitz et al. (1991) consider to be the effective end-date for the use of wrought nails. The sheared cut nails in the shoulders (Figure 28.6a) do not narrow the date greatly as such fasteners were used in the UK from the late eighteenth century. This is well before the likely date of the mannequin and such nails have,

Summary Although conservation work on this mannequin has only just begun, radiography has proved extremely useful in providing information about an object for which the Leeds conservation team, and conservators elsewhere, have had little previous experience. This will allow greater confidence in selecting appropriate conservation treatments and will allow a much sounder basis for describing the object and in suggesting its age. This is now thought to be mid- to late nineteenth century. Scanning and merging the

324

Case studies

radiographic images from the separate plates produced a spectacular composite image which will undoubtedly prove useful for presentations on both radiographic techniques and on the scientific examination of historic objects in general.

Acknowledgements Many thanks to Sonia O’Connor, University of Bradford, for providing the superb high-resolution digital images from the radiographs and to Joanne Plenderleith, Royal Armouries Wardrobe Mistress, for her observations on the textile.

Note 1.

Personal communication, Joanne Plenderleith, Royal Armouries Wardrobe Mistress.

References Dickinson, H. W. (1941/2). Origin and manufacture of wood screws. Transactions. Newcomen Society, 22, 79–89. Geselowitz, M. N., Westcott, T. R. and Wang, S. (1991). For want of a nail: archaeometallurgy and dating in historical archaeology. MASCA Research Papers in Science and Archaeology, 8(II), 45–55.

29 X-radiography of Rembrandt’s paintings on canvas Ernst van de Wetering Précis by Mary M. Brooks and Sonia O’Connor

Editors’ note The following is a précis of research carried out by Ernst van de Wetering into the use of radiographs in the study of canvas in seventeenth century oil paintings, particularly by Rembrandt (van de Wetering, 1986, 1997). Only the key results of his thorough and well-referenced research are highlighted; statistical data, detailed case studies, images and references are not included. Wetering’s work also contains a detailed account of the development of the use of canvases by artists and a survey of the canvases used by Rembrandt. The editors are most grateful to Ernst van de Wetering for permission to publish this brief summary of this pioneering research.

Rembrandt’s oil paintings on canvas The Rembrandt Research Project had access to radiographs of about 250 paintings on canvas attributed to Rembrandt by the art historian Bredius (1969); the main period covered is 1631 to 1642. Of these, 130 were complete radiographic surveys while the remainder were radiographs of only part of paintings. The scope of this collection of radiographs warranted research into canvas as a support for paintings.

Objectives of research into canvas supports Two reasons prompted this investigation of fabric supports. First, to explore the validity of the suggestion that thread counts could provide an indication of date or provenance and, second, to investigate whether the absence of cusping (distortion due to stretching) might

indicate that a painting has been altered in size. Some studies have attempted to explore local and regional variations, schools and periods in the use of fabrics for oil painting but very few studies have been undertaken of the canvases used by a single artist. Conservators usually record details of fabrics and seams. MeierSiem has published one of the few systematic analyses of radiographs undertaken of seventeenth century Dutch canvases from a single museum. This work, which included thread count analysis, guided a study of Rembrandt paintings in the Centraal Museum, Utrecht (Houtzager et al., 1967). Despite this study, information relating to the canvases is rarely mentioned in art historical research into paintings, probably because they are not easily accessible. The unpainted borders of canvases are commonly missing as they tend to tear along the edge of the stretcher and are lost in restoration and reframing. Obscured by paint on the front and with additional lining fabrics on the back, it is only through radiography that the original canvas structure becomes visible. However, radiographs have usually been taken to answer questions relating to the painted image and its genesis layer rather than the canvas.

Radiographs as a means of studying canvas In radiographs of oil paintings, it is not so much the canvas that produces an image but its imprint in the radio-opaque ground layer below the paint surface. The variation in thickness produced by the threads in the surface of the ground layer form features of varying darkness within the image of the ground, so revealing the structure of the canvas. Viewing the radiograph in a positive form, as a contact print using the X-ray film 325

326

Case studies

as a negative, resulted in the more radio-opaque elements (the ground in the interstices between the threads) producing the darker images. This sometimes made it easier to observe and interpret the information recorded as the imprint of the canvas in the ground showed up lighter. As canvas is barely radio-opaque, lining fabrics are not usually visible on radiographs except at the edges beyond the ground and painted area, unless these too have been coated with a radio-opaque substance such as paint. In some cases, this substance was so radio-opaque that the image it formed of the lining fabric actually overwhelmed the image of the original canvas, confusing the thread counts. In such instances, it may have been possible to detect the thread count of the original canvas only where it overlies the stretcher as the back of the fabric may have remained untouched by paint below the battens. Transferring a painting to a new canvas using a facing technique implies the removal of the original canvas. The imprint of the original canvas would be preserved in the ground layer and, in such cases, it is possible to take a thread count from its radiograph unless the ground had been sanded to such an extent that this imprint was removed. If this has happened, the replacement lining may become visible instead due to the use of a radio-opaque adhesive. There were limitations to the evidence that could be seen in radiographs because it was the imprint produced in the ground by the textile rather than the textile itself that was observed. For example, fibre identification was not possible nor could yarn structures be analysed. Although, in theory, the conical projection of the X-ray beam will produce some magnification of the image of the canvas, the close contact between the canvas and the film means that such distortion did not affect measurements significantly.

Research methods and results Thread count Thread counts of both warp and weft were taken in a systematic way to ensure statistical validity. Ten to over twenty counts (depending on the size of the piece of fabric) were taken on the radiographs in both warp and weft directions, after which the average values could be worked out. The spread of the counts was recorded so as to form an impression of the degree of irregularity of the thread count. The maximum and minimum values obtained provide an indication of the range of variation in the thread count. The choice of

the places where measurements were taken was dictated by the degree of visibility of the canvas structure in the radiograph, though a reasonable spread was aimed at. Yarn qualities Irregular and non-uniform thickening of yarns could be observed in the radiographs which, nevertheless, had a degree of constancy and formed characteristic patterns. These features, probably the result of different spinning and weaving methods, could be recognised and comparisons made. Warp direction Establishing warp direction was a precondition for the proper comparison of the thread counts obtained. Original selvedges had usually been lost so this was difficult. However, it was generally assumed that warps run parallel to seams. In smaller canvases with no seams, it was not safe to assume that the lengthwise dimension of the painting indicated the warp direction. In these cases, which were in the majority, criteria for detecting warp direction had to be developed through the evidence from the radiographs in other ways. First, warp yarns are occasionally more regularly spun and, second, there is greater constancy in the number of threads in the warp direction than in the weft direction. The resulting thread counts were sufficiently constant to conclude that, if the warp count varied by more than one thread per centimetre, it was more than likely that the canvases came from different bolts of cloth. Given the nature of the weaving process, evidence from weft counts was less reliable. In canvases with the same number of warp threads and approximately the same number of weft threads per centimetre, an assessment of the general thread and weave characteristics is necessary in answering the question whether the canvases stem from the same bolt. Examining the nature of the irregularities in the thickness of warp and weft threads may be conclusive. No correlation between canvas size and density of thread count was noted. Loom widths and painting formats Selvedges are extremely rare as hardly any canvas paintings have their original edges. As noted already, it can be safely assumed that seams in larger canvases are in the warp direction. Evidence from the radiographs showed that the majority of paintings are on canvases with a loom width of 1-12 ells (i.e. approximately 1 m).

