Investigations into the Dyeing Industry in Pompeii: Experimental Archaeology and Computer Simulation Techniques 1789697425, 9781789697421, 9781789697438

Citation preview

Archaeopress Roman Archaeology 86

Investigations into the Dyeing Industry in Pompeii Experimental Archaeology and Computer Simulation Techniques

Heather Hopkins Pepper

Investigations into the Dyeing Industry in Pompeii Experimental Archaeology and Computer Simulation Techniques

Heather Hopkins Pepper

Archaeopress Roman Archaeology 86

Archaeopress Publishing Ltd Summertown Pavilion 18-24 Middle Way Summertown Oxford OX2 7LG www.archaeopress.com

ISBN 978-1-78969-742-1 ISBN 978-1-78969-743-8 (e-Pdf) © Heather Hopkins Pepper and Archaeopress 2022

Cover: Replica dyeing apparatus, constructed 2002, based on Vat 5 in Property VII ii 11 then Vat 5 in Property VII xiv 17 in Pompeii. Photograph: author.

All rights reserved. No part of this book may be reproduced, or transmitted, in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior written permission of the copyright owners. This book is available direct from Archaeopress or from our website www.archaeopress.com

To my family and to my friends, who encouraged and supported me to be whatever I wanted to be.

To everyone who got me through in tempestuous times... Thank you. I wouldn’t be here without you.

‘Your years at Bradford will be the most exciting of your life.’ Freshers Guide, 2002.

‘The temptation, when a problem arose, was always to reach for the volume of Vitruvius, conveniently to hand, rather than go to the site, often inconveniently distant, to look at the actual remains, much less, to cross the quad to that, psychologically at least, even more distant ultima Thule, the Faculty of Engineering.’ A. Trevor Hodge, 1992.

‘If a Stone Age Man could have come to life again and seen the models I made, I expect he would have flung himself down and laughed until he cried. That is why I call it experiment in reconstruction. No one was going to catch me saying that I intend to build a Stone Age house as it really had been, but only as it might have been. Now here is the problem that I was faced. Try it for yourself before I go on to tell the story of what happened and how I fared. Figure 2 is an archaeologist’s plan of the site on which I based my experiment, and includes information given in his report on his excavations. See if you can reckon from this how the house would have looked, which is what I had to do. Figure 3 shows how I made a model of house B on the site. If you think that you can do better than mine, all right, or if you feel mine are better than yours, I won’t quarrel with that either. It will just show how difficult it is to arrive at the truth, or how many possible ways there are of interpreting facts.’ Hans Ole Hansen, I built a Stone Age house, 1959.

Archaeological experiments … ‘generate powerful memories and infectious enthusiasm that persists for life’. Marion Blockley, 1999.

Contents Acknowledgements to this publication�������������������������������������������������������������������������������������������������������������� vi Preface����������������������������������������������������������������������������������������������������������������������������������������������������������������� vii Chapter Zero Preface to the published thesis�������������������������������������������������������������������������������������������������������1 Introduction����������������������������������������������������������������������������������������������������������������������������������������������������������������������������1 How and why the approach of this study differed������������������������������������������������������������������������������������������������������������1 How location and time affected this study�������������������������������������������������������������������������������������������������������������������������2 Location������������������������������������������������������������������������������������������������������������������������������������������������������������������������������2 Study in the context of time and technological developments��������������������������������������������������������������������������������3 Development of experimental archaeology ����������������������������������������������������������������������������������������������������������������������3 Overview of the study������������������������������������������������������������������������������������������������������������������������������������������������������������4 Background to this study������������������������������������������������������������������������������������������������������������������������������������������������4 New approaches within this study��������������������������������������������������������������������������������������������������������������������������������5 Surveying the original apparatus in 2002��������������������������������������������������������������������������������������������������������������5 The survey in 2002: water supply����������������������������������������������������������������������������������������������������������������������������5 The survey in 2002: finding a ‘Real Roman’�����������������������������������������������������������������������������������������������������������6 The survey in 2002: defining ‘Recording’, ‘Conservation’, ‘Reconstruction’����������������������������������������������������7 Redefining ‘experiment’��������������������������������������������������������������������������������������������������������������������������������������������������8 Finite Element Analysis���������������������������������������������������������������������������������������������������������������������������������������������������8 ‘Roman’ vs ‘Modern’ manufacturing�����������������������������������������������������������������������������������������������������������������������������8 Intangible questions���������������������������������������������������������������������������������������������������������������������������������������������������������9 Expansion since the doctorate was awarded�������������������������������������������������������������������������������������������������������������10 Presentations and publications since this thesis was submitted����������������������������������������������������������������������������������11 Publications �������������������������������������������������������������������������������������������������������������������������������������������������������������������11 In Press�����������������������������������������������������������������������������������������������������������������������������������������������������������������������������12 Conferences presentations��������������������������������������������������������������������������������������������������������������������������������������������12 Seminars given����������������������������������������������������������������������������������������������������������������������������������������������������������������13 Acknowledgements to the published thesis�������������������������������������������������������������������������������������������������������16 Timeline of previous work relevant to this study����������������������������������������������������������������������������������������������17 Chapter One Introduction to the dyeing industry of Pompeii���������������������������������������������������������������������������19 1.1 The significance of the scale of manufacture of textiles in Pompeii����������������������������������������������������������������������19 1.1.1. Aims and Objectives���������������������������������������������������������������������������������������������������������������������������������������������20 1.1.2. Nature of this investigation �������������������������������������������������������������������������������������������������������������������������������21 1.2 Literature Review�����������������������������������������������������������������������������������������������������������������������������������������������������������21 1.2.1 Research background��������������������������������������������������������������������������������������������������������������������������������������������21 1.2.2 Roman Dye vat design�������������������������������������������������������������������������������������������������������������������������������������������22 1.2.3 The debate so far���������������������������������������������������������������������������������������������������������������������������������������������������22 I viii 19������������������������������������������������������������������������������������������������������������������������������������������������������������������������������22 Previous work: A Strengths and Weakness Analysis�������������������������������������������������������������������������������������������23 Moeller�����������������������������������������������������������������������������������������������������������������������������������������������������������������������23 Jongman���������������������������������������������������������������������������������������������������������������������������������������������������������������������24 Mann ��������������������������������������������������������������������������������������������������������������������������������������������������������������������������24 Laurence���������������������������������������������������������������������������������������������������������������������������������������������������������������������24 Summary��������������������������������������������������������������������������������������������������������������������������������������������������������������������24 1.2.4 Quantifying archaeological writing��������������������������������������������������������������������������������������������������������������������25 Related work�������������������������������������������������������������������������������������������������������������������������������������������������������������25 Watling, 2004�������������������������������������������������������������������������������������������������������������������������������������������������������������25 1.3 The significance of this study���������������������������������������������������������������������������������������������������������������������������������������26 1.3.1 To conclude������������������������������������������������������������������������������������������������������������������������������������������������������������26 1.4 Thesis Outline:�����������������������������������������������������������������������������������������������������������������������������������������������������������������27

i

Chapter Two Literature Review���������������������������������������������������������������������������������������������������������������������������28 2.1 Introduction��������������������������������������������������������������������������������������������������������������������������������������������������������������������28 2.2 Textile processing�����������������������������������������������������������������������������������������������������������������������������������������������������������28 2.3 Dye types��������������������������������������������������������������������������������������������������������������������������������������������������������������������������28 2.4 When to dye: Stage at which dyeing takes place�������������������������������������������������������������������������������������������������������29 2.5 Roman textiles����������������������������������������������������������������������������������������������������������������������������������������������������������������29 2.6 Requirements for dyeing�����������������������������������������������������������������������������������������������������������������������������������������������30 2.7 Consumables used in manufacture������������������������������������������������������������������������������������������������������������������������������30 2.7.1 Fleece�����������������������������������������������������������������������������������������������������������������������������������������������������������������������30 2.7.2 Mordants�����������������������������������������������������������������������������������������������������������������������������������������������������������������31 2.7.3 Roman dyes������������������������������������������������������������������������������������������������������������������������������������������������������������32 2.7.4 To summarise ��������������������������������������������������������������������������������������������������������������������������������������������������������33 2.8 Undyed textile�����������������������������������������������������������������������������������������������������������������������������������������������������������������33 2.9 Process consumables������������������������������������������������������������������������������������������������������������������������������������������������������33 2.9.1. Water supply���������������������������������������������������������������������������������������������������������������������������������������������������������33 2.9.2 Water content��������������������������������������������������������������������������������������������������������������������������������������������������������33 2.9.3 Water quality����������������������������������������������������������������������������������������������������������������������������������������������������������34 2.9.4 Reliability of ancient and modern sources on water���������������������������������������������������������������������������������������35 2.9.5 Fuel��������������������������������������������������������������������������������������������������������������������������������������������������������������������������35 2.10 Recipe directions����������������������������������������������������������������������������������������������������������������������������������������������������������36 2.10.1The Recipe used in this study�����������������������������������������������������������������������������������������������������������������������������36 Recipe for mordanting���������������������������������������������������������������������������������������������������������������������������������������������36 Recipe for mordanting:��������������������������������������������������������������������������������������������������������������������������������������������36 Recipe for dyeing with madder:�����������������������������������������������������������������������������������������������������������������������������37 Times for recipe��������������������������������������������������������������������������������������������������������������������������������������������������������37 2.11 Removing the water�����������������������������������������������������������������������������������������������������������������������������������������������������37 2.12 Quantity of textile��������������������������������������������������������������������������������������������������������������������������������������������������������37 2.12.1 Roman garments ������������������������������������������������������������������������������������������������������������������������������������������������38 2.12.2 Pictorial representation of Roman dress���������������������������������������������������������������������������������������������������������38 2.12.3 Problems with depiction������������������������������������������������������������������������������������������������������������������������������������39 2.13 Population����������������������������������������������������������������������������������������������������������������������������������������������������������������������39 2.14 Intangible evidence������������������������������������������������������������������������������������������������������������������������������������������������������39 2.15 Summary������������������������������������������������������������������������������������������������������������������������������������������������������������������������40 Chapter Three Experimental Replica������������������������������������������������������������������������������������������������������������������41 3.1 Introduction��������������������������������������������������������������������������������������������������������������������������������������������������������������������41 3.2 A differing approach������������������������������������������������������������������������������������������������������������������������������������������������������41 3.3 Experimental archaeology��������������������������������������������������������������������������������������������������������������������������������������������41 3.4 Constructing the apparatus �����������������������������������������������������������������������������������������������������������������������������������������43 3.5 Experiment One: Preliminary experimentation��������������������������������������������������������������������������������������������������������44 3.6 Experiment Two��������������������������������������������������������������������������������������������������������������������������������������������������������������45 3.6.1 Hypothesis��������������������������������������������������������������������������������������������������������������������������������������������������������������45 3.6.2 Apparatus����������������������������������������������������������������������������������������������������������������������������������������������������������������46 3.6.3 Diagram�������������������������������������������������������������������������������������������������������������������������������������������������������������������46 3.6.4 Method��������������������������������������������������������������������������������������������������������������������������������������������������������������������47 3.6.5 Results���������������������������������������������������������������������������������������������������������������������������������������������������������������������48 3.6.6 Discussion���������������������������������������������������������������������������������������������������������������������������������������������������������������48 3.6.7 Conclusions from practical experiment �����������������������������������������������������������������������������������������������������������48 3.6.8 Summary�����������������������������������������������������������������������������������������������������������������������������������������������������������������49 3.7 Discussion �����������������������������������������������������������������������������������������������������������������������������������������������������������������������50 3.7.1 Reconstruction�������������������������������������������������������������������������������������������������������������������������������������������������������50 3.7.2 The use of wood or charcoal��������������������������������������������������������������������������������������������������������������������������������50 3.7.3 Relative energy of wood and charcoal ��������������������������������������������������������������������������������������������������������������50 3.8 Further work�������������������������������������������������������������������������������������������������������������������������������������������������������������������52 Chapter Four Review of Remains in situ���������������������������������������������������������������������������������������������������������������������������� 53 4.1 Fieldwork in Pompeii�����������������������������������������������������������������������������������������������������������������������������������������������������53 4.2 Gazetteer of the dye vats discovered in Pompeii to date�����������������������������������������������������������������������������������������53 4.2.1 Defining the dye vats��������������������������������������������������������������������������������������������������������������������������������������������54 ii

Attributes of a dye vat following the operation of a replica:����������������������������������������������������������������������������54 4.2.2 Extent of excavation in Pompeii.������������������������������������������������������������������������������������������������������������������������55 4.3 The Survey and Gazetteer���������������������������������������������������������������������������������������������������������������������������������������������55 4.3.1 Summary of dye vats���������������������������������������������������������������������������������������������������������������������������������������������55 Property I viii 19�������������������������������������������������������������������������������������������������������������������������������������������������������55 Properties V I 4 and V I 5�����������������������������������������������������������������������������������������������������������������������������������������55 Property VII xiv 17���������������������������������������������������������������������������������������������������������������������������������������������������55 Property VII ii 11������������������������������������������������������������������������������������������������������������������������������������������������������55 Property IX iii 2 ��������������������������������������������������������������������������������������������������������������������������������������������������������56 4.3.2 The discounted ‘dye vat’���������������������������������������������������������������������������������������������������������������������������������������56 4.3.3 Flued vats����������������������������������������������������������������������������������������������������������������������������������������������������������������57 4.3.4 Vats and steps��������������������������������������������������������������������������������������������������������������������������������������������������������57 4.4 Water supply to dye works��������������������������������������������������������������������������������������������������������������������������������������������57 4.5 Discrepancy between the digital map and the aerial photograph of Pompeii�����������������������������������������������������59 4.6 Bowing �����������������������������������������������������������������������������������������������������������������������������������������������������������������������������62 4.7 Recent excavation�����������������������������������������������������������������������������������������������������������������������������������������������������������63 4.8 Further work�������������������������������������������������������������������������������������������������������������������������������������������������������������������63 Full Gazetteer of Dyeing Apparatus in Pompeii��������������������������������������������������������������������������������������������������65 Chapter Five Application of Ergonomics to Apparatus and Skeletal data���������������������������������������������������������98 5.1 Ergonomics of a dyeing apparatus�������������������������������������������������������������������������������������������������������������������������������98 5.2 The height of the average Roman��������������������������������������������������������������������������������������������������������������������������������98 5.3 Ergonomics����������������������������������������������������������������������������������������������������������������������������������������������������������������������99 5.4 Skeletal evidence from Herculaneum�������������������������������������������������������������������������������������������������������������������������99 5.5 Build��������������������������������������������������������������������������������������������������������������������������������������������������������������������������������101 5.6 Criticism of data �����������������������������������������������������������������������������������������������������������������������������������������������������������101 5.7 Difference of approach ������������������������������������������������������������������������������������������������������������������������������������������������101 5.8 Studying Pompeii and Herculaneum�������������������������������������������������������������������������������������������������������������������������102 5.9 Understanding modern ergonomics�������������������������������������������������������������������������������������������������������������������������102 5.10 Lifting���������������������������������������������������������������������������������������������������������������������������������������������������������������������������103 5.11 Working day����������������������������������������������������������������������������������������������������������������������������������������������������������������104 5.12 Nutrition����������������������������������������������������������������������������������������������������������������������������������������������������������������������104 5.13 Necessary steps�����������������������������������������������������������������������������������������������������������������������������������������������������������106 Property I viii 19�����������������������������������������������������������������������������������������������������������������������������������������������������107 Property Vi4������������������������������������������������������������������������������������������������������������������������������������������������������������107 Property Vi5������������������������������������������������������������������������������������������������������������������������������������������������������������108 Property VII ii 11����������������������������������������������������������������������������������������������������������������������������������������������������108 Property VII xiv 17�������������������������������������������������������������������������������������������������������������������������������������������������109 Property IX iii 2�������������������������������������������������������������������������������������������������������������������������������������������������������109 5.14 Summary����������������������������������������������������������������������������������������������������������������������������������������������������������������������110 5.15 Further work���������������������������������������������������������������������������������������������������������������������������������������������������������������110 Chapter Six Flued Experimental replica�����������������������������������������������������������������������������������������������������������111 6.1 Introduction �����������������������������������������������������������������������������������������������������������������������������������������������������������������111 6.2 Flued vats�����������������������������������������������������������������������������������������������������������������������������������������������������������������������111 6.3 Vat size discrepancies in the replicas������������������������������������������������������������������������������������������������������������������������111 6.4 Experiment Three���������������������������������������������������������������������������������������������������������������������������������������������������������112 6.4.1 Hypothesis������������������������������������������������������������������������������������������������������������������������������������������������������������112 6.4.2 Diagram�����������������������������������������������������������������������������������������������������������������������������������������������������������������112 6.4.3 Apparatus��������������������������������������������������������������������������������������������������������������������������������������������������������������113 6.4.4 Method������������������������������������������������������������������������������������������������������������������������������������������������������������������113 6.4.5 Results�������������������������������������������������������������������������������������������������������������������������������������������������������������������113 6.4.6 Discussion�������������������������������������������������������������������������������������������������������������������������������������������������������������114 6.4.7 Conclusion to Experiment Three����������������������������������������������������������������������������������������������������������������������114 6.5 Stress on the vats ���������������������������������������������������������������������������������������������������������������������������������������������������������114 6.6 Summary������������������������������������������������������������������������������������������������������������������������������������������������������������������������116

iii

Chapter Seven The Finite Element model���������������������������������������������������������������������������������������������������������117 7.1 Introduction������������������������������������������������������������������������������������������������������������������������������������������������������������������117 7.2 Replicating the material behaviour of the apparatus���������������������������������������������������������������������������������������������117 7.3 Creep�������������������������������������������������������������������������������������������������������������������������������������������������������������������������������117 7.4 Changes over time��������������������������������������������������������������������������������������������������������������������������������������������������������118 7.5 A virtual replica������������������������������������������������������������������������������������������������������������������������������������������������������������119 7.5.1 Modelling the dye vat ����������������������������������������������������������������������������������������������������������������������������������������120 7.5.2 Finite element analysis���������������������������������������������������������������������������������������������������������������������������������������121 7.5.3 Modelling creep���������������������������������������������������������������������������������������������������������������������������������������������������122 7.6 Requirements for the model���������������������������������������������������������������������������������������������������������������������������������������124 7.6.1 Experiment Four �������������������������������������������������������������������������������������������������������������������������������������������������124 7.6.1.1 Hypothesis���������������������������������������������������������������������������������������������������������������������������������������������������124 7.6.1.2 Apparatus����������������������������������������������������������������������������������������������������������������������������������������������������124 7.6.1.3 Diagram��������������������������������������������������������������������������������������������������������������������������������������������������������124 7.6.1.4 Method���������������������������������������������������������������������������������������������������������������������������������������������������������124 7.6.1.5 Results����������������������������������������������������������������������������������������������������������������������������������������������������������125 7.6.1.6 Discussion of Experiment Four����������������������������������������������������������������������������������������������������������������129 7.6.1.7 Conclusion to Experiment Four���������������������������������������������������������������������������������������������������������������129 7.6.2 Average temperature profile�����������������������������������������������������������������������������������������������������������������������������129 7.7 Constructing the finite element model���������������������������������������������������������������������������������������������������������������������130 7.7.1 Modelling the apparatus in the computer������������������������������������������������������������������������������������������������������130 7.7.2 Geometric model�������������������������������������������������������������������������������������������������������������������������������������������������132 7.7.3 Material properties ��������������������������������������������������������������������������������������������������������������������������������������������133 7.7.4 Loading������������������������������������������������������������������������������������������������������������������������������������������������������������������133 7.7.4.1 Self-weight��������������������������������������������������������������������������������������������������������������������������������������������������133 7.7.4.2 Hydrostatic loading������������������������������������������������������������������������������������������������������������������������������������134 7.7.4.3 Static load����������������������������������������������������������������������������������������������������������������������������������������������������134 7.7.5 Introduction of temperature�����������������������������������������������������������������������������������������������������������������������������135 7.8 Results����������������������������������������������������������������������������������������������������������������������������������������������������������������������������135 7.9 The constructed dyeing apparatus ���������������������������������������������������������������������������������������������������������������������������140 7.10 Summary����������������������������������������������������������������������������������������������������������������������������������������������������������������������142 Chapter Eight Discussion����������������������������������������������������������������������������������������������������������������������������������� 143 8.1 Introduction������������������������������������������������������������������������������������������������������������������������������������������������������������������143 8.2 This work in context����������������������������������������������������������������������������������������������������������������������������������������������������143 8.2.1. Experimental archaeology��������������������������������������������������������������������������������������������������������������������������������143 8.2.2 Definition of experiment������������������������������������������������������������������������������������������������������������������������������������144 8.3 Review of standing remains����������������������������������������������������������������������������������������������������������������������������������������145 8.4 Ergonomics��������������������������������������������������������������������������������������������������������������������������������������������������������������������145 8.5 Summary of each section of work������������������������������������������������������������������������������������������������������������������������������145 8.5.1 Original work, the foundation of study ����������������������������������������������������������������������������������������������������������146 8.5.2 Preliminary work�������������������������������������������������������������������������������������������������������������������������������������������������146 8.6 Assumptions applied to the industry������������������������������������������������������������������������������������������������������������������������147 8.7 The new findings from this study �����������������������������������������������������������������������������������������������������������������������������147 8.7.1 Significance of lead���������������������������������������������������������������������������������������������������������������������������������������������147 8.7.2 Comparison of modern manufacturing systems to Roman dyeing �����������������������������������������������������������149 8.7.2.1 Inputs and outputs�������������������������������������������������������������������������������������������������������������������������������������154 8.7.2.2 Buffers����������������������������������������������������������������������������������������������������������������������������������������������������������154 8.7.2.3 Storage���������������������������������������������������������������������������������������������������������������������������������������������������������155 8.7.2.4 Information Process Cycle �����������������������������������������������������������������������������������������������������������������������155 8.8 Conclusions��������������������������������������������������������������������������������������������������������������������������������������������������������������������156 8.9 Further work�����������������������������������������������������������������������������������������������������������������������������������������������������������������157 Glossary�������������������������������������������������������������������������������������������������������������������������������������������������������������� 159 References���������������������������������������������������������������������������������������������������������������������������������������������������������� 162

iv

Appendices��������������������������������������������������������������������������������������������������������������������������������������������������������� 167 Appendix One����������������������������������������������������������������������������������������������������������������������������������������������������������������������168 Coding Pompeii: The layout of the city and address description�������������������������������������������������������������������������168 Location of dye works in Pompeii������������������������������������������������������������������������������������������������������������������������168 Location of properties�������������������������������������������������������������������������������������������������������������������������������������������168 Property I viii 19�����������������������������������������������������������������������������������������������������������������������������������������������168 Properties Vi4 and Vi5������������������������������������������������������������������������������������������������������������������������������������168 Property VII ii 11����������������������������������������������������������������������������������������������������������������������������������������������169 Property IX iii 2������������������������������������������������������������������������������������������������������������������������������������������������169 Property VII xiv 17�������������������������������������������������������������������������������������������������������������������������������������������169 Appendix Two���������������������������������������������������������������������������������������������������������������������������������������������������������������������170 Understanding the economic influence of the dyeing industry in Pompeii through the application of experimental archaeology and thermodynamics���������������������������������������������������������������������������������������������������170 Abstract��������������������������������������������������������������������������������������������������������������������������������������������������������������������170 Introduction������������������������������������������������������������������������������������������������������������������������������������������������������������170 Dye vat design���������������������������������������������������������������������������������������������������������������������������������������������������������171 Methodology�����������������������������������������������������������������������������������������������������������������������������������������������������������172 Implementation of Engineering Theory ������������������������������������������������������������������������������������������������������������172 Results����������������������������������������������������������������������������������������������������������������������������������������������������������������������175 Conclusion���������������������������������������������������������������������������������������������������������������������������������������������������������������176 Bibliography������������������������������������������������������������������������������������������������������������������������������������������������������������178 Appendix Three�������������������������������������������������������������������������������������������������������������������������������������������������������������������180 Appendix Four: Assembling lead data for model�����������������������������������������������������������������������������������������������������������184 Appendix Five: Abaqus input decks���������������������������������������������������������������������������������������������������������������������������������190 Input deck for the lead column at 20oC, used to gain the data for the input deck for lead kettle.�����������������190 Input deck for the lead column at 40oC, used to gain the data for the input deck for lead kettle.������������������192 Input deck for the lead kettle�������������������������������������������������������������������������������������������������������������������������������������195 Online Content��������������������������������������������������������������������������������������������������������������������������������������������������� 200 Online Content�������������������������������������������������������������� avaliable at https://doi.org/10.32028/9781789697421-online

v

Acknowledgements to this publication There are far too many people to thank individually here but a special mention to my husband Christopher.

vi

Preface The following text constitutes the publication of my PhD thesis, submitted and awarded by the School of Engineering at the University of Bradford in 2007. The PhD expanded from a BSc dissertation that was submitted and awarded by the Department of Archaeology at the University of Bradford in 2002. Following the award of the PhD, research has continued, which has been subsequently published, but this is the first time that the contents of the PhD have been made available. The PhD material includes the previously unpublished full data that formed a basis for further research, such as the gazetteer of the dyeing apparatus and workshops in Pompeii. Every study is undertaken within a theoretical, cultural and technological context, which influences its approach and findings. This context is intangible and if left unrecorded will become lost over time. To explore and record the context of this thesis, a ‘Chapter Zero’ has been included in this publication, which follows this preface. This could only be written after time had passed, allowing reflection through hindsight. Throughout Chapter Zero, there is reference to the ‘study’ and the ‘thesis’. The ‘study’ is viewed as the whole study, beginning with the BSc, which then expanded into the PhD, then continued through further research until the present. The ‘thesis’ refers to the PhD only, part of the wider study. The PhD stands alone as a discrete body of work submitted for examination at PhD level in 2007. The thesis refers to itself as the ‘thesis’ throughout Chapters OneEight. At the time of writing the PhD (2002-2007), the BSc had been awarded (2002) and an article published from it (2005). It was not known at this time that further research would occur after the PhD was awarded, so reference to the ‘study’ within the thesis refers to only the BSc and PhD together or the PhD, depending on context. Heather Hopkins November 2019

vii

viii

Chapter Zero

Preface to the published thesis No study stands alone – all are affected by the context in which they are undertaken, which may not be explicitly recorded in the final work. This ‘Chapter Zero’ aims to give an overview of the context of this study so that these intangible factors that affected its approach and methods do not become lost.