X-radiography of Rembrandt’s paintings on canvas 327

This loom width was often used singly or doubly, making the most economic use of the fabric. Loom widths of 2 or even 3 ells were also measured. It came as something of a surprise that the dimensions of a painting were governed partly by the width of a weaver’s loom. This knowledge may be helpful in estimating the original size of altered canvases. Cusping Cusping was frequently seen in the radiographs which included the edges of a painting. When unprepared canvas is stretched, the fabric is distorted by the locally greater stress exerted at points where the canvas is attached to the oversized strainer in which the stretched canvas was prepared with ground. This results in deformations known as ‘primary’ cusping. The pitch of the primary cusping often varied. It extended in different degrees into the body of the canvas and was more or less marked at the edges. In some cases there were breaks in the regular succession of curves in the threads or the cusping became more and more marked from one corner of a painting to the other. ‘Secondary’ cusping, which does not extend as far into the body of a painting, occurs when a primed canvas was restretched on its final stretcher, either before or after it was painted. The loss of secondary cusping occurs when the edges of a canvas have been trimmed while the absence of any cusping may suggest that the size of a painting has been altered. Seventeenth century canvas may, however, also have been cut from long, semi-industrially prepared and primed strips of canvas. This would imply that canvases cut from such strips show only two, or occasionally three, edges with cusping. Information from the radiographs, combined with visual and documentary evidence, allowed for deeper understanding of methods of priming and stretching canvases as well as the later physical history of the paintings.

Characteristics of canvases by, or attributed to, Rembrandt Radiographs of Rembrandt’s canvases, whatever their date or quality, show that the ground only occasionally penetrated between the threads. This is because these canvases were first treated with glue size as a sealant. Where there was local penetration of the canvas by the ground, this showed up on the radiographs as isolated or clustered white spots. The thread counts taken of canvases known to be by, or attributed to, Rembrandt conformed to the

norm for Dutch paintings. There were some striking variations. Some variation was to be expected but finding this within the work of one artist was not expected as it seemed likely that an artist would use bolts of canvas kept in the studio. This research showed that canvases from Rembrandt’s workshop between 1632 and 1634 came from at least twenty-five bolts of fabric, suggesting that they were probably bought separately (or in batches) from professional primers. On the basis of thread counts and yarn qualities, several small groups of paintings could be identified as coming from the same bolt. They apparently belonged to the same batch of canvases bought at the same time. These groupings could be confirmed through provenance or historical links; for example, some were painted as companion pieces. Weaving faults, unusual thread counts and loom widths also provided corroborative evidence.

Conclusion This research, based on careful and time-consuming analysis of numerous radiographs, demonstrated the importance of studying fabric supports with both negative and positive results. It established that thread counts cannot be used for dating canvases and specific weave counts cannot be identified with particular artist’s studios, including Rembrandt’s. There are clearly implications for understanding the means by which seventeenth century artists obtained their canvases and suggests that the person commissioning the painting may have supplied the canvas, thus controlling the quality and format of the support. Radiographs showing cusping can provide insights into trade practices and the physical history of a painting. Studying canvas also illuminated some specific issues relating to Rembrandt’s paintings. Identifying some groups of paintings as being on the same canvas established that they originated in the same workshop and clarified some attribution and dating problems. Information from radiographs of canvases used for oil paintings can thus support and enhance evidence from other sources. It is hoped that such radiography will become established practice. As more radiographs are taken of artists’ canvases, the more it will become possible to assess the significance of the evidence which they reveal.

Acknowledgements The editors would like to thank Alan Phenix, then of University of Northumbria, for drawing their

328

Case studies

attention to Professor Wetering’s research. Ernst van de Wetering was assisted by a number of colleagues, temporary assistants and interns including Emil Bosshard, Greet van Duyn, Brigitte Blauwhoff, Michiel Franken, Loutje den Tex and Koos Levy-van Halm.

References Bredius, A. (1969). Rembrandt: The Complete Edition of the Paintings (first edition 1935, revised by H. Gerson). Phaidon.

Houtzager, M. E., Meier-Siem, M., Stark, H. and de Smedt, H. J. (1967). Röntgenonderzoek van de oude schilderijen in het Centraal Museum te Utrecht. Centraal Museum. Van de Wetering, E. (1986). The canvas support. In A Corpus of Rembrandt Paintings (Rembrandt Research Project Foundation; J. Bruyn, ed.) II, pp. 15–43; republished and extended in van de Wetering, E. (1997). Rembrandt. The Painter at Work, pp. 90–130, 298–304, Amsterdam University Press.

Index

Acrylic fabric, 8 Adhesive, 84, 145, 187, 193–194, 238, 262, 264, 326 Air: absorption, 18 attenuation, 25–26 filtration, 25, 45, 51 Alteration, 4, 163, 168, 205, 209 Aluminium cassette, 20, 43, 221, 304 see also Film cassette Aluminium filter, 40–41, 206, 274 Anaglyph, 52, 70 Analogue image, 58 Apollo space suit, 7 Appliqué, 41, 76, 83, 153, 243–244, 273–274, 282, 284 reverse, 35, 37 Archaeological: Archbishop Walter de Gray, 5, 51, 266 footwear, 166, 169, 294–301 Godfrey de Ludham, 49 matrix, 46 pre-Columbian, 5 Peruvian, 5, 9 Queen Arnegonde of France, 5 shoe, 166, 169 soil block, 297 textile, 26, 31, 46, 126, 128–129, 133, 134, 136, 137, 164–165, 266, 294–301, 302–306, Armature, 4, 6, 8, 309–311 Assessing equipment sensitivity, 28 Atomic number, 13 Attenuation, beam, 13, 15, 16 Baby clothes, 54 Background density variation, 76 Background radiation, 94, 100 Backscattered radiation, 19 Backstitch, 142, 229 Baleen, 80–81, 92, 119–120, 121, 158, 163–164, 166, 168, 204–205, 206–209, 214 structure of, 119 Bamboo, 192 Banner, 43 Barium solution see Contrast agent Basketry, 308 Bead, 6, 187, 199, 310–311 glass, 28–29, 158, 238 ceramic, 92