This introductory chapter complements the thesis which provides more detailed explanation of term and ideas cited below, and a full list of references. How and why the approach of this study differed The study took a new approach, moving from a classical question into archaeology then on into engineering, first theoretically, then literally as the author moved from a Department of Archaeological Science to a School of Engineering. To ensure academic integrity, the study was supervised throughout by academics in Archaeology (experiment and textiles) and Engineering (materials and thermodynamics) both at the University of Bradford, and Classics (Roman Pompeii) at the University of Oxford. After graduating with BSc in Archaeology the author sat modules in engineering and mathematics in the School of Engineering at the beginning of MPhil. Applying experimental archaeology to understand Pompeii’s dyeing industry was new, but moving between departments to undertake the study was itself also an experiment. Each part of the study had to be accessible to supervisors with no background in that area, and the author, with no A level mathematics or physics, had to research and write a thesis in engineering. This explains why a thesis that would ordinarily have been five-six chapters in length is eight chapters and five appendices. As this includes the origin of the data used, explanations of engineering equations that are accessible to classicists and a publication of early results, the thesis is a whole greater than the sum of its parts and has been heralded as a good step-by-step introduction for anyone with a phobia to science.

This context shall be explored through: • • • • • •

Introduction How and why the approach of this study differed How location and time affected this study Overview of the study New approaches within this study Expansion since the doctorate was awarded

Introduction In the time that has passed since this thesis was written much has changed, allowing reflection on its findings and its context. This publication gives the opportunity to update it, presenting it with a more modern approach to a new audience who would receive it differently. But that would be a dis-service. This thesis stands as a single discrete body of work, submitted for examination at PhD level. It was written within the wider context of academic knowledge and belief of the time, which it now forms part of. This study’s findings withstood scrutiny from three academic fields, when ordinarily it would have been required to withstand only one. Any update now could be disproved in the future, while this study stands as it is, a solid foundation for further work. Each doctorate should ‘add to the body of humanity’ through findings and approach – the first question during the viva was whether this study had done that. This study used a new approach triangulated between three academic disciplines to answer a previously unanswerable question, while being undertaken in a university with a unique history from which the study could benefit. This also coincided with an exciting time in the evolution of technology, which included the advance of home computing and the birth of the internet. At the time of writing this thesis, these were ongoing developments with an outcome that was not yet known, so it is only now that it is possible to look back and place the study in the context of the time and place it was undertaken. Without doing this, the wider factors and events that affected the approach and findings would be lost.

The inclusion of multiple disciplines allowed the study to self-challenge, triangulate and calibrate its data and findings. Early criticism made the study stronger as each was deconstructed and robustly challenged. These diverse inclusions allowed this study to form a robust foundation and springboard for further work, moving in unpredictable ways, continuing to reassess and develop while remaining unsurpassed. The publication of this thesis allowed an opportunity to update the thesis, but it was decided that as it is a discrete body of work forming a single examination piece it should stand unchanged. Instead, this has been taken as an opportunity to explore what has altered within the study and its context. Experimental archaeology is a discipline in its own right, falling under the umbrella of archaeology. 1

Investigations into the Dyeing Industry in Pompeii Previous examples of applying physics to archaeological questions do exist, but the majority are so scattered, specific or dense that they cannot be used or applied to other artefacts. This study provided a broad walkthrough of a worked example that can be applied to other work and during planning of the replica it was discovered how much such examples are needed: participants from the school of engineering had to be informed that cooking potatoes in the firebox was inauthentic as the Romans did not have potatoes, while classicists had to be shown that painting wood to match brick aesthetically does not help find the fuel required to heat brick as wood and brick do not react the same thermally, or to a naked flame. In the School of Engineering, the study has inspired exam questions and dissertations in heat and air flow, as an unseen example that requires application instead of simple recitation from a textbook. The apparatus was found to be 20% efficient – the archaeologists felt this was wonderful, the engineering felt this was shockingly inefficient, showing the difference in viewpoint from different schools of the same finding. Dyeing in Pompeii was ‘pre-industrial’ but was still an industry, with skilled business professionals operating specific apparatus in dedicated workshops. A paradigm shift allowed archaeology to appreciate the skills involved and engineering the difference in scale.

artefact on the micro and macro level, but at the time this study began, moving to engineering was a new idea, part of a change in wider mindset without which experimental archaeology would not view ‘engineering’ techniques as accessible or normal. How location and time affected this study Location Where a study is undertaken is important. This study was undertaken at the University of Bradford and would not have progressed as it did if it had taken place elsewhere. Ordinarily, the experiment would have been situated on moorland owned by the university, meaning that all participants would require ferrying to site. The author could not drive so the experiment was sited on campus land within the boundaries of the inner city. The first apparatus was constructed in 2001, at the time of the Bradford riots and no objection was made by civil or university authorities to a small, controlled fire within a smokeless zone when the daily news was of other areas within Bradford already being alight. The experiment required full wool fleeces. The Wool Board storage depot of the North of England was on the road running through campus and due to a recent economic downturn they were giving fleeces away. This provided the wool and also meant it was possible to see the size and scale of the storage required for one year’s fleeces. Bradford was documented as having the cheapest cost of living in a university town or city in the country which meant that after AHRB (Arts and Humanities Research Board) converted to AHRC (Arts and Humanities Research Council) and cut funding of projects involving experimental archaeology it was possible to continue as a self-funded student. Although distressing at the time, this financial release gave the author freedom to move between academic departments, following the development of the study. Furthermore, the school of engineering specialised in automotive engineering and has existed since the 1960s, an era of change in automotive design, so possessed the original ergonomic data from a demographic match of Pompeii. The presence of the discipline of ‘forensic archaeology’ also allowed complementary interpretation of skeletal remains.

When Finite Element Analysis was introduced into the study to allow a digital reconstruction of the physical apparatus, the study became too technically complex to be supported in the Department of Archaeological Science and so moved to the School of Engineering. Issues with compartmentalising and publishing are problematic, especially as the first person in a new field has nowhere to publish – the analogy is the Wright brothers’ first flight being documented in ‘Gleanings in beekeeping’. Moving to engineering allowed an expansion and greater in-depth exploration of technical questions that had an overall impact on the dyeing industry. The linguistic differences between archaeology and engineering, both constructed on empirical science, highlighted differences in approach, method and how comprehensive answers could be: when asked the method of ‘How does a car work?’ an archaeologist could describe inserting petrol, while an engineer could lift the bonnet and describe the engine. When asked how a dyeing apparatus worked, an experimental archaeologist can add fuel to a replica and watch the heat move through it, while an engineer can see how the heat moves through the lead at a molecular level. This highlights the difference in mindset between understanding and approaching an artefact on the ‘macro’ (archaeological) or ‘micro’ (engineering) level. To construct a more complete picture, this study had to explore and combine both. Now it is possible that an experimental archaeologist would also explore an

The only immediate disadvantage with the choice of experiment location was that it was between a children’s nursery and a cancer research department, so releasing fumes and steam from chemicals boiled in lead was not feasible. The aim of reconstructing the apparatus was to see how heat transferred through it, so this issue was resolved by using a physical, thermal replica of lead: stainless steel. There was criticism that it was not an aesthetic match, but this was not required as this had no bearing on its physical behaviour. Modern 2

Preface to the published thesis

bricklaying mortar was used instead of lime mortar: this had known properties, was a thermal match to the Roman lime mortar and had the advantage of being less corrosive to skin. Later work, subsequent to the thesis, found that a stainless steel replica could not be used to explore creep or the chemical effect on dyes, but these factors were not part of the original question so the replica was valid when it was constructed. The closest previous reconstructions were by John Edmonds at the Chiltern Open Air Museum who explored dye chemistry and a replica of cooking apparatus from the Mary Rose constructed by the Mary Rose Trust as a public demonstration of cooking methods below decks, to inform and entertain. This study was the first to construct and use a replica dyeing apparatus from Pompeii, and was the first attempt to investigate cycle time: it was discovered that the apparatus could be used once per day, due to it taking at least eight hours to cool, used 7.5kg of pine to heat, and allowed 8.8kg of fleece to be dyed per person annually in Pompeii.

Now, having presented at conferences, it is possible to see the new generation of classicists, who even the most technophobic of which are used to having emails, facts, photos and answers literally ‘to hand’ through their phones and who have grown up seeing computer games. These same researchers now value network simulation and visual reconstructions as they have purpose and are familiar. Archaeological Sciences is respected and complements wider scientific disciplines, while experimental archaeology and digital humanities are now taught as disciplines in their own right, so other disciplines, including classics, are more accepting of them. There is still confusion as to the role of physical digital reconstructions, possibly because the extra specialism required still falls within the realm of engineering, not digital arts. There is still a reliance on older theories and the belief that an older answer written in the Victorian era or by Pliny must be right as it is already ‘established’, regardless of whether it withstands scrutiny, which means that digital studies may still be based on erroneous data. Moving away from physically testing artefacts raises the question of what has been lost and what may be lost in the future.

Study in the context of time and technological developments During the first week of the first year of the undergraduate degree of BSc Archaeology in 1998, the class were taken into the library, sat in front of computers and taught how to write an ‘e-mail’. None of us had used ‘the internet’ before. The author’s dissertation, written in 2002, was typed on a word processor and saved on 3.5 inch 1.44mb floppy disk, which unusually for the time were DOS compatible and so could be edited using a computer with Microsoft word. Sat typing at a laptop now, the last twenty years of work is backed up on a 16gb datastick. At the beginning of the undergraduate degree only students of computer science owned computers, but at graduation four years later the author was one of very few to not own one. Private internet use was also still beyond the finances of most students – to use the internet, students went into the library on campus to use the institution’s computers and internet connection. The author was ordered to buy a computer before beginning the MPhil in 2002 and bought a 20gb hard drive laptop affectionately called ‘Doorstop’ as the screen was easier to see – prior to flatscreen monitors this was a serious consideration. It was one of the last laptops to feature an inbuilt floppy disk drive. The author’s first data stick was bought in 2006 near the end of the PhD and during the world cup in Germany: after waiting for the technology to stabilise, a spherical football shaped datastick was bought, in the hope it would not get lost.

This study taught to beware of sweeping statements and belief that any finding, method or technology will be the last of its kind. In 2001 Monaghan said that contemporary dyeing recipes had not been found, a statement which may be true at the time, but becomes more unsafe as the internet becomes ubiquitous. This is a mixed blessing as data and publications that were thought to be lost and which were otherwise unavailable can now be uploaded and downloaded to be kept in perpetuity. In 2007, when this PhD was awarded, internet use was in its infancy with national institutions having the same online presence as private individuals writing unfounded opinions as ‘fact’, so relying on internet sources was an automatic academic fail. In 2017, when studying an unrelated MSc, the roles had reversed so that reliance on books and journals was deemed as failing academically as information gained from the internet can be continually updated and demonstrably linked to authenticated resources. ‘Digital’ photographs were also in their infancy in 2002, so the survey was recorded digitally and with film. The digital copies were initially allowed for publication but now early digital photographs do not have high enough resolution and instead scans of film photographs are sometimes used. This highlights how unpredictable the future is.

At the start of this study, the classicists involved were technophobic, distrusting any answer acquired through the aid of computing. This extended to Finite Element Analysis, hence the move to engineering.

True experiments, that seek to prove whether something was physically possible, can fail. In the 1960s the Ra I sank, revealing that the design had not been properly understood – the Ra II voyage was successful. The Sea

Development of experimental archaeology

3

Investigations into the Dyeing Industry in Pompeii Stallion, built in 2008, continued this tradition when it sailed across the Atlantic, finding that although the ship moved at speed it could not stop without resting against a quayside or on a beach as the design meant it would fall over if it moved too slowly. This contrasted with the contemporaneous Matthew, which sailed the Atlantic with great publicity, meeting the Queen when it docked. Public involvement in experiments of such large scale and cost mean that they cannot be allowed to fail, which arguably undermines their experimental quality, turning them into inaccurate replicas. Both ships had to meet modern navigation requirements. The Sea Stallion did this by sailing with a support ship, allowing the replica ship to be authentic. The Matthew, a replica of a 15th century ship, had satellite navigation fitted, which did not alter its handling, storage or living space, and an engine with fuel which did.

industry was so small that Pompeii relied on imports, but each study used the same survey, without challenge or review. Some authors, such as Laurence, side-stepped the issue by focusing on the economic context. The diversity of the findings demonstrated that a significant factor was missing from the wider understanding. This study realised that each author had been attempting to calculate the size and scale of the industry through a purely theoretical understanding of the processes involved applied to a superficial measurement of remains. When this study re-surveyed the remains, it highlighted a further issue of authors not revisiting the site with practical knowledge: Moeller had found 33 dyeing apparatus in 1976, Janaway and Robinson found 40 in 1994, this study found 35 in 2002, not including one that Moeller had erroneously included. This study began as an undergraduate dissertation, academic year 2001-2, then expanded into MPhil then PhD, awarded December 2007. The original aim of this study was to discover the size of the dyeing industry of Pompeii to allow an understanding of its economic role in Pompeii and Pompeii’s role in the wider Roman world. It was realised early on that many influencing factors were intangible and ephemeral, such as laws, economics, customs and social convention, each of which could remain undiscovered or change immediately with a new find. To develop an unchanging answer, a solid foundation, this study decided instead to develop an understanding of the physical capabilities of the apparatus, then define how each tangible or intangible factor could affect it. This study was the first to reconstruct and use a replica of a dyeing apparatus from Pompeii, to see its capabilities, limitations and method of operation. The figures gained, such as fuel required, temperatures reached and time taken, will stand until intangible findings are able to refine them, without time limit. The Romans did not dye whole fleece, use pine offcuts for fuel or dye with modern tap water – these were ‘holding values’, known analogies to be replaced when the real quantity of wool, calorific value of fuel or content of water were become known. It still is not known if wool was dyed in the fleece, yarn or fabric, or the rate of re-dyeing. The fuel type was believed to be charcoal on economic grounds, but practical experiment and engineering theory independently demonstrated that there was insufficient oxygen to burn it. Understanding the water quantity and type required allowed exploration of the water content, source and storage needs. The recipe was an amalgamation of pre-industrial methods derived from textile finds, texts and practice to dye wool with madder and alum, the most common materials believed to have been used, as discovered through contemporaneously textile finds from across the Roman world. These were deconstructed to find the common activation points – dyeing will not work without the correct time,

Modern experiments must meet modern health and safety requirements, which may limit or alter which experiments can be undertaken. FEA (Finite Element Analysis) allows replicas to be tested, with failure speeded up or slowed down, and poisonous or prohibitively expensive materials to be avoided. During this study it was not possible to make replicas from lead, but it was possible to use thermal replicas to get the data regarding heat transfer then digitally model them to be made from lead. Digitally it was also possible to change the shape and size of the apparatus, alter the ambient and firing temperatures, speed or slow time, then repeat dyeing cycles until the kettle broke, slowing the moment of breakage to understand how it happened. This would not be possible with a physical replica and shows a positive way forward for experiments to develop, allowing greater understanding overall. When this study was presented at an Experimental workshop in Edinburgh, two railway specialists commented that, ‘Lead doesn’t act like that.’ Their views as experienced craftspersons must be considered, but when asked they could not say how it would act instead. To date, no one has built a physical replica and used it to destruction – that is an experiment yet to be undertaken. Overview of the study Background to this study Prior to this study, all previous understanding of the dyeing industry of Pompeii was based on a single survey undertaken by Moeller, published in 1976. Moeller’s conclusion that Pompeii’s dyeing industry was large enough to export was immediately challenged by Wild who said that Moeller had not produced the evidence for this. Subsequent studies, such as Jongman, used the same survey evidence to demonstrate that the dyeing 4

Preface to the published thesis

temperature, pH, water content or kettle shape. The study began before widespread use of the internet so the author only had access to British publications or selected German or Italian translations. The scale of an industry can only be known from comparing its size to the size of population it supplies. After reviewing contemporary literature, this study based comparison on Pompeii having a population of 12,000 people, despite figures in the academic literature varying between 8000 and 20,000 people. Each new finding meant recalculating the industry, but this just made the study more accurate.

assessed, which allowed identification of apparatus that were missing features, had additions or had otherwise been amended. All remaining kettles were lead. Dyeing apparatus braziers were either flued or unflued. An apparatus that Moeller identified was discounted as it could not have held a kettle. Three apparatus were found to have been heightened, indicated by a change in cement and a geometry that would have resulted in too high a pressure on the lead kettle during use. Sympathetic conservation was identified where walls had collapsed and kettles cemented into place to prevent their loss.

Since completion this work has been presented and published widely. This has highlighted the importance of disseminating work, allowing other disciplines to evaluate it: mysteries have been remedied, mistakes realised and unrecognised assumptions challenged. Presenting has also allowed expansion, making a stronger study with fuller answers. One person’s mystery may be another’s daily routine, which can bring answers or spark new questions to answer together (see below).

Pompeii can be compared to a ship wrecked at sea: everything required for it to function had been present – the size and number of dyeing apparatus and workshops present would have been the number the city required, the number altering with economic viability. Archaeology explored the kettle geometry as a complementary interest, while engineering actively sought to understand it. The geometry of the kettles showed that they were of comparable design regardless of size, with comparable pressure on the base when in use. This appeared to be a deliberate design. Prior to inclusion of computer simulation into the study, this was calculated by making paper templates. It is unknown to what extent the dyers knew of creep and whether the chamfered edge of the brick support in some apparatus is an original feature. Later it was noted that the design matched purpose-designed kettle for dyeing yarn, not fleece or fabric – again it is unknown if this was deliberate. Several kettles had taps passing through the brazier which meant that they were emptied, cleaned and filled in situ.

New approaches within this study Surveying the original apparatus in 2002 Constructing and using a replica based on the 1994 survey gave a practical knowledge and findings that allowed the dyeing apparatus in Pompeii to be resurveyed in 2002. This formed the basis of expanding the dissertation (2002) into MPhil. The survey included where properties were in the city, the property size and layout, water supply, apparatus number and their location. These were measured accurately, documented in writing and recorded through photographs. If all factors influencing output could be known or allowed for, the process would then only be defined by time. Understanding the apparatus size and dyeing cycle time allowed storage requirements and supply process to both be explored, such as whether the industry relied on ‘Just In Time’ supply – organisational arrangements that mean only the immediate resources required for a manufacturing process are present, with next resources delivered ‘Just in time’ to allow the process to continue, meaning that storage requirements for each work station (e.g. dyeing apparatus) are kept to a minimum. Findings indicated that if dyed whole, space would be required for up to 54,000 fleeces annually to supply Pompeii, but in reality only the finer quality parts of fleece were dyed, suggesting storage requirements for a far larger number.

The survey in 2002: water supply Pompeii’s water supply is well preserved, resulting in diverse studies, one of which is used (by Hodge) to show the importance of returning to the original remains before reporting findings. Hodge is quoted at the front of this thesis. New findings should challenge or complement existing theory and sometimes there is no apparent alternative. Hodge and the author share learned behaviour that would have made using Pompeii’s lead-lined water supply safe: to ensure that the water is flowing fresh from a mains supply before drinking from it by running the lead contaminated water through the pipes first. In Pompeii the water supply was continually flowing, while also cleaning the streets as it overflowed from fountains, two actions that improved the population’s safety from lead poisoning and disease. Running the water for long periods before drinking it is a behaviour now lost to a modern generation who view it as wasteful, especially since the advent of water meters, showing how dangerous it is (theoretically and literally) to apply modern squeamishness or values to

Understanding operating parameters allowed a set of polythetic entities of a dyeing apparatus to be identified. Apparatus were surveyed and ergonomically 5

Investigations into the Dyeing Industry in Pompeii an ancient problem and solution. In 2002 during the survey in Pompeii, when faced with a choice of drinking from the flowing lead-piped fountains or the modern supply containing stationary sections and sewage from a recent flood, the author chose to drink from the lead. Like witnessing the birth of the internet, the author has also witnessed the loss of lead use in plumbing and the learned behaviour that made it safe – the author’s housemates did not know to run the water in a Victorian house with suspected lead pipes in 2002 and felt it wasteful until they realised.

bridging data, that matches the skeletal data but comes from a living population, must be used. Coincidentally, Pompeii and Herculaneum’s populations matched the USA population from 1900-1960. In the 1960s, through improved food and lifestyle, the USA population increased in height. This coincided with an increase of women driving and entering the workforce. To cater for this, there was a sudden need for an ergonomic understanding of an increasingly diverging population. One result is that there is now a complete dataset available that can be used to represent Pompeii’s population, gathered from car manufacturers, the World Health Organisation (WHO) and the US army.

Originally it was believed that the Roman population were at risk of lead poisoning through their water supply. During the 20th century this was disproved through a better understanding of the water supply, focusing on the use of hard water which coated the lead pipes with cinter. Subsequent excavation in Pompeii and newly available tests undertaken since this thesis was submitted have again disproved this by showing that skeletons in Pompeii did contain lead which is believed to have accumulated before cinter was able to develop after new lead pipes were added to the system (e.g. Keenan-Jones et al, 2011). This does not change the context of Vitruvius’ findings: lead workers are notably unwell through working with lead, but dyers are not. It is probable that lead leaching from new sections of pipe into the water supply and being absorbed by the population was at such a low level that its accumulation in the population went unnoticed. It is possible that Pompeii’s population may have absorbed lead from other sources.

During the survey of dyeing workshops undertaken in 2002 as part of this study, some of the dyeing apparatus were found to have steps. These could only have been used by children. This finding was rejected for publication (Papers for the Institute of Archaeology in 2013) because it was already assumed that children had been present, so publishing evidence demonstrating that they had was felt to be unnecessary. This shows the need for publication of all findings – such presumptions remain either unwritten or without evidence and unchallenged. This study is the first to identify steps at all, let alone to provide physical evidence of the presence of children. Finding that there had been children in the dye works fits Temin’s theory that dyers were released on contract and remained within the workshop – close to their family and prevented from competing. This poses the question of the financial and social value of dyers, and if this was altered by needing additional business skills. An example cited is Pliny’s speech writer who although a slave was sent to recuperate from illness at Pliny’s country villa – the treatment of a skilled but stained professional slave or freedman is unknown. The skill required to dye accurately without thermometer, pH meter or precision control over the apparatus should not be underestimated – theory changes but the skills required do not. Children born to dyers would grow up within the environment proficient in dyeing, becoming skilled adults.

The survey undertaken in 2002 allowed the location and provision of public fountains to be explored in relation to the location of the dyeing workshops, the water storage requirements and possibilities of workshops to be investigated, and the identification of drainage slots in workshop doors that allowed dye liquor to be disposed of into the streets. The liquor would have been washed away by the fountain overflow. Further mysteries about water supply remain, such as pipework being found below a dyeing workshop (by Borgard): it is not possible to tell where it runs from or to, just that it is present.

Saying that a Roman dyer cannot be found but a Roman can sounds reminiscent of Functionalism and Middle Range Theory: assuming that a modern person will respond to a situation with the same logic as someone in the past. This study produced a physical replica of a Roman, not a skills replica. Again, like experimental archaeology overall, it does not show what happened but what could have happened – it gives boundaries to what is physically possible, which the true answer will be contained within. These boundaries should also allow that dyers would have been acclimatised to the dye works and that WHO recommendations may be broken in practice, meaning that dyers may have managed more than theorised possible. But conversely,

The survey in 2002: finding a ‘Real Roman’ One criticism levelled at the study was, ‘You can build what you like – you can’t find a real Roman to use it.’ Answering this was irritating at the time, but allowed a better analysis and understanding of the Roman population and the apparatus through an ergonomic assessment, leading to a stronger outcome. Skeletal data from Pompeii and Herculaneum was reviewed and the forensic application of ergonomics was examined. Ergonomic data is not the same as skeletal data – 6

Preface to the published thesis

dyers may not reach their full genotype height if a reduced dietary intake as a slave stunted their growth and a modern individual may have a more sedentary lifestyle than a Roman dyer or American driver from the 1960s. An allowance for differences in capabilities between individuals must be made.

A wider question posed by the alteration of the apparatus is what ‘restoration’ is and means. During the 1970s in the UK architectural restoration had to be obvious, to ensure it was not mistaken as original. This leads to examples, such as Tattershall College, where the restored areas are brighter and more robust to erosion than the original, overwhelming it visually. Now reconstruction is more muted, but can be mistaken for the original. Examining the apparatus showed that historic reconstruction was completed by people who did not understand the apparatus but who were sympathetic to the remains and without their work the remains would not have survived well enough to be examined.

The survey in 2002: defining ‘Recording’, ‘Conservation’, ‘Reconstruction’ Undertaking the survey of dyeing workshop in 2002 highlighted a glaring omission: no complete plan of Pompeii existed until 2004. Pompeii is so large that each part had been surveyed individually, but a single detailed plan could not be drawn. In 2004 a satellite photograph and single plan drawn from the photograph were published. It was so large that standard computers of the time would crash when attempting to open it. This study had access to this plan and satellite image, but as this came through the AAPP (Anglo-American Pompeii Project) it no longer does. The new version was accepted at the time unquestioningly – it was forgotten that any new technology is only as accurate as its programming. When comparing the digital plan with the satellite photo and the survey from this study it was found that some walls, doors and windows had been moved, which was problematic for exploring storage space and airflow. This highlighted issues in older plans: maps showed fountains in different locations than where they were in reality and there was no standard way of recording multi-storey buildings or architectural features that had been known to be present but since lost. Today technology has moved further and now plans of this size may be multi-layered and 3-dimensional visual reconstructions may be devised that allow viewers further levels of understanding as they ‘walk’ through the city, with computers able to support this.