Bead work, 275 Beating pattern, 135 Beatle wing, 258, 261 Beta radiography (ß-radiography), 50 source, carbon 14, 50 Bi pack film, 18 Binocular stereoscope, 52 Bird, 238–240 Beak, 238, 240 furcula Bird, Junius, 5 Bisque, 251–252, 253, 261, 262, 264 Black thread, 154, 166, 242, 245 Blanket stitch, 142 Blood, 6, 120 Bobbin lace, 138 Bodice, 81, 86–88, 147–148, 157–158, 168, 170 Swiss, 163–164 Bone, 55, 298–299, 308 Bonnet, 7, 122 Book cover, 167 Boot, 5, 6, 147, 214–215, 294–301, 320 ankle, 214–215 button, 5 construction, 149, 168–169, 214–215 safety, 214–215 Braid, 5, 86–88, 190 stitch, 243 woven, 138 Bran, 123 Brehmsstrahlung radiation, 15 Bridgman, Charles F., 5 Brocade, 135, 190 Brown paper, 83–84 Burial environment: frozen, 295–296 permafrost, 46, 295–296 waterlogged, 48 wet, 295 Burnt textile, 164–165 Busk, 206–207 Cane, 122 Canvas, 24, 54, 81, 82, 135, 150, 264, 325–328 Cap, sprang, 49, Cape, 146–147 Carbon 14 dating, 92

Card, 40, 54, 83–84, 122, 204–205, 206, 231, 233, 258, 264 washer, 119, 269–272 Cardboard, 81, 85, 123, 164–165, 187, 192, 194–195 mount, 193 Casing, 121, 147–148, 163 Cassette see Film Cassette Ceramic bead, 92 Chain line, 194–195 Chair see Upholstered furniture Chalice veil, 24, 69, 76–77, 123, 138, 141, 153, 164, 169, 225–230 Chasuble, 46, 47, 69, 71, 154–156, 163, 187, 188–190 Chicken wishbone, 4 Chinese screen, 187, 188, 190–193 ‘Chip’, 7 Choosing X-ray facilities, 25–26 Clinical radiography, 8, 10, 16, 24, 52 Clog, 289, 291–293 Clothes, 5, 28, 45 Coir, 111 Colour change, 91 gemstone, 91 Combining images, 44–45 Commissioning radiographs, 12, 23 Complex textile, 39–40 Composition, 249, 251, 252, 254, 261–262, 264, 265 Computed radiography, 67–69 imaging plate, 68 Computed tomography (CT), 7, 30, 45–46, 49, 52–56, 201, 210 Concealed: garment, 26, 203–204, 206 thread, 145 Conical beam projection, 186–187, 258–260, 326 Conservation: mount, 238 support, 71, 73, 170–171 Continuous spectrum, 14 Contrast agent, 5 Conventional radiography, 15 Conventional tomography, 52 Cord, 41, 74–75, 76, 126, 148 Cord quilting, 74–75, 80, 84, 86, 163, 274, 282–284, 286 Correct exposure parameters, 35–39

329

330

Index

Corrosion, 122, 164, 176, 223–224, 238, 253, 262, 263, 311 Corset, 119, 122, 205, 213, 214, 215, 216 Cost, 10, 66, 72, 194 Costume, 54 Cotton: boll, 111, 113 fabric, 8, 187 seed, 111, 113, 278 tape, 86–88 twill, 86–88 wadding, 84 Couching, 235 Coverlet, 127, 131, 141, 150, 153, 163 Cowry shell, 41 CR see Computed radiography Crack, 8, 192 Cracquelure, 158, 238 Crease, 76, 122, 134, 224, 226, 228, 280, 282 Crêpe, 5 Crier, 24, 263–264 CT see Computed Tomography CT number see Hounsfield unit CT scanner, 52, 54 Cupar hat, 123, 141, 164–165 Curved surface, 18, 43 Cushion, 51, 115, 117 Cusping, 325, 327 Cut velvet, 195 Darn, 170–171, 218, 223–224 Dart, 86–88, 120, 145 Dating, 92, 109, 122, 135, 163, 238, 249, 274, 323, 327 Degradation of image quality, 20 Dendrochronology, 238 Densitometer, 21 Density, material, 13 Dental radiography, 7 Diamond, 55, 156 Dichroic fog, 20 Digital image, 58–73 archiving, 62, 64, 301 benefits, 58–59 bit depth, 59–61 capture, 59, 61, 64–66, 214 compression, 62–63 cost, 72 data migration, 64 dynamic range, 59, 61–62 enhancement, 70–71, 73 file format, 62–64 high dynamic range (HDR), 61 histogram manipulation, 69–70 manipulation, 58, 69–71, 238, 242 metadata, 62–63 mosaicing, 69 negative transform, 69–70 output, 59, 61 pre-processing, 65, 67, 69 processing (DIP), 69–71, 75–76, 183, 260 resolution, 58–61

stereo pair, 52 storage, 62, 64 viewing, 61 Direct radiography (DR), 67 Dish developing, 20, 35 Display: strategy, 190, 196, 275, 308, 311, 318 stress, 30, 144, 196 Dmax, 19–20, 61, 66 Dmin, 21 DNA, 92 Dock plant, 229–230 Document scanner, 65–66 Documentation by radiography, 231 Doll, 3, 4, 6, 24, 32, 45, 81, 114, 115, 126, 149–165, 249–265, 266–267 bisque, 251–252, 253, 261, 262, 264 cloth, 171–172 cloth body, 252 clothes, 4, 252, 254, 258, 261, 264 composition, 249, 251, 252, 254, 261–262, 264, 265 see also Doll, wax on composition damage, 249, 252, 258, 261, 262, 264–265 eye, 249, 251, 252, 259, 262, 263, 264 eye mechanism, 28, 249, 251, 252–253, 254, 256, 260, 262, 264, 265 image interpretation, 258–260 joint, 4, 28, 45, 249, 251, 252, 256, 262 kid body, 252 limb, 249, 251, 252, 253, 258, 262, 265 Lord Clapham, 266–267 nut, 251, 258, 260, 261 papier-mâché, 6 plastic, 28, 40, 60, 64, 65, 68 radiography procedure, 258, 260 repair, 171–172, 258, 264–265 sound mechanism, 251, 252, 253, 256, 262, 263–264 stuffing, 251, 252, 258, 262, 263 tea, 123–124 test object, 37–38, 40, 64, 65, 67, 68, 123, 158–159 walking mechanism, 83, 251, 253–254, 256, 258, 264 wax, 249, 251, 252, 253, 261, 262, 265 see also Doll, wax on composition wax on composition, 88–90, 249, 251, 260, 264 wig, 262 wood, 4, 249, 251, 252, 258, 261, 262 Dosemeter, 102–103 Dosimeter, 38, 186 Dots per inch, 61 Down feather skirt, 113–114, 150, 152 DR see Direct radiography Drawing, 243 see also Underdrawing Dress construction, 148