‘Restoration’ overlaps with ‘conservation’ and ‘recording’: each may be of questionable accuracy and alter future understanding of remains so are a mixed blessing. The survey required a faithful record and representation of misunderstood remains. Engineering required photographs and a design schematic to record what the apparatus currently looked like, any visible changes and to show how it could have operated, allowing highlight of alterations or missing parts. This was rejected in archaeology where it was felt a drawing of each apparatus would be more accurate and last longer. In addition, the ivy present on the apparatus should be drawn too even though if the apparatus had been used with it present it would have burned instantly. There were similar arguments over whether the apparatus should be painted white – there is no evidence that they were, but Borgard’s reconstructions were painted white. Undertaking the survey and wider research also highlighted recording mysteries that had either been overlooked or presumed endemic knowledge. Borgard speaks of ‘The House of the Queen of England’ but this was not shown on any plan this author had access to so its location remains a mystery and conclusions made about it unusable. This house may be known by a different name and already included in this study. Borgard also refers to newly discovered pipework which is not on any plan, is of unknown age and flow direction. This demonstrated the need for a single comprehensive map of Pompeii, but highlighted difficulties such as what to include in such a monumental work and how to show features that were known or believed to exist but have been lost. It is not possible to reconstruct a feature that there are insufficient records of and likewise it is not possible to excavate twice to find remains that were removed or discarded in an earlier excavation as unimportant or unrecognised. Differing styles and priorities mean that some evidence or working are not published, just the end result, so data and calculations cannot be checked. This was problematic when reconstructing Pompeii’s population through its skeletons during this study and when reviewing Borgard’s reconstructions during and after this study. Differing priorities causes

In 2002, this study undertook the last survey of remains. The apparatus were compromised through ‘reconstruction’ by 2010 when replica kettles were cemented to apparatus walls, altering airflow and heat transfer compared to the original remains. The survey and digital physical reconstruction have allowed the study to fulfil a criteria of UK commercial archaeology, to ‘Preserve By Record’ (Hopkins 2010). In commercial archaeology the artefacts are faithfully recorded, with the artefact and records going into the national archive where they remain unaltered, but in Pompeii the artefacts remain in situ, in their altered state without accompanying explanation, allowing future viewers to be misled into thinking they are seeing unaltered originals. After twenty years of erosion they will appear original or a sympathetic conservation. This does raise the question of how original the remains were when reconstructed in this study, but this was allowed for when examining them. 7

Investigations into the Dyeing Industry in Pompeii relevant subjects to be excluded as their importance is not understood: Flohr stated (Pers. Comm 2018) that defining the physics of how the apparatus worked was completely irrelevant as it did not show the social issues involved. This assertion missed that other researchers may be studying the physics and that the physics gives a solid framework that intangible, changeable subjects, such as social issues, fit around – only by knowing how a dye works operated is it possible to see how many people would be required.

and the first to model temperature. FEA had been used only twice before: modelling of strength in chimpanzee wrist bones and modelling pressure fractures in amphorae. Abaqus was the FEA program used to model the dyeing apparatus. This computer program has been used for modelling aircraft wings since 1982, so had a proven record of modelling metal failure. Modelling lead proved problematic as the data required did not exist – a study from 2000 that explored the compression strength of lead was used and data was extracted and inserted into the model by the programmer Mark Robinson.

Redefining ‘experiment’ When reconstructing the apparatus, where to place the flue led to a redefining of ‘experiment’ for this study. The replica was unflued, copying an unflued original apparatus. To replicate a flued apparatus would mean adding a replica flue to the replica apparatus. A matching flued apparatus was found but when reviewing the survey it was discovered that apparatus with flues also had taller fire boxes as part of their overall design. In an unflued apparatus, the exhaust gases rose from the firebox, upwards between the supports holding the kettle aloft, then through the gap between the kettle and the brazier wall, being released at the top under the kettle flange. In the flued apparatus, some exhaust gases were lost this way, but the majority were drawn upwards between the supports holding the kettle aloft then upwards into the opening of the flue, which was at the bottom of the kettle. To add a flue to the unflued apparatus at the correct height but leaving the fire box unaltered would mean the flue would open into the side of the kettle, so could not draw gases into it. The apparatus would not work and it would not replicate the design of an apparatus – the experiment would be void. To alter the firebox before adding the flue would mean making two changes to the apparatus. An experiment only changes one thing at a time – the decision was made to amend the flue height to continue the experiment. The aim of the experiment was to explore the design of the apparatus, so the flue was put at a height to be workable, even though an apparatus has not yet been discovered that matches these dimensions. A matching apparatus may yet be discovered – one third of Pompeii has yet to be excavated and other examples may be found elsewhere in the Roman world. Further experiments could alter the design incrementally allowing changes to be understood. The presence of taller fireboxes on flued apparatus suggests that there had been difficulties with ventilation of the originals. The survey appears to suggest that flued apparatus were in more enclosed environments than unflued apparatus. This issue was presented at ExArc 2008 which has yet to be published. Further works are developing this idea.

FEA and a wider understanding of the properties of lead showed that not only had lead been chosen for its chemical properties (see below), but that the wider physical properties protected it. Lead has a different grainsize to other metals, which allows movement during temperature change, meaning that the lead will creep but during cyclical creep/strain it will not break. Instead lead will display a ratchetting effect, demonstrating that the same stress can cause different strain in different metals. If the change in lead doesn’t exceed 4% it will not break. The significance of this factor in Roman dyeing is new.

Finite Element Analysis

Moving to the School of Engineering allowed engineers to explore Pompeii’s dyeing industry in new ways. One was to draw back from the apparatus and explore its context within the dyeing workshop and each property

When collecting data from the physical replica to use in the FEA model a counterintuitive change had to be allowed for: as fuel was added the fire initially got colder. This was because the fire decreases as the energy is used to break up the wood releasing new energy. As each dyeing cycle differed slightly, in ambient temperature, fuel size, rate of combustion for example, taking mean temperatures at certain times would not provide an accurate dataset for the model. Instead the data was examined for its activation points – tasks, times and temperatures – and a representative amalgamation was used in the model instead. Finite Element Analysis is now used widely to test the design of objects before they are constructed physically, avoiding cost, risk and time if they fail. The test can be repeated to see accumulative damage and the object is constructed in a controllable environment, allowing external changes such as ambient temperature to be modelled. This study was able to use the digital physical replica to repeat the dyeing cycle of heating and cooling, within different temperatures of between 20oC to 40oC, mimicking the use of dyeing apparatus in Pompeii. The FEA was judged to be accurate as it replicated changes in the lead and the apparatus discovered in Pompeii. ‘Roman’ vs ‘Modern’ manufacturing

This study constructed the first physical digital model of an archaeological artefact of more than one material 8

Preface to the published thesis

as a whole. It was realised that each workshop was set out as a modern ‘U-shaped’ layout devised by Toyota in 1985 for vehicle production. This allowed parallel workstations (individual dyeing apparatus) to operate within a single space, while simultaneously under observation by a single person. The quantity of storage at each property suggested the use of ‘JIT’ (Just-In-Time) delivery and processing. For example, harvesting fleece was an annual occurrence and storing up to 54,000 fleece before use, not counting fleeces that would not have been dyed, would require a larger area than that offered in the city’s workshops – they would have been stored elsewhere and brought to the workshops for processing.

have been linked to the next process, economically or by ownership, each producing, storing and feeding into the next stage. A delay in one workstation would not delay the whole workshop, a delay in one workshop would not delay the whole city as other workshops could supply the next stage. Ownership of materials and workshops would have structured production. This could have increased production making it as efficient as possible. When dyeing, each stage is dependent on the last. To dye, fleeces should be rinsed after mordanting, but still be damp. Starting up a dyeing cycle requires 1-3 people, depending on size of apparatus and experience, but once all dyeing apparatus are alight only one person is required to watch them. Using the experimental replica suggested that if each apparatus could be used once per day, staggering start up would mean all dyers were at work, able to process each apparatus in turn as it finished, creating most output with fewest labour and people. This raises the question of the role each person took, whether there was a single dyer overseeing a workshop with an assistant and if the dyer simultaneously managed ‘front of house’, the customer interface and book keeping. Workshops Vi4 and Vi5 were on to the street so the public would meet the dyer, while other workshops were located within properties away from public gaze.

Modern manufacturing uses three main methods: continuous, batch and one-off. Continuous was not possible in Roman Pompeii – it is the method by which food is manufactured and packaged on a conveyor belt. ‘One-off ’ specialist craft production may have occurred but not been economically feasible for everyday production, ‘Batch’ production was most likely to be the method used – manufacturing a quantity together, such as pots within one kiln firing. Defining the method of manufacture is possible from understanding the remains, but ownership of the materials involved at each stage would dictate the rate of dyed material produced. If the dyers owned the material to be dyed, they could dye in batches, restocking as required. This may have allowed for slight variations in dyeing outcome to be acceptable. If private individuals owned the materials and requested specific colours, greater precision may be required from the dyer, otherwise the owner may reject the outcome or require it to be redyed. The question of ownership affects the quantity of material that could be dyed and would need to be answered for the output of the industry to be fully understood but as yet is intangible. This harks back to the aim of the study: to create an objective physical framework that can be used to explore the subjective questions involved.

Intangible questions Economic, social and pragmatic questions about dyeing operations highlight the intangible side of an industry. This study sought to develop a physical framework and understanding, but the industry would have operated in an intangible ever-changing framework of law, economics, material supply, culture, social norms and fashion, each of which will remain unknown or change as new discoveries are made. Only the contemporaneous public and dyers of Pompeii would know these endemic entities and their effect may differ between people and workshops. An example would be the learned behaviour that makes using a lead-containing water supply safe to use. Pompeii’s populace did not need to know this, as the water supply of the city was constructed with the behaviour in-built – the water was continually running so was mainly uncontaminated by lead. As discovered by the author, this behaviour is nearly unknown in the modern era, giving an example of how a modern archaeologist may not understand how to use a hazardous artefact is safely, instead coming to an erroneous understanding that cannot be challenged as the knowledge has been lost.

The use of ‘U-shaped’ dyeing workshops and ‘JIT’ manufacturing was called ‘old’ and ‘out of date’ by some engineers in 2007. This highlighted again the difference in mindset: it had been forgotten that the Romans had empirically devised a system that it took manufacturers more than 1900 years to devise a second time. Line balancing is used in modern manufacturing and was also used in Pompeii. Line balancing is understanding the time taken for each part of a process, then using parallel workstations to complete tasks that take a longer time, so that no workstation is left idly waiting and the whole manufacturing process is as efficient as possible. Each workshop appears to have had multiple apparatus and specialised in a single task overall, meaning that line balancing occurred not within a single property but between properties across a city. Each workshop would

The legal and financial framework of Pompeii will remain broadly unknown. This study had to rely on Diocletian’s edict of 301 AD for a basis which could only be used to compare relative prices due to its being written more 9

Investigations into the Dyeing Industry in Pompeii than 200 years after Pompeii was buried. Temin likens understanding finance in the Roman world to trying to understand the workings of the London Stock Exchange from reading a single copy of the Financial Times. The Roman civilisation had an empirical understanding of economics without the modern vocabulary to describe it, but much remains lost.

inherently unstable when heated or placed under load. Engineering had no data about the loading properties of lead as there was no modern use for it. Monteix and Pernot had confirmed through analysis that the kettle was lead and the survey in 2002 had found that all surviving kettles had been lead – they kindly sent their unpublished data confirming it. Experiment design and understanding of outcome is affected by background and approach of the designer. This was demonstrated in 2007 when the engineers and archaeologists could not believe lead was used, despite evidence from the survey, contrasting with finding in 2012 that lead was the ‘neutral kettle’, that reliably would not alter the expected outcome and had actively been sought for use as a kettle. Experiments replicating dyeing in kettles constructed from different metals found that iron and copper resulted in altering the colour outcome, while lead and in particular lead with an oxide layer on the surface, were ‘neutral’ resulting in a bright colour, the choice of dyers. Lead had been dismissed in 2007 as a horrendous choice due to its physical properties, but in 2012 at European Textile Forum, held as the LEA (Labor für Experimentelle Archäologie) inaugural event in Mayen, lead was found to be the perfect choice chemically. The influence of mordants and dyes had been previously investigated and understood, and although dyeing handbooks (including in the UK) note that use of iron or copper as a mordant would affect dyeing outcome, this study provided exploration of the phenomenon of the difference kettle material made. This appeared to be the first study to focus on the kettle material. Theoretically, premordanted fabric should not be affected by kettle material as only the dye should bind to the mordant, but empirically it does. Findings were published as Kania et al 2014. An addendum took place in 2013 that highlighted the need for reference samples and in 2016 the experiment was repeated broken into stages. This showed that metals absorbed during mordanting, during dyeing, during both procedures when wool is exposed to both, but that most absorption happens during dyeing.1 LEA is part of RGZM (Römisch-Germanisches Zentralmuseum) and now hosts European Textile Forum annually – these ongoing experiments take place in Germany as part of European Textile Forum so at the time of writing ‘Brexit’ is causing great interest.

Fashion remains another endemic entity and links to wider economic issues. Fabric was overdyed and redyed but the rate of this was unknown. Dyers were skilled but the rate of failure of dyeing outcome (e.g. ruining the fleece, dyeing an unwanted colour) was unknown. It may be possible to devise a figure from modern historic dyers, but without knowing the ownership of materials in Pompeii the rate will remain unknown. New discoveries that explore empirical entities are welcome, as they may increase, limit or change findings. Expansion since the doctorate was awarded This study was presented to three Experimental Archaeology conferences where it was well received. After this it was presented at the first European Textile Forum in 2009, which specialises in presenting new textile-related questions and findings to academics, professional re-enactors and skilled enthusiasts to explore together. The science, archaeology and textiles knowledge were valid, but this study had not included the knowledge, skills and experience of a professional historic dyer. The study had begun before the widespread use of the internet so was built on works published in English available in the UK. The dyer, Sabine Ringenberg, physically demonstrated there and then that with the same equipment but a different method it was possible to produce triple the dyed output. The study had already found demographic Romans that could use the apparatus but now it had found a Roman dyer to use it properly. This demonstrated the importance of including craft knowledge from the beginning of the study, which had not been appreciated before. It showed the importance of presenting to diverse audiences, as one person’s mystery or overlooked factor is another’s daily routine. Sabine also noted that despite the assertions of the British authors, a professional dyer would not dye the whole fleece, just the best quality parts, which would alter the fleece requirement and quantity of textile produced, especially as soft furnishings were dyed whole while clothing could be undyed with features picked out in dyed wool. Sabine’s reaction to the study also showed the importance of presenting before the craftspeople with skills and learned behaviours are lost.

The choice of materials for replicating the apparatus in the laboratory has been queried as brass and other alloys were contemporaneously available in Pompeii but this study used only elements. Elements were chosen as alloys can have an infinite variation of ratios of the constituent metals and would require a correspondingly large group of samples to test the effect of each one. Madder, wool and alum were used as they were the most commonly used materials for

The professional historic dyer noted that the dyeing kettles had been made from lead. During the survey the archaeologists had noted the presence of lead as a curio, the engineers with disbelief and dismissal as lead was so

The whole experiment was repeated in stages in 2018, in order to collect fresh samples of each type to allow complete dye analysis.

1 

10

Preface to the published thesis

dyeing in the Roman civilisation contemporaneously to Pompeii, with samples found in Israel and Vindolanda. Further experiments took part with madder and birch, crossing over to birch through changing one variable at a time. Birch showed colour subtleties that madder could not, so was deemed more suitable for experiments investigating the effects on colour variation. Birch has been reliably available to participants in the Textile Forum experiments, allowing consistency. It is arguable that birch was available to Roman dyers – birch was growing in northern Europe – but whether it was available to dyers in Pompeii is more questionable. Like other quantities in this study such as using pine offcuts for fuel, birch gives a ‘holding’ value to be used until the actual dyes used in Pompeii can be determined and substituted in. Repeating the experiment also highlighted the need for repetition in experiments – each dyeing batch is slightly different as plants contain natural variations depending on many factors, which can lead to very different results and possible misinterpretations. The first experiment using birch resulted in a yellow yarn with a reddish tinge – this tinge has not been replicated since.

Hopkins, H. 2015. The importance to archaeology of undertaking and presenting experimental research: a case study using dyeing in Pompeii. In: Archaeological and Anthropological Sciences. March 2015, Vol 7, Issue 1 pp131-40. http://link.springer. com/article/10.1007/s12520-013-0159-y Katrin, K. Hopkins, H. Ringenberg, S. 2014. Investigating the influence of the kettle material on dyeing in the industry of Pompeii. EXARC Journal Digest 2014, Issue 2, pp 6-9. http://exarc.net/history/exarc-journaldigest-2014-autumn-printed Kania, K. Hopkins, H. Ringenberg, S. 2014. Investigating the Influence of the Kettle Material on Dyeing in the Industry of Pompeii. Presented at the 7th Experimental Archaeology Conference, Cardiff, 2013. EXARC 2014/3. http://journal.exarc.net/ issue-2014-3 Hopkins, H. J. 2013. Reconstructing the Dyeing Industry of Pompeii: the Importance of Understanding the Dyers’ Craft within a Multidisciplinary Approach. Nesat XI The North European Symposium for Archaeological Textiles XI. 10-13 May 2011 in Esslingen am Neckar. VML Verlag Marie Leidorf GmbH. Hopkins, H. J. 2013. The importance to archaeology of undertaking and presenting experimental research: a case study using dyeing in Pompeii. In: Archaeological and Anthropological Sciences. Proceedings of the 6th Experimental Archaeology Conference. October 2013. DOI: 10.1007/s12520-0130159-y Ed: Hopkins, H. J. 2013. Ancient Textiles, Modern Science. Proceedings of the First and Second European Textilforum. Oxbow. Hopkins, H. J. 2013. Reconstructing the dyeing industry of Pompeii through experimental archaeology: the challenges and rewards of a new approach. In: Ancient Textiles, Modern Science. Proceedings of the First and Second European Textilforum. Ed: Hopkins, H. J. Oxbow. Hopkins, H. J. 2012. Discussion on archaeological reconstruction in situ: a view of Heather Hopkins (UK). EXARC Journal Digest 2012. http://journal.exarc.net/issue-2012-2/mm/ discussion-archaeological-reconstruction-situ Hopkins, H. J. 2012. Discussion on archaeological reconstruction in situ: a view of Heather Hopkins (UK). EuroREA, Journal of EXARC, ‘International Association of Archaeological Open Air Museums’. 2012/2. http://journal.exarc.net/issue-2012-2/ mm/discussion-archaeological-reconstruction-situ Hopkins, H. J. 2012. Review of ‘Experimental Archaeology’ by Johns Coles, (2010 reprint). EXARC Journal 2012/1. ‘International Association of Archaeological Open Air Museums’. http:// journal.exarc.net/issue-2012-1/mm/experimentalarchaeology-john-coles Hopkins, H. J. 2012. Textilforum 2010. Conference review in EXARC Journal 2012/1. ‘International Association

The importance of using lead as a kettle would have been unrecognised by the archaeologists without the input of the dyers and professional dyers cannot know every facet of enquiry without archaeologists presenting it. This raises the question of what other influences the dyers and archaeologists do not yet know of and which specialism could reveal yet more about the industry. Currently further experiments and analysis are being undertaken from older samples made in 2012 and fresh samples gained in 2018. Presentations and publications since this thesis was submitted Since graduation, this study has formed the basis for further research, presentation and publication. Please note that some items may have similar titles but were rewritten with a new focus for difference audiences. Further information may be viewed at: http://bradford. academia.edu/HeatherHopkins Publications Eds: Hopkins, H. J. and Kania, K. 2018. Ancient Textiles, Modern Science II. Oxbow. Hopkins, H. J. and Kania, K. 2018. Introduction. In: Ancient Textiles, Modern Science II. Oxbow. Kania, K., Hopkins, H. J. and Ringenberg, S. 2018. The influence of metal kettle materials on the mordanting and dyeing outcome. In: Ancient Textiles, Modern Science II. Oxbow. Chapter 8: pp 99-121. Hopkins, H. J. 2018. The supply of water to the dyeing workshops of Pompeii. In: Ancient Textiles, Modern Science II. Oxbow. Chapter 9: pp 122-140. 11

Investigations into the Dyeing Industry in Pompeii influence output. PhD thesis, University of Bradford, 2007. Hopkins, H. Willimott, L. Janaway, R. Robinson, D. Seale, W. 2005. Understanding the economic influence of the dyeing industry in Pompeii through the application of experimental archaeology and thermodynamics. In Scientific Analysis of Ancient and Historic Textiles, Informing Preservation, Display and interpretation Eds R. C. Janaway and P. Wyeth. Archetype Publications, London.

of Archaeological Open Air Museums’. http:// journal.exarc.net/issue-2012-1/mm/conferencetextilforum-2010 Hopkins, H. J. 2011. Extending the paradigm to answer the ‘unanswerable’: a case study using Roman dyeing. In: Experiments with past materialities. Eds: Gheorgio and Children, 2011. BAR international series. S2302. 2011. Proceedings of the 14th Annual General Meeting of The European Association of Archaeologists 2008, Malta Hopkins, H. J. 2011. Creep of lead dyeing vessels in Pompeii: the use of finite element analysis to answer ‘unanswerable’ questions. Archaeometry, ‘Research Laboratory for Archaeology and the History of Art, Oxford’. Vol 53, Issue 6, pp 1231-1248, Dec 2011. First published online: Apr 2011 Archaeometry, 53:  no. doi: 10.1111/j.1475-4754.2011.00593.x Hopkins, H. J. 2011. Using experimental archaeology to answer the ‘unanswerable’: a case study using Roman Dyeing. In Reconstructions Recreating science and technology of the past. Ed: Staubermann, K. National Museums Scotland. pp21-49 Hopkins, H. J. 2011. Reconstructing the past to prevent future loss: the dyeing industry of Pompeii. MRS Proceedings, 1319, mrsf10-1319-ww08-03 doi:10.1557/opl.2011.740. Proceedings of the MRS (Materials Research Society) 2010 Fall Meeting, Boston, Massachusetts. Symposium WW – Materials Issues in Art and Archaeology IX. MRS Online Proceedings Library (2011), 1319: mrsf10-1319-ww08-03 Hopkins, H. J. 2011. Creep of lead dyeing vessels in Pompeii: the use of finite element analysis to answer ‘unanswerable’ questions. Archaeometry, ‘Research Laboratory for Archaeology and the History of Art, Oxford’. Online preview: http://onlinelibrary.wiley. com/doi/10.1111/j.1475-4754.2011.00593.x/abstract Citation: Hopkins, H. 2011. Creep of Lead dyeing vessels in Pompeii: the use of finite element analysis to answer ‘unanswerable’ questions. Archaeometry, 53: no. doi: 10.1111/j.1475-4754.2011.00593.x Hopkins, H. J. 2010. Reconstructing the dyeing industry of Pompeii through experimental archaeology: the challenges and rewards of a new approach. Published online as part of proceedings of the First European Textilforum, Openlucht Museum, Eindhoven. www. textilforum.org Hopkins, H. J. 2008. Using experimental archaeology to answer the unanswerable: A case study using Roman dyeing. In Experiencing Archaeology by Experiment. Proceedings from the Second Conference of Experimental Archaeology. Eds: Heeb, J. Paardekooper, R.Oxbow Book, Oxford. Pp103-118. Hopkins, H. J. 2007. An investigation of the parameters that would influence the scale of the dyeing industry in Pompeii An application of experimental archaeology and computer simulation techniques to investigate the scale of manufacture of the dyeing industry and the factors that

In Press Hopkins Pepper, H. J. The role of interdisciplinary experiments in developing untested theories: The application of engineering and digital approaches in the study of Roman dyeing vats from Pompeii’ In: Grana, L., Ivleva, T. and Griffiths, B. (eds.) Forthcoming A Guide to Roman Experimental Archaeology. Bloomsbury. Hopkins, H. J. Experience versus Experiment: differing disciplines’ definitions leading to the answering of ‘unanswerable’ questions, a case-study using Roman dyeing. In Proceedings of the Third Conference of Experimental Archaeology, Edinburgh University. Conferences presentations Archaeology Sciences UK 2019, 24-26th April 2019: Vavle, A. Taylor, G. Hopkins Pepper, H. and Kania, K. Presented by Vavle, A. Poster presented: Micromorphological analysis of textiles fibres from experimental dyeing vats. European Textile Forum. Laboratory of Experimental Archaeology, Mayen (part of RGZM RömischGermanisches Zentralmuseum, Mainz). 2018: Paper presented: Reconstructing dyeing in Pompeii: molecular answers to industrial questions. TRACamp 2018. Vindolanda: Paper presented: The diverse role of experiments in reconstructing dyeing. RAC/TRAC 2018. Edinburgh University. 2018: Paper presented: Contrasting the roles of experience, experiment and expertise in experimental archaeology: a case study of reconstructing the dyeing industry of Pompeii Tenth Experimental Archaeology Conference EAC 10. ExArc. Leiden University. 2017: Paper presented: The effect of discipline bias on the accuracy of reconstruction: a case study of dyeing in Pompeii Seventh European Textilforum. Laboratory of Experimental Archaeology, Mayen (part of RGZM Römisch-Germanisches Zentralmuseum, Mainz). 2016: 12

Preface to the published thesis



Paper presented: Potatoes, paint and photographs: why making mistakes is vital to accuracy, a case study of the dyeing industry of Pompeii. Sixth European Textilforum. Laboratory of Experimental Archaeology, Mayen (part of RGZM RömischGermanisches Zentralmuseum, Mainz). 2015: Lightening talk: The supply of water to the dyeing workshops of Pompeii. Lightening talk: Who were the dyers? Populating the dye works according to Strabo. Lightening talk: The camera never lies? Even with modern technology, why it’s always best to check things that are ‘accurate’.

Experimental Archaeology Workshop. National Museums Scotland, 2009: Paper presented: Using experimental archaeology to answer the ‘unanswerable’: a case study using Roman Dyeing. First European Textilforum. Openlucht Museum, Eindhoven, 2009: Paper presented: Reconstructing the dyeing industry of Pompeii through experimental archaeology: the challenges and rewards of a new approach. Third Experimental Archaeology Conference. University of Edinburgh, 2008: Paper presented: Experience versus Experiment: differing disciplines’ definitions leading to the answering of ‘unanswerable’ questions, a case-study using Roman dyeing.

Seventh Experimental Archaeology Conference, University of Cardiff, 2013: Paper presented: The influence of the dyers’ craft on experimental context: investigating the affect of metals in the dyeing industry of Pompeii.

Fourteenth Annual General Meeting of the European Association of Archaeologists. University of Malta, 2008: Paper presented: An investigation of parameters that would influence the scale of the dyeing industry in Pompeii. (Presented on my behalf by Roeland Paardekooper).

The Historical Metallurgy Society: Research in Progress 2012. Newcastle University: Paper presented: The unforeseen consequences of Roman metal choice: The far-reaching influence on dyeing in Pompeii and modern experimental archaeology.

Second Experimental Archaeology Conference. University of Exeter, 2007: Paper presented: An investigation of parameters that would influence the scale of the dyeing industry in Pompeii An application of experimental archaeology and computer simulation techniques to investigate the scale of manufacture of the dyeing industry and the factors that influence output.

Third European Textilforum. Laboratory of Experimental Archaeology, Mayen (part of RGZM RömischGermanisches Zentralmuseum, Mainz). 2012: Paper presented: An investigation of parameters that would influence the scale of the dyeing industry in Pompeii: Background to the dyeing experiment. Introduction to experiment undertaken during Textilforum 2012.