Dressed statue, 187, 188, 196–199 joint, 197–199 undergarment, 197–199 Dressed wax sculpture, 187, 188, 199 Dressmaker, 163 Dyes, 107, 133, 134, 150, 154, 221, 242, 278 Dynamic range, 18, 21, 68 Earth apple see Globe Edge: cut, 129, 135 finished, 140 hemmed, 129 selvedge, 133, 135, 141, 190, 227–229, 275, 320, 326 Edge enhancement, 50, 71, 72 Effective focal spot, 14, 18, 25, 26, 50–51 Eiderdown, 278, 280 Elastic band, 28 Electromagnetic radiation, 12–13, 91 energy, 12 frequency, 12 wavelength, 12 Electromagnetic spectrum, 12–13 Electron beam, 14 Electron potential (keV), 15 Electron transmission radiography, 50 Embroidered book cover, 167 Embroidery, 73, 76, 190, 199–200, 217, 225–230, 235, 237–248, 273 feather stitch, 280–281 machine, 163 mixed media, 83–85 opus senese, 188–190 Energy dispersive X-ray fluorescence (EDXRF), 93 Enhancing images, 66, 69, 70, 72, 75–76 see also Digital image processing Epoxy resin, 8 Equipment selection, 20 Ethical implications, 307, 311 Ethnographic mask, 41, 42 Excelsior see Wood-wool Exhibition evaluation, 216 Exposure: contrast, 19 duration, 20, 25, 37, 40 latitude, 16, 18, 19, 40, 68 parameters, 18, 20, 34, 83 meter, 38–39 Fabric, 86, 135 acrylic fabric, 8 cotton, 107, 109, 314–315, 318 felted, 261 fold, 261 fulled, 133, 137 horsehair, 111 image density variation, 134–136 interlining, 146–148 linen, 134, 204–205 lining, 146–148

Index 331 mohair plush, 268–269 net, 28, 181, 182 painted, 145–146, 150–151 pile, 127, 164 printed, 150, 152–153, 204–205 silk, 131, 132, 147, 204–205, 289–293 synthetic, 28 velvet, 135, 314–318 wool, 107, 109, 205 Fancy dress, 168, 169, 219, 223 Fast Fourier transform, 71 Fatty stain, 195 Feather, 92, 308 see also Rachis boning, 122 down, 113–114, 278–280 Felt, 83, 133, 136, 261, 298 FFD see Focus to film distance Fibre, 107–119 acrylic, 139 cotton, 107, 110, 111 flax, 181 horsehair, 107 jute, 187, 199 linen, 107, 110 man-made, 107, 126, 128 polyester, 116 reinforced composite, 55 silk, 107, 110 synthetic, 181 tow, 200 wool, 110 Filling, 81, 82, 107, 109–119, 178, 179, 182 see also Padding, Stuffing and Wadding coir, 112 cotton, 111, 112, 114 feather, 113–114 hair, 176, 177, 181 horsehair, 112, 113–114 kapok, 112, 114–115 mohair, 118–119 polyester, 115, 116 soya bean fibre, 115–116 straw, 251 wood-wool, 118, 119–120 wool, 117–118 ‘sub’, 268–269 Filling cover, 176–177, 178, 180, 182 Film see Radiographic film Film cassette, 20, 68–69, 238 cellulose acetate window, 35 distortion, 35 gelatine window, 35 re-useable flexible plastic, 35–36, 322 Film enclosure 20, 34–35 see also Film cassette black plastic bag, 258 paper envelope, 275, 289 pre-packed film, 20, 35, 289 laminated, 35 Filmless digital radiography, 58–59 benefits of, 64, 66–67

cost, 66 direct, 67 indirect, 67 X-ray scanner, 67, 68 Filter, 18, 238, 237–238 aluminium foil, 28, 40 holder, 41 Filtration, 17, 18, 41, 68 Fish scale, 158, 160 Flag, 43 Flexible plastic cassette, 322 Flores, Elia, 8–10 Flower, 28–29, 123 Fluorescent screen, 10, 19, 42 Fluoroscopic X-ray machine, 5 Foam rubber, 115, 117 Foamed resin, 8 Focal spot see Focus Focus, 14, 16, 17, 18 Focus to film distance (FFD), 16–18, 25, 26–28 Fogging, 18–19, 20, 43 Fold, 81, 122 Folding damage, 233 Footwear, 166, 169, 294–301 see also Shoes and Boots Fourier transform, 71, 73, 240 Freeze/thaw cycling, 295, 299 French knot, 242 Fringe, 133 Fungi, 231–234 Fur, 83 Furnishing fabric, 145–146, 150–151 Furniture construction, 164, 175–184 see also Upholstered furniture Gamma radiation, 12–14 Gaps, 84, 86–88 Gemstone, 91, 156, 187, 195 Geometric unsharpness, 18, 23, 25, 50–51 Germanisches Nationalmusuem Nürnberg, 6 Gesso, 178, 180, 261–262, 309–310 Gilded copper, 187 Gilding, 175, 176 Glass, 54, 156–157, 249 Globe, 6, 52–53 Glove, 6 Glue Size, 264, 327 Godfrey de Ludham, 49 Gold: powder, 193, 194 thread, 5, 76, 154, 187, 190, 196, 199, 302–306 Grenz ray, 25, 32 Greyscale image, 61 Growler, 269, 270, 271, 288 ‘Guestimation’, 36 Gun cartridge, 308 Hair, 308 Hand stitching, 86–87, 141 Quality, 229

Hanging wall pockets, 7, 157, 166, 169, 231–236 ‘Hard’ X-rays, 13 see also High energy X-rays Hat, 123, 141, 164–165 Hay see Straw Helical wire trim, 233 Helium filled chamber, 39 Hem, 129 Heterogeneous object, 38, 40–41 Hidden thread, 153 High contrast: film, 19 image, 17 High definition radiograph, 23 High energy X-rays, 13, 16, 91 images, 23 Hoof, 308 Hoop, 7 Horsehair, 83–84, 262, 323 Hounsfield unit, 54 Human remains, 49, 54, 298–299 Ice, 46 Image: artefact, see Image interpretation brightness, 16 continuity, 45, 258 contrast, 16, 18, 19, 25, 26, 37, 40–41, 43 darkness, 16 definition, 35 degradation, 21, 46 density, 16, 20, 23, 76, 78, 80, 83 distortion, 17–18, 28, 44 hardening, 43 interference pattern, 81, 82, 86–88, 127–129, 226 layout, 32 permanency, 21 receptor, 13, 15, 18 reference, 74 sensitivity, 19, 31 sharpness, 19, 41, 43 unsharpness, 18 viewing, 21–22, 61, 76, 107 Image interpretation, 28, 54, 81, 74–90 Artefact, 20, 41, 52, 54, 88–90, 322 Gap, 84–88 overlaid lines, 78, 79, 282–283 reference, 74 systematic survey, 76 Image quality, 17–18, 23, 30, 32, 51 Degradation, 30 indicator (IQI), 19, 32–33, 76 optimising, 25 Indian hanging, 156 Indigo pencilling, 150 Industrial radiographic film scanner, 65–66 Industrial X-ray: film, 19, 20 radiography, 19 X-ray unit, 19, 25, 27, 178, 183