AHRB Research Centre for Textile Conservation and Textile Studies First Annual Conference ‘Scientific Analysis of Ancient and Historic Textiles: Informing Preservation, Display and Interpretation’, Winchester, 2004: Poster presented: Understanding the economic influence of the dyeing industry in Pompeii through the application of experimental archaeology and thermodynamics.

Interface 2011. The 3rd International Symposium for Humanities and Technology. University College London, 2011: Paper presented: Unlocking the potential of virtual replication in experimental archaeology NESAT XI, Esslingen 2011. (North European Symposium for Archaeological Textiles): Paper presented: Reconstructing the Dyeing Industry of Pompeii: the Importance of Understanding the Dyers’ Craft within a Multidisciplinary Approach (Presented on my behalf by Katrin Kania).

Seminars given Department of Classics and Ancient History, University of Durham, 2009. Using experimental archaeology to answer the ‘unanswerable’: A case study using Roman dyeing How exploring differing disciplines’ definitions and ‘Experiment versus Experience’ lead to the reconstruction of the dyeing industry of Pompeii and its scale of manufacture. Department of Archaeology, University of Reading, 2009. Using experimental archaeology to answer the ‘unanswerable’: a case study using Roman Dyeing.

MRS (Materials Research Society) 2010 Fall Meeting, Boston Mass: Paper presented: Reconstructing the past to prevent future loss: the dyeing industry of Pompeii. (Presented on behalf of H. J. Hopkins by Lesley Frame). Interface 2010. The 2nd International Symposium for Humanities and Technology. International Digital Laboratory, University of Warwick, 2010: Paper presented: Establishing the scale and significance of an industry through physical and virtual reconstruction, a case study using Roman dyeing.

13

14

An investigation of the parameters that would influence the scale of the dyeing industry in Pompeii An application of experimental archaeology and computer simulation techniques to investigate the scale of manufacture of the dyeing industry and the factors that influence output.

15

Acknowledgements to the published thesis Thank you to Rob Janaway, Dr Steve Wright, Dr Damian Robinson and Dr William Seale for providing patient and attentive supervision throughout this diverse project. Thank you to Mark Robinson for support in Finite Element Analysis and to Leigh Mulvaney-Johnson for technical support. Thank you to Dr Patricia Andrew and Phil Meise for providing translations for text. Thank you to Prof Andrew Wilson for the use of the photographs in Chapter One. I would like to thank ‘Abrahams and Carlisle’ and the ‘British Wool Board’ for providing materials for the project. Thank you to Thomas Fisher and Thomas Halliwell for providing both technical support and transport throughout the project. Thank you to Liz Parton, Dr Natalie Kay, Dean Summat, Peter Bray, Nick Smith, Rebecca Ramsey, Philip Meise, Philip Corner, Bob Bonwick, Dr Timothy Simmons, Judy Watson, Ruth McBroom, Judith Cauldwell and Rachel Heelas for technical advice, animated discussion, assistance and encouragement regarding the practicalities of the project in a range of areas. Thank you to my parents Dr Richard and Charmian Hopkins for proofreading and encouragement throughout the project. Thank you to Dr Alexander Blustin and Yumiko Okada for proofreading the final version. And thank you to Lorna “Munchie” Willimott (née Watling), involved in the project from day one, for providing support, encouragement, and much animated discussion, who innocently said one day, many years ago “you know, I think they might be able to answer this over in engineering…” (Pers. Comm. Watling, 2002).

16

Timeline of previous work relevant to this study Date

Occurrence

62AD

Earthquake destroyed Pompeii. Pompeii was rebuilt, partly because of interest from the Emperor Nero whose wife Poppaea was from Pompeii.

79AD, August 24th Following previous smaller tremors, Vesuvius erupted. Herculaneum: Exposed to sulphurous gas, buried in a pyroclastic flow. Pompeii: Exposed to sulphurous gas, buried in rain of pumice up to first storey level, remaining city buried in pyroclastic flow. Baiae and surrounding smaller towns: Obliterated. c1750

Discovery of Pompeii and Herculaneum. Plundered as treasure trove. Some restoration of remains occurred at this point.

c1924-1961

As Director of Archaeology, Maiuri excavated and methodically recorded Pompeii. Some restoration of remains occurred at this point.

1976

Publication of ‘wool trade of ancient Pompeii’ by Moeller, W.

1977

Critical review by Wild of ‘wool trade of ancient Pompeii’by Moeller.

1988

Publication of ‘The Economy and Society of Pompeii’ by Jongman.

1994

Preliminary fieldwork undertaken by Janaway and Robinson (unpublished)

1998-9

Laboratory investigation of the dyeing outcome of madder when used with differing mordants, temperatures and pH. Hopkins, unpublished.

2002

Reconstruction and operation of dyeing apparatus undertaken as part of Final Year Archaeological Dissertation, data taken from fieldwork undertaken by Janaway and Robinson, 1994, unpublished.

2002

Survey of dyeing apparatus in Pompeii undertaken as part of this study, based on original list published by Moeller, 1976, with a practical understanding of the dyeing apparatus gained from 2002 reconstruction and operation.

2003

Amendment of dyeing apparatus by addition of a flue

2003

Fieldwork observation undertaken by Robinson to clarify ambiguities arising from 2002 fieldwork, unpublished

2004

Theoretical calculation calibrated by experimental work undertaken in 2002 undertaken as part of Final Year Engineering Dissertation. (Watling, 2004).

2004

Presentation of findings to date at AHRB First Conference in Textile Conservation, Southampton University at Winchester.

2005

Publication of findings to date in conference proceedings.

2005

Rerun of experimental dye runs with flued vat with thermocouples to record temperature at specific locations within the apparatus

17

18

Chapter One

Introduction to the dyeing industry of Pompeii It may be argued that as Herculaneum was encased in solidified lava to a depth greater than Pompeii and so has survived with greater preservation, with less looting or salvage attempts following the eruption, (Allison, 1992), that Herculaneum may be the better city to study. However, Herculaneum is only 1/8 of the size of Pompeii, Pompeii contains a greater number of textile manufacturing apparatus, Pompeii has been excavated to a greater extent than Herculaneum and more intensively studied, (De Franciscis, 1978; WallaceHadrill, 1990).

1.1 The significance of the scale of manufacture of textiles in Pompeii The Roman city of Pompeii is located at the base of the volcano Vesuvius in Italy. Pompeii was destroyed by a volcanic eruption in 79 AD. The city of Pompeii was buried in raining pumice to the first floor level which allowed the preservation of all buildings to that level. Herculaneum and its surroundings were exposed to heated sulphur clouds, killing every inhabitant. Both cities were subject to the Plinian stage of the pyroclastic eruption (named after Pliny the Younger who witnessed the eruption at Pompeii) and were buried in seconds in a lava flow that moved with a force greater than that of a nuclear explosion, (Capasso, 2001). The destruction was sudden and complete, causing the city’s isolation as a whole, encased in solidified lava.

Scholars have debated the scale of manufacture within a Roman city, (Moeller, 1976; Jongman, 1988; Laurence 1994). Understanding the scale of manufacturing would give a usable value to allow the extent and complexity of the economy of Pompeii to be understood. This would allow a greater understanding of the economic functions of the Roman world and the role of Pompeii within this. To determine the size and complexity of an economy it is necessary to select an industry to study in detail. The dyeing industry has the advantage that dyeing apparatus are uniquely identifiable, allowing the equipment to be studied in isolation. Dyeing apparatus has not survived well from Roman sites as a whole. In Pompeii the dyeing industry has left tangible evidence that is identifiable as purely dyeing apparatus. The contrast between the preservation of dye vats in Pompeii and dye vats within the rest of the Roman world as a whole may be seen in Figures 1.1 and 1.2 below. The dyeing apparatus excavated at Sabratha, Libya, (Figure 1.1) have been preserved to an extraordinary degree, yet the preservation of the dyeing apparatus at Pompeii (Figure 1.2) is still superior.

The lava prevented gross taphonomic changes allowing brilliant preservation of structures and artefacts. Where artefacts have decayed their presence may be discerned from remaining outlines, (Allison, 1992). This preserved the city, thereby allowing archaeologists to view complete artefacts that elsewhere have not survived, and to view a city that had ‘stopped-intime’. Most archaeological sites are uninterrupted in development which leads to an obliteration of previous remains through disturbance during the construction of foundations for subsequent buildings. For example, the cities of York, London, Ostia and Rome, contemporary to Pompeii, while having well preserved Roman remains, do not provide the quality of evidence obtainable from Pompeii or Herculaneum. The quality of evidence that survives in Pompeii allows the location and possible heights of walls to be discerned, a demarcation and discernment of the use of buildings, with the exploration of preserved houses and study of equipment from workshops, bakeries and fulleries. Pompeii’s encapsulation and ‘loss’ ensured that such disturbance did not take place, thereby preserving artefacts in situ and each context’s relation to each other. This allowed the relationship of artefacts and context to be understood. Furthermore, the degree of preservation allowed a full exploration of the identification and use of each artefact. Not only were domestic artefacts discovered, but industrial and manufacturing facilities were preserved, allowing their operation to be discerned. This has provided a unique insight into the Roman World, (Laurence, 1994; Wallace-Hadrill, 1990).

At Pompeii preservation is so complete that it is possible to measure and reconstruct the design of each apparatus, in particular the volume of each dye vat and the number and location of the dye vats within each workshop. The reconstruction allows a better understanding of their operation, the amount of cloth produced to be determined and an understanding of the size and significance of the industry as a whole through an understanding of the scale at which manufacturing could have taken place. Once the significance of the industry has been determined it will be possible to gauge its place in the economy of Pompeii, leading to an understanding of the size and functions of the economy as a whole. Through this it shall be possible to gain a better understanding of the Roman economy as a whole.

19

Investigations into the Dyeing Industry in Pompeii

Figure 1.1. Dyeing apparatus at Sabratha, Libya. Measure: 0.5m. Courtesy of Prof. Wilson, 2007, unpublished (University of Oxford)

The wider discussion of the structure and function of the economy of Pompeii is beyond the remit of this study. To understand the structure of the economy itself and to translate this to a modern equivalent is a complex undertaking that has yet to reach a final solution. Instead, this study seeks to understand the scale of manufacturing in one industry which can then be used within the understanding of the economy. The aim of this study is not to determine the intricate workings of the economy as a whole.

Step One: To construct a replica dye vat and use it to determine operating parameters. Step Two: To conduct a review of extant dyeing apparatus in situ in Pompeii, to survey structure, construction and to determine volume. The findings from the replica work shall influence the interpretation of the remains. Step Three: To modify the replica to study the effects that differences in design have on ventilation through the apparatus and its effect on operating parameters.

1.1.1. Aims and Objectives

Step Four: Construct a computer model of a vat and compare it with the two replicas.

The aim of this investigation is to understand the scale of manufacturing of the dyeing industry in Pompeii through the application of experimental archaeology, the principles of thermodynamics and systems theory. These principles will be used to examine the efficiency of the apparatus to estimate the throughput to determine a theoretical maximum capacity of individual dye vats and by extension the capacity of each workshop and the industry as a whole.

Step Five: To apply systems theory to the apparatus and models to show physical limitations in production Step six: To apply systems theory to the Pompeii economy to establish production capacity. Experimental archaeology and thermodynamic principles will be used in conjunction with literary and archaeological evidence, to quantify the possible annual output of dyed

The project objectives can be stated in terms of a number of steps.

20

Introduction to the dyeing industry of Pompeii

1.2 Literature Review 1.2.1 Research background In order to understand the dyeing industry there is a need to understand the operation of the dyeing apparatus and the place of the apparatus within in the wider context of the economy. The role of the apparatus within the workshop and the factors that may increase or decrease the output of the apparatus must also be understood. Each part of the manufacturing process, together within its economic context determines the size and significance of the dyeing industry. To determine what is already understood about the size and significance of the dyeing industry in Pompeii it is necessary to carry out a literature review, a review of previous work. The subject of the size of the dyeing industry and its economic influence has led to vigorous debate in the archaeological literature. There are many stages in textile production, but only dyeing and fulling leave significant tangible evidence of the processes in the archaeological record. A fullonicae is a fullery, where the textile Figure 1.2. Dyeing apparatus five from property VII xiv 17, Pompeii. Note was fulled, receiving its final finish in the intact brazier, firebox and the presence of the original lead kettle. its manufacturing. However a fullonicae Vertical measure: 1.0m, Horizontal measure: 0.2m. was also used to clean garments as well Source: Author, unpublished as to finish them so it is not possible to gauge the quantity of textiles that may be manufactured through a fullonicae, (Jongman, textile in Pompeii, the ‘size and scale’ of the dyeing industry. 1988). However, dye works are also recognisable and Following this, systems theory will be used to gauge the were only used for dyeing materials in the textile quantity of consumables used within the dyeing industry, manufacturing process. Therefore an examination identify the limiting factors of the process and to determine of dye works would result in an accurate estimation the effect these limitations would have had on the total of the size of the dyeing industry, and, consequently, amount of dyed textiles, and through this determine the size, the textile manufacturing industry. Unfortunately, so relative scale and thereby the importance of the industry. far most of the research carried out in Pompeii has been theoretical, based on assumptions applied to 1.1.2. Nature of this investigation the remains of the dye vats in situ. This has resulted in varying answers to the questions of the size of the While the initial question that the project is seeking industry and its significance within the economy. It to answer may be seen to fall within the remit of may be noted that not all of the textiles manufactured archaeology, the project is not a purely archaeological were uniformly dyed, a textile may be plain with detail project. The approach of this project, the questions made from dyed material or the process may not have asked, the methods used and the interpretation of the resulted in a uniform colour, (Bender Jørgensen and answers all use the principles of engineering science Mannering, 2001; Cardon, 2001) but an estimation of and so only through the application of engineering the upward limits of the dyeing industry provides a principles can this problem be fully understood. foundation for a more accurate understanding of the Archaeology alone can not address the research aim textile industry as a whole. and thus far no such answer has been determined.