332

Index

Inherent filtration, 19 Ink, 243, 278 Insect: cocoon, 209 damage, 164–165, 181, 208–209, 231 frass, 208–209 Interactive display, 212–213 Interference pattern, 81, 82, 86–88, 127–129, 226 Interleaving film in object, 189, 195–196 Inverse square law, 16 Ionising radiation, 13 risks, 28 IQI see Image quality indicator Iron wire, 4 Armature, 200 Ivory, 315, 316, 318 ‘Jap gold’ thread, 154 Japanese paper, 190 Jewellery, 6 Joint, 320 Kapok, 268–269, 271 Seed, 115 Kid, leather, 147, 252 Knitting, 139–140, 217–224 damage, 223–224 knot, 223 machine, 217 stitches, 139–140 yarn join, 139, 221, 223 Knot, 126, 154, 223, 229, 242 Knotless netting, 136, 138 Koenig doll collection, 6 Lace, 147–148, 158 bobbin, 138, 314, 316–317 metal, 6, 225, 231, 314, 316–317 Laid and couched work, 217, 219, 225 Lamb’s skin, 269 Lanolin, 117–118 Large films processing, 187 Lavender, 28–29, 123 Layered textile, 39–40, 78 Layout of pattern pieces, 135 Lead: numbers, 32, 43 pigment, 178, 179, 180 putty, 4 sheet, 19, 35 Lead screen intensifier, 25, 41–43, 46, 47, 86, 121–122, 221, 304 Leather, 40, 43, 48, 54, 83–84, 114, 125, 215, 289–293, 314–318 archaeological, 294–301 kid, 147, 252 lamb’s skin, 269 synthetic, 125 Lifejacket, 114 Light box, 21, 213 Line spectrum, 14

Linen, 54, 83, 126, 128–129 fabric, 181, 187 interlining, 204–205 lining, 204–205 printed, 204–205 yarn, 5 Lining fabric, 169–170, 325–326 Lock stitch, 142–144, 281 Loom width, 227–229, 326–327 Lord Clapham doll, 266–267 Lost thread see Missing thread Low energy X-ray, 12, 16, 17, 19 beam, 26 high definition radiography, 24–25 Machine knitting, 217 Machine sewn: chain stitch, 81, 142–143 embroidery, 163 irregularity, 143 lock stitch, 142–144, 281 seam, 81 Magnification, 20, 21, 107 Maker, 273 domestic, 141, 154 group, 143, 163, 168, 274, 278, 282 professional, 141, 154, 236 single, 163, 168, 274 Mammography, 8, 24–25 film, 10, 19 X-ray unit, 19 Man-made thread, 128 Mannequin, historic, 319–324 Manual processing, 20 Mapping damage, 166 Mapping wear, 55 Marionette, 308–311 construction, 309–311 Martin Behaim’s Erdapfel (earth apple) 6, 52–53 Mask, 41, 42 Material: Density, 13, 15, 16 Identification, 78 Maximum optical density (Dmax), 19–20, 61, 66 Minimum optical density (Dmin), 21 Mechanism, 32, 45, 249, 251, 254, 258, 264 clockwork, 254, 256 eye, 25, 45, 251, 252, 256, 260, 264, 265 movement, 251, 262, 264, 265, 269, 271–272, 288 sound, 24, 251, 252, 256, 262, 263–264, 268–269, 270, 271, 288, 314–318 walking, 251, 256, 258, 264 Medical radiography, see also Clinical radiography equipment, 19, 25 film, 10, 19, 21 film scanner, 65–66

Metal, 28, 54, 187 armature, 6 braid, 83, 310–311 component, 308 concretion, 48 corrosion, 122, 164, 176, 223–224, 238, 253, 262, 263, 311 eyelet, 264 fastening, 316–317 fixing, 269–272 lace, 6, 225, 231, 314, 316–317 pin, 5 Metal thread, 5, 46, 83, 154–156, 187, 195–197, 199, 302–306 coiled wire, 154 complex, 154–155 damage, 235 embroidery, 5, 7, 51, 248 fabric, 43, 44 fibre core, 80, 154, 156, 229, 302 gold, 76, 154 ‘Jap gold’, 154 joins in wrapping, 234 modern, 156 plastic support, 156 purl, 154–155, 223 strip, 304–306 tack, 176, 177, 178, 179, 180 wrapped, 80, 126, 302–304 wrapping, 154–155, 234, 235 Metal thread work, 24, 43, 46, 47, 76, 78, 83–84, 166–167, 169, 217–224, 225, 231–236 Metalised paper, 310–311 Methodological unity, 185–186 Micro-CT (computed tomography), 49–50, 54–55 Micro-focus radiography, 50–51, 154, 159 Mineral: filler, 28, 123 inclusion, 278 Mineral preserved organic remains (MPO), 49 Textile, 49 Mirrored panel, 41 Missing thread, 145, 164 Mitre, 6, 32, 187, 194–196 Mixed media object, 7, 41, 54 Mobile X-ray unit, 28, 101, 175, 182–183 Moiré effect, 86–88, 127–129, 226 Moiré silk, 225 Moon boots, Neil Armstrong, 6 Mordant, 107, 133, 134, 154, 221, 242, 261, 278 Mosaicing, 44–45, 69, 71 Motorcycle helmet, 214–215 Mottern, Robert W., 5 Mummy, 8 Mummy wrapping, 126, 128–129, 134 Musette, 314–318 Nail, 177, 178, 179, 190–192, 193, 197–199, 310–311, 320, 322–323

Index 333 Nano-CT (computed tomography), 55 National Aeronautics Space Administration (NASA), 6 Natural pearl, 159, 161 Necktie, 166, 168 Needle holes, 143, 145, 300 Negative image, 74–75 Neoprene, 214–215 Net, 28, 181, 182 Neutron activation, 5 Nkisi, 308–309 Non-flat object, 39 Non-invasive investigation, 185–186 Non-woven fibre, 199 Non-woven structure, 55, 136, 138–140, 311 Nut, 258, 260, 261 Nutmeg, 246–247 Nylon net, 181, 182 ‘Old Woman in the Shoe’, 3–4, 71–72, 126, 168–169 Optical density, 21 Optically stimulated luminescence dating (OSL), 92 Optimising film viewing conditions, 21 Optimising image quality, 16, 18 Over exposure, 16, 19, 20, 37 Overcast stitching, 227, 229 Overlaid lines, 78–79, 229 Overstitching, 226 Padding, 192, 235, 243–244, 245 see also Filling, Stuffing and Wadding Paint, 4, 150, 175, 176, 310 Painted textile, 145–146, 150–151, 176, 178, 225, 320 Paintings, 24–25, 28, 43, 74, 93 lining, 325–326 Rembrandt, 325–328 Pallium, 49 Paper, 4, 25, 74, 93, 122, 136, 187, 192–193 brown paper, 83–84 metallised, 310–311 structure of, 50 template, 78, 107, 109, 122, 131, 141, 163, 274, 275, 278 Par speed screen 10, see also Fluorescent screen Parchment, 195, 247 Passing thread, 217, 219 Patch, 6, 127, 131, 143 Patchwork, 109, 122–123, 141, 163, 273 –287 chevron, 273 clamshell, 273, 275–275, 278 coverlet, 78, 107 crazy, 273, 278, 281 doll’s bedspread, 273, 275, 277, 283 hexagon, 273 log cabin, 273, 278, 280 Pearl, 159, 161, 195, 238, 242–243 Peg joint, 249 Pesticide, 170

Petticoat, 81 Photo degradation, 91, 93 Photographic density, 21 see also Optical density Photographic effect of X-rays, 13 Photon, 12, 14 Pigment, 76, 150, 193, 226, 278, 322 Pile fabric, 127 Pin, 28, 145, 176, 238, 247, 292–293 Pin holes, 123, 145 Pintuck, 141 see also Tuck Piping, 48, 126, 148 Pixel, 58, 59 Plain weave, 127, 176, 199, 217, 219, 225–226, 242, 275, 278, 323 Plait, 7–8, 126, 130 Plant remains, 123, 229–230 Plaster of Paris, 261 Plastic, 28, 123 mesh support, 156 sequin, 28–29 Pocket, 76 Polyester wadding, 116 Polyurethane foam, 115 Porcupine quill, 80, 122 Portable X-ray unit see Mobile X-ray unit Positioning objects, 30–31 Positive image, 74–75, 325–326 Pre-exposed test film, 21 Pre-packed film, 20, 35, 289 Pressure suit see Space suit Pressurised flight suit, Wiley Post, 7 Primary beam X-rays, 18 Printed: fabric, 150, 152–153 paper, 192–194 Processing monitoring control strip (PMC), 21 Public engagement, 10, 210, 212–216, 265, 274, 284–285, 287, 306 Puppet, 8, 308–311 Purl, metal thread, 154–155, 223 Purse, 48, 123, 126, 130 Quality control: processing, 20–21 processing monitoring control strip (PMC), 21 over date film, 322 Quilt, 43, 111, 143, 163, 273–287 corded, 163 damage, 283 Durham, 274, 280–281, 282 Elizabeth Watson, 153, 163, 274, 282–284 filling, 278 log cabin, 163 Mary Burnett, 111, 119, 150, 168, 273, 278–279, 281, 283 recycled, 281, 283 ‘YGF’, 274, 282–285 Quilting, 126 frame, 282 pattern, 282–285

Rabbit fur, 8 Rachis, 113, 278, 308 see also Feather Radiation: Damage (EDXRF), 93–94 legislation, 96–98 monitoring, 101–103 protection, 100–101 regulation, 97–98 Radiation risk, see also Safety assessment, 97–100 dose limits, 98 dose units, 99 harm, 97, 99–100 justification, 96–97 limitation, 96–97 optimisation, 96–97 reduction, 96–97 Radiograph, 15 Radiographic exposure, 15 Comparing, 19 Radiographic film, 13, 15, 18, 34, 61, 64–66, 68, 69, 73 archival standards, 22 archive, 21–22 automatic processing, 20, 35 cassette see Film cassette contact with, 23, 31 contrast, 19, 20 copying, 21 damage, 20 see also Image artefact, 20 degradation, 21 digitising, 58–59, 64–66, 201, 300 emulsion, 19 enclosure see Film enclosure grain, 19, 20 handling, 20 industrial, 34 large, 64, 45, 187, 192, 197 mammography, 34 medical, 10, 19 positioning, 30–32 pre-packed, 20, 35, 289 processing, 19, 20–21, 35, 187 selection, 19–20 shelf life, 20, 322 size, 26, 43 slow, 26 speed, 18, 19 storage, 20, 21, 22 viewing, unenclosed, 34 Radiographic image, 13, 15 Digital, 215 negative form, 74–75 parallax, 322 penumbra, 18 positive form, 74–75, 325–326 Radiography: clinical see Radiography medical computed (CR), 67–69 conventional, 15 dental, 7 direct (DR), 67

334

Index

Radiography (contd) industrial, 21, 24, 26, 52 medical, 8, 19, 26, 42, 74, 213–214 security, 92 transmission see Radiography, conventional Radiography of textile objects: (see individual entries for specific materials of object types e.g. Shoes, linen, threads etc.) engaging with the public, 10, 210, 212–216, 265, 274, 284–285, 287, 306 commissioning, 12, 23 conservation treatment timing, 247 cost benefits, 194 degradation test, 94 documentation of textiles, 30, 231 ethical implications, 307, 311 exposure parameters, 183–184, 188, 206, 219–221, 228, 251, 268, 273–274, 289, 297, 304, 307, 316 handling, 30 large textiles, 186–187, 275 monitoring change, 247 rationale, 30 record keeping, 32, 34 risk to object, 10, 35, 91–95, 185, 188 ritual objects, 307–313 survey planning, 43 Radio-opacity, 15, 16, 18, 19 Raffia, 308 Raised work embroidery, 30, 83–84, 85, 154 Real-time radioscopy, 51 micro-focus, 51 Reconstructing depth, 78–80 Recording missing threads, 145 Recycled material, 119 ‘sub’, 268–269 Reference: collection, 81 images, 74 Reinforcement: patch, 243 textile, 125 Reliquary bags, Saint Wolfgang of St Emmeram, 6 Repair, 4, 8, 9–10, 140, 169, 171–172, 176, 188, 190, 193–194, 205, 219, 235, 314, 317–318 patch, 205, 225–226, 229 stitches, 283, 286 thread, 4, 126 Resist technique, 150 Reticule, 158, 160 Reverse appliqué, 35, 37 Ribbed plain weave, 226 Ribbed silk fabric, 147 Ribbon, 41, 54–56, 88–89, 133, 275, 283, 286 Risk: to object, 10, 35, 91–95, 185, 188 to personnel, 28 see also Safety

Rolled textile, 26, 54–56, 123–124 Röntgen, Wilhelm, 5 Rose bud, 123 Rust 3, 122 Safety, ionising radiation, 96–103 Satin stitch, 225 Sawdust, 115, 117, 171–172 Scattered radiation, 13, 18, 19, 20, 23, 43, 46, 100–101 Screw, 323 Sculpture, wax, 187, 188, 199–201 Sealing wax, 265 Seam, 6, 30, 39, 71–72, 76, 78, 81, 133, 134, 135, 138, 140–141, 142, 147–148, 169, 217, 219, 268–269, 270, 278–280, 281, 300, 316, 317, 320, 325, 326 allowance, 78 see also Turnings butted, overcast, 141, 144 distorted, 144 flat, 140–141, 144 flat and fell, 141–142 French, 140–141 hand-sewn, 140, 163, 225–226 machine butted, 144 machine sewn, 141, 163 stressed, 144 turnings, 227–229 Securing threads, 141–142, 148, 153, 282, 284 Seed, 49, 55, 115, 123, 258, 261 Selection of X-ray facilities, 16, 26 Selvedge, 133, 135, 141, 190, 227–229, 275, 320, 326 Sequin, 6, 78, 166, 192, 310–311 gelatine, 158 metal, 157–159, 231, 233, 234 plastic, 28–29, 65, 158, 159 Settee see Upholstered furniture Sewing holes see Needle holes Sewing thread, 300 Sheep wool, 107 Shell, 258, 261 Shoe, 3, 4, 5, 43, 44, 71–72, 86, 126 archaeological, 166, 169 construction, 43–44, 84, 86, 168–169, 288–293, 294–301 damage, 166, 168–169 material, 43–44, 289–293 radiographic procedure, 297 stitching, 290–293 wear, 166, 168–169 Shoulder plate, doll, 6 Shroud of Turin, 5–6, 135 Shut-eye mechanism, 28, 45, 250, 251, 252–253, 254, 256, 260, 262, 264, 265 Signature, 282–384 Silk, 3, 4, 54–56, 83–84, 126, 130, 133, 147–148, 164–165, 166, 168, 217–224, 225–230 degradation test, 94 fabric, 134–136, 187

fibre, 50–51 fragment, 187 painted, 145–146, 150–151, 176, 178, 225, 320 satin, 240 thread, 127 unweighted, 39 velvet, 187 Silver, 41, 46, 83, 153, 154, 187, 188, 190, 217, 225, 231, 314, 316–317 gilt, 231 thread, 187 Skin, 308 Skirt, down feather, 113–114, 150, 152 ‘Soft’ X-rays 19, see also Low energy X-rays Soil block, 304 Soiling, 107, 122, 150, 207, 209, 262 Soya bean fibre, 115–116 Space suit, 6–7 boots, Neil Armstrong’s, 6 glove, Michael Collin’s, 6 Spangle see Sequin Spatial resolution, 54 Spectral lines, 15 Spectrum, continuous, 14 Spin direction, 50 Spinning irregularities, 126 Splicing, 126–128 Sprang, 136, 138 cap, 49 Squeaker, 251, 252, 262, 263–264, 268–269, 270 Standing screen see Chinese screen Starch paste, 190 Statue see Dressed statue Stain: wax, 195, 196 fatty acid, 195 Stay, 119–122, 147–148, 158, 168 baleen, 80–81, 120, 122, 163, 164, 166, 204–209 casing, 86, 119–120, 121, 147, 163, 208, 209 channel, 205, 208 steel, 86–88, 120–122, 214 stitching, 120 Stays, 119, 205, 208–209 see also Corset Stepwedge, 19, 32 Stereolithography, 55 Stereoradiography, 5, 45, 51–52 Stitched support, 170–171 Stitching, 6, 30, 69, 76, 78, 141–144, 163, 164, 182, 271, 273–274, 278–280, 290, 320 see also Machine sewn back, 142 blanket, 142 bracing, 292–293 braid, 243 channel, 205, 208 direction, 143 holes, 298, 310–311

Index 335 laid and couched, 242, 243 long and short, 153, 242, 247 machine, 86–87, 278–280 overcast, 86 patterns, 163 quality, 143 quilting, 78 repair, 169–170 running, 78, 142, 282 stay, 120 tacking, 78, 126, 143, 163 threads, 233 zigzag, 147, 278, 281, 293 Stocking, 31, 169, 217–224 Stomacher, 92, 120, 168, 203–211, 213, 214 construction, 208–209 dating, 208 Storage, 193, 296 damage, 163, 166, 168, 209, 267, 311 strategies, 4, 8, 275 Straw, 115, 116, 251, 252, 262, 320, 323 Stress, 30, 144, 166, 195, 196, 209, 218, 327 Stuffing, 81, 251, 252, 258, 262, 263, 268–269, 320, 323 see also Filling, Wadding and Padding Substructure, 3, 5, 6, 7, 8, 10, 81, 119–123, 164, 175, 176, 180, 198, 321–322 Suitcase, 54–55 Support, 119–123 board, 240 see also Conservation mount fabric, 240 Survey, planning, 43 Swinegate purse, 48, 123, 126, 130 Synthetic: pearl, 159–161 fibre, 181 textile, 39 Table cover, 275–276, 278 Tablet weaving, 49, 304–306 Tack holes, 176 Tack, metal, 176, 177, 178, 179, 180 Tacking stitches, 126 Tape measure, 246–247 Tassel, 48 Tea, 123 Tear, 122, 171, 190, 279, 310–311 Teddy bear, 51, 115, 117, 119, 169, 266–272, 288 assembly, 269–272 construction, 269–272 damage, 267–268 eyes, 270 fixings, 269–272 joint, 266–272 lamb’s fleece 269 radiographic procedures, 268, 271 repair, 269–270, 271–272 stuffing, 268–269

Test: doll, 37–38, 40, 64, 65, 67, 68, 123, 158–159 exposure, 35–38 object, 28–30, 54–55 Textile Conservation Centre, 7–8 Textile object: archaeological, 5, 9, 26, 31, 46, 51, 126, 128–129, 133, 134, 136, 137, 164–165, 266, 294–301, 302–306 burnt, 164–165 complex, 39–40 composite object, 24 construction quality, 140–141 construction technique, 140–148 crease, 166, 168 curved, 18, 43 damage, 9–10, 135, 164–170, 205–209 degradation, 163, 164, 209 difficulty of radiographing, 23 ethnographic, 307–313 folded, 166, 168 frozen, 46 handling, 188–189 heterogeneous, 13, 19, 40–41 irregular, 18 large, 43 layered, 39–40, 78 leather composite footwear, 296–301 mixed media object, 7, 41, 54 moiré finish, 225 moisture distribution, 55 multilayered, 54 painted, 145–146, 150–151, 225, 320 re-use, 141 rolled, 123–124 single layer, 30 structure, 5 thick, 40 thin flat multiple layer, 30 thin homogenous, 39 three dimensional, 18, 45–46, 187 tubular, 219, 221 waterlogged, 48 wear, 166, 168 Thread, 5, 23, 46, 68, 71–72, 84, 86, 126–127, 145 concealed, 126 count, 127, 135, 190, 325–326 damaged, 129 diameter, 50 irregularities, 126, 127 join, 126 knot, 126 linen, 141 missing, 127 repair, 169, 272 silk, 126, 187 spliced, 126–128 weighted, 126, 169 Three dimensional: imaging, 30, 49, 50, 51–56 information, 45

layering, 55–56 object, 28, 45–46, 55, 187 Thermoluminescence dating (TL), 92 Toe puff, 147 Tog 8, see also Puppet Toy, 3, 4, 45, 51, 71–72, 115, 126, 149–165, 249–265 dating, 249 Transmission radiography, 15 Transparent overlay, 78, 207 Tri pack film, 18 Trouble shooting: equipment fault, 19 film processing fault, 19 image artefact, 20, 41, 52, 54, 88–90, 322 Tube: current (mA), 15, 16 peak voltage (kVp), 15 voltage (kV), 15 Tuck 6, see also Pintuck Turning, 86–87, 127, 140–141, 282, 284 cut edge, 227 frayed edge, 227 selvedge, 133, 135, 141, 190, 227–229, 275, 320, 326 Twine, 178 Under exposure, 16, 20, 38 Underdrawing, 84, 153, 225, 247 see also Drawing Undergarment, 197–199 Underskirt, 28, 40, 41 Understructure see Substructure Underwear see Undergarment Unencapsulated film, 20, 26, 188, 275 Unweighted silk, 39 Upholstery, 114 fabric, 145–146, 150–151, 176, 178 see also Upholstered furniture, top cover Upholstered furniture, 3, 24, 28, 164, 175–184 alteration, 176 Audley End settee, 180–183 base cloth, 182 bridal ties, 182 condition, 178 converted nail, 182 damage, 180 exposure parameters, 183–184 filling, 176, 178, 179, 181, 182 frame, 176, 180, 182 frame joints, 176 gesso, 178, 180 gilding, 175, 176 insect damage, 164, 181 lead pigment, 178, 179, 180 metal fastener, 176 nail, 177, 178, 179 nylon net, 181, 182 paint, 175, 176 radiographic procedure, 175 repair, 176

336

Index

Upholstered furniture (contd) Seehof suite, 176, 178–180 tack, 176, 177, 178, 179, 180 tack holes, 176 timber profile, 176 top cover, 177, 178, 180, 181, 182 understructure, 175, 180 webbing, 176–177, 182 Ultra violet light (UV), 12, 91, 92–93 Vacuum chamber, 39 Vacuum packed film, 43, 224 Viewing radiographic images, 21–22, 61, 76, 107 Velvet, 46, 83, 133, 141, 187, 193–195, 314–318 Vinyl cassette see Flexible plastic cassette Visual interference, 127, 129, 226 see also Moiré effect Voxel, 54 Wadding, 111, 112, 115, 116, 117, 273, 274, 278, 281–282, 283 Waist cloth, 59, 70 Walking doll, 264 Warp: direction, 135, 326 faced plain weave, 176 irregularity, 134, 326 thread, 129 Water, 23, 46, 48–49, 54 ice, 46 stain, 6 Waterlogged, 48 Watermark, 50, 74, 93 Wax, 187, 188, 195, 196, 199–200, 249, 251, 252, 253, 261, 262, 265 Wear, 9–10 Weave, 86–88 see also Warp and Weft beating pattern, 135 brocatelle, 107–108 complex, 133 damask, 190, 197–198 ermesini, 190 fault, 6, 127, 133, 134, 136, 194, 327 flaw see Weave fault irregularity, 129, 136 lampas, 188–190, 197–198 net, 28, 181, 182 pile, 127, 164 plain, 129, 133, 176, 217, 219, 314, 323 ribbed plain weave, 226 satin, 129, 132, 135, 190, 193, 199 satin self patterned, 133, 135 structure, 28, 50, 76, 81 tablet, 49, 304–306 taffeta, 188–190

technique, 10 twill, 129, 131, 132, 135, 136, 205, 298 variation see Weave irregularity velvet, 135, 314–318 warp faced plain weave, 176 Webbing, 176–177, 182 Weft: count, 326 irregularity, 134–135, 326 thread, 24, 78, 129 Weighted silk, 4, 86, 93, 126, 166, 168, 169 thread, 168, 169 Weighting, 107 Whalebone see Baleen Whalebone corset, 5, 214, 216 Wire, 8, 122, 238, 247, 262, 309–311 Wood, 54, 55, 187, 192, 261, 289–290, 308, 309–311, 321–322 backboard, 238 ‘chip’, 7 damage, 192 dendrochronology, 238 frame, 179, 180 grain, 238 knot, 193 rot, 164 Wood-wool, 115, 119, 262, 268–269, 271 see also Excelsior Woodworm, 164 Wool: fabric, 133, 137, 140–141 fibre, 50 fulled, 140–141 thread, 127 wadding, 83, 274 yarn, 133 Woven textiles, 126–136 Wrapped metal thread, 223 Xeroradiography, 50, 307, 308–309, 312, 315–316 X-ray: absorption, 13, 18 analysis see X-ray fluorescence spectroscopy dose, 15, 16, 18, 91, 93 see also X-ray safety electrical generation, 13, 14 flux, 13, 16 photon energy, 15 production, 13–15 properties, 12–13 safety, 96–103 spectrum, 14, 15 X-ray beam: attenuation, 20, 25–26, 86 centre, 78, 80

conical projection, 31 divergence, 18, 26 energy, 15, 16, 83 filtration, 25, 40–41, 45, 51, 206, 274 focus, 17 see also Focal spot and Effective focal spot hardening, 25, 41 intensity, 13, 15, 18, 25, 26 penumbra, 18 positioning, 17, 78, 80 primary, 18 spectrum, 18, 238 width, 18 X-ray facilities: access, 26 cleanliness, 26 dual energy, 17 industrial, 17 medical, 16 museum, 26 ‘real time’, 17 room, 18, 101 security, 17 selection, 17, 19 X-ray film see Radiographic film X-ray fluorescence spectroscopy (XRF), 91, 93, 154, 185 X-ray room, 18, 101 X-ray tube, 14, 15, 25, 26 anode, 17 current (mA), 15, 16 peak voltage (kVp), 15 target, 14, 25 voltage (kV), 15 window, 14, 17, 25 X-ray unit: cabinet type, 18, 26, 101, 219, 258 ceiling mounted, 28 continuous running, 26 cooling, 187 electrical generation, 13, 14 free-standing, 26 filtered, 237–238 fluoroscopic, 5 industrial, 178, 183 medical, 182 mobile head type, 28 specification, 25–26 stand mounted, 28 XRF see X-ray fluorescence spectroscopy Yarn, 5, 80, 86, 126–127, 133 irregularity, 126, 326 join, 126 knot, 126 spliced, 126, 128 quality, 326