21

Investigations into the Dyeing Industry in Pompeii It should be noted that the addresses in Pompeii are not given for the properties themselves but refer instead to the doors to each. Therefore it is possible that a single property with doors that face on to more than one street could have one address in one set of records and a wholly different one in a differing set. For example, property VII xiv 17 is actually located between Via dell’Abbondanza and Vicolo degli Scheletri, with doors on to each. The address of VII xiv 17 is the main door leading on to the Via dell’Abbondanza and not the door from the dye works within the building. It is also possible that differing doors on the same street may be listed in differing records, causing a single property to be listed as being at two different address locations. This possibility has been noted and care has been taken to avoid it in this study.

1.2.2 Roman Dye vat design A dye vat is an apparatus used for dyeing fibre, textile or yarn. It consists of a metal kettle, containing the material to be dyed, dye and water. This is supported in a brazier so that the kettle is held above a fire (which provides heat for the reaction to take place). Figures 1.3 and 1.4 show schematic representations of the design of dyeing apparatus. Figure 1.3 shows an unflued vat, Figure 1.4 shows a flued vat. Control of the fire through the firebox is extremely important, as while the dye liquor must simmer, it must not be allowed to boil. Boiling ruins the fleece and causes changes to the dye, possibly resulting in a different colour, (Storey, 1978). The fire (and therefore temperature) is influenced through the fuel amount and the airflow, (Rossotti, 1993). Ultimate control of the temperature relies on a combination of the vat design, the fuel and the skill of the operator.

At no point did Moeller undertake experimental work to determine how the apparatus operated or calculate the theoretical throughput. His findings were estimates based on his conjecture following his application of general principles of dyeing to the remains of the dyeing apparatus in situ. The application was based solely on observing the remains in situ, including the altered remains, and not any practical experimentation into how they were constructed or operated. He concluded that the throughput of the dyeing industry would allow the textile industry of Pompeii to be large enough to export to the hinterland and beyond.

1.2.3 The debate so far The most substantial assessment of the textile industry so far was carried out by Moeller, (1976). He studied the city first hand and identified premises that were used in the various stages of textile production. This included fullonicae and dye works. He used this identification, in combination with a theoretical knowledge of the dyeing process, to determine the size and scale of the dyeing industry. Below is the list of properties that Moeller identified as dye works (Moeller, 1976:35-39). An explanation of the coding of addresses within Pompeii may be found in Appendix One.

Moeller’s work was heavily criticised in subsequent studies, including the original review of his findings by Wild who stated that ‘Moeller regularly goes beyond the point to which his sources can take him’, (Wild, 1977). Wild believed that there was insufficient evidence presented in Moeller’s work to allow the conclusions about the size of the dyeing industry that Moeller had made, (Wild, 1977). Moeller had concluded that the textile industry was large enough to supply Pompeii and also to export. Jongman concluded, using the same

I viii 19 Vi 4-5 VII ii 11 VII xiv 17 IX iii 1-2 IX vii 2

Figure 1.3. The parts of a dyeing apparatus. Source: Author

Figure 1.4. The parts of a dyeing apparatus with flue. Source: Author

22

Introduction to the dyeing industry of Pompeii

evidence as Moeller, that the textile industry was small and that it could only supply the population of Pompeii, possibly requiring supplementation with imports, (Jongman, 1988). Mann (1994:87-106) questioned Jongman’s statistical methods and discovered that his evidence should result in a larger industry than even Moeller claimed. Mann further concluded that the true scale and importance of the dyeing industry may be impossible to determine.

Moeller, although criticised, did collect the data that he used first-hand, (Mann, 1994; Robinson, 1999). Robinson examined the use of space in Pompeii and the techniques for recognition of different industrial areas, (Robinson, 1999). Mann carried out a literary review, (Mann, 1994). The amount of dyed material that may be produced by a single dye works has been calculated purely from examination of the vats in situ, and the application of recipes and the presumed effects of factors identified through literature review as being influential to the dyeing process. This, understandably, has led to different interpretations of the size and significance of the dyeing industry.

The spatial distribution of the textile producing workshops and the significance of their locations has also been examined. Robinson re-examined the identification of the textile workshops of Pompeii, challenging many of the findings of Moeller, (Robinson, 1999). This re-assessment agreed with Jongman and Mann that the textile industry was smaller than Moeller believed it to be, and may have only been operating on a scale large enough to supply Pompeii.

It should be noted that although each subsequent author has been critical of Moeller, (Jongman, 1988; Laurence, 1994; Mann, 1994), Moeller’s identification of the dyeing apparatus has not been questioned, even by authors who have examined the dye works in situ. It is only Moeller’s theoretical assumptions and economic conclusions that have been questioned. Moeller’s identification of the standing remains has withheld scrutiny from each subsequent author, although no examination of the standing remains in situ has been undertaken as thoroughly as Moeller’s original survey (Moeller, 1976).

Laurence, (1994), examined production in Pompeii from an economic standpoint and within the social context, rather than through the structural remains, and so did not undertake an examination of the physical practicalities involved. Laurence followed Jongman’s reasoning in deconstructing Moeller’s argument by exposing a lack of substance in Moeller’s ideology. However, Laurence felt that Jongman applied Finley’s ‘Consumer City’ model to Pompeii purely to challenge and criticise others, notably Moeller, rather than as an attempt to examine Pompeii, (Finley, 1973; Jongman, 1988; Laurence, 1994). Laurence concluded that despite problems in its application, a ‘consumer city’ model was the most appropriate for an ancient city in economic terms, such as Pompeii, (Laurence, 1994). While this may be of interest in placing the industrial output of Pompeii into an economic context, the exploration of the implications of Laurence’s argument falls outside the remit of this study. While Laurence concluded that Pompeii possessed a relatively small textile industry and could not have supported export, he did this through economic argument and without examination of the physical factors involved, (Laurence, 1994).

Previous work: A Strengths and Weakness Analysis A breakdown of the strengths and weaknesses of the conclusions of each previous study on the subject of the size and significance of the dyeing industry in Pompeii is given below. Moeller Moeller surveyed the standing remains of the structures associated with the textiles industry as a whole in 1976. Following his survey he concluded that the dyeing industry was large enough to supply the textile industry with an excess to export to the surrounding hinterland and beyond. It should be noted that although the criticism may be levelled at subsequent authors that they relied on data collected by Moeller, Moeller’s survey was comprehensive.

Whilst it must be noted that previous research has been detailed and broad, it relied mainly on data collected in the field by others and on reviews of previous literature.

Table 1.1. The Strengths and Weaknesses of Moeller’s findings Strengths

Weaknesses

• Moeller collected his own data. • It is the largest survey to date. • Moeller provided a conclusion to his work.

• Moeller’s conclusions go beyond the reach of the evidence (Wild, 1977). • Moeller did not undertake experimental work. His conclusions are theoretical. • It is unlikely that the industry was as large as Moeller’s conclusion suggested

23

Investigations into the Dyeing Industry in Pompeii Table 1.2. The Strengths and Weaknesses of Jongman’s findings Strengths

Weaknesses

• Reassessed Moeller’s data. • Concluded that the dyeing industry was small, supplying the domestic market.

• Used Moeller’s data without collecting his own. • Didn’t undertake experimental work – made assumptions over operation • Statistical methods questioned by Mann (1994)

Table 1.3. The Strengths and Weaknesses of Mann’s findings Strengths

Weaknesses

• Reassessed Moeller and Jongman’s findings. • Used literature review and pictorial sources to assess textile requirements of Pompeii. • Used additional methods to analyse previous work (statistics) • Stated that it is not possible to determine the size of the dyeing industry.

• Used Moeller’s data without collecting her own. • Undertook no experimental work. Made theoretical assumptions about operation of apparatus.

Table 1.4. The Strengths and Weaknesses of Laurence’s findings Strengths

Weaknesses

• Reassessed Moeller and Jongman, agrees with the findings of Jongman. • Familiar with Pompeii and its economic context

Jongman

• Didn’t specifically collect data but used Moeller’s. • No experimental work was undertaken, theoretical assumptions were made regarding throughput

previous findings, for example her use of statistics. Mann stated that it would never be possible to determine the size of the dyeing industry.

Jongman used Moeller’s data and theoretical assumptions to determine the size and significance of the dyeing industry of Pompeii. He concluded that Moeller had miscalculated the size of the dyeing industry and that it was in fact smaller than Moeller had stated, only allowing for domestic supply and possibly relying on imports (Jongman, 1988).

Laurence Laurence reassessed the findings of Moeller and Jongman, (Laurence, 1994). While he did not collect his own specific data concerning the dyeing industry of Pompeii, but instead relied on Moeller’s original survey, he was familiar with Pompeii and the economic context of the textile industry in which Moeller’s findings were placed. He used Moeller’s data, but questioned Moeller’s findings, and agrees with the conclusions of Jongman. Laurence did not undertake any experimental work, so his conclusions regarding the throughput of the dyeing industry and the operation of a dyeing apparatus were entirely theoretical.

Mann Mann questioned both the findings of Moeller (1976) and Jongman (1988). She re-examined Moeller’s finding and assessed whether Moeller had overestimated the amount of textile processing equipment identified and re-examined the statistical methods used by Jongman, (Mann, 1994). She concluded that Moeller had over-estimated the size of the dyeing industry. However she also concluded that Jongman had incorrectly used his statistical methods and that in fact if his arguments were taken to their full conclusion he would have concluded that the dyeing industry was even larger than Moeller argued. Mann also used pictorial sources and literature review in assessing the quantities of textile that would be required in Pompeii and put this in context with the arguments of Moeller and Jongman. She assessed both the textile production and requirement of Pompeii. She used additional methods to assess the validity of

Summary While Moeller’s findings and conclusions have been criticised, his actual survey of Pompeii was the most comprehensive undertaken to date. However, his survey was extremely broad and relied on a theoretical knowledge of the dyeing process to identify the remains that were of interest to his study. Moeller stated the geographical location of each of the dyeing apparatus within the dye works and the location of the dye works

24

Introduction to the dyeing industry of Pompeii

within the city, but did not examine these relationships further. He did not expand the context with a practical understanding of the needs for delivery and storage of resources such as fuel and raw materials or the space required for processing.

It may even be possible to draw comparison between the approach of Borgard’s report and the view of Moeller’s findings (Moeller, 1976) formed during the original peer review of Moeller’s publication, (Wild, 1977, reviewing Moeller, 1976).

1.2.4 Quantifying archaeological writing

The aim of the study, which was achieved, was to investigate the architectural evolution of properties associated with dyeing. However, although the title suggested an investigation into production and equipment, no exploration of the function or measurements of equipment was published. Conclusions regarding production were generalised and presented without evidence to support them. There was no attempt to place findings within a wider economic or social context or to determine output or the context of the workshops within the industry.

An unfortunate element of archaeological reporting is the tendency to report the author’s findings without the reasoning behind it. For example, Jongman states that there are six dye works in Pompeii containing a total of 32 dyeing apparatus, (Jongman, 1988). He does not cite the source for this data, but the context is in a review of the work of Moeller, (Moeller, 1976). He does not state whether this is his finding or Moeller’s. The importance of accurate reporting may be seen in this example as Jongman is incorrect, there are 40 dyeing apparatus in Pompeii (confirmed through fieldwork by Janaway and Robinson, 1994, unpublished).1 This tendency makes quantifying and proving archaeological data difficult and shows the need for returning to the original remains to allow an accurate survey on which to build further findings. Through reporting, re-reporting and the addition of further findings it is possible to lead to discrepancies. The foundation of a new study of the dyeing industry should be an independent survey of the standing remains, allowing inaccuracies and discrepancies in past reporting to be discarded.

The approach and subject of the study appears superficially relevant to the present investigation, but on closer examination may only make an indirect contribution. The architectural history of the dye workshops may be relevant, but the incomplete data and sweeping generalisations mean that the study can not contribute greatly to a wider economic understanding. Watling, 2004 Watling undertook a final year dissertation in 2004 which investigated the processes behind a dyeing apparatus, (Watling, 2004). She used the Specific Heat Capacity of each constituent part to determine the theoretical energy required to heat an apparatus during a dye run. The results were calibrated with practical experimental work and used to predict the energy required to heat other dyeing apparatus of differing design. The results were presented as a spreadsheet allowing investigation of the effects of a change in each value. Watling presumed that no physical change occurred to the materials during heating. The study was of how to approach a problem rather than the actual determination of a solution, using the dyeing apparatus as a case study. Watling undertook a theoretical study which was validated through experiment. This was undertaken as part of the wider question that this project addresses. Watling’s approach and findings are discussed in more detail in Chapter Three.

Related work Recent studies have been undertaken, and although each bears a similar title to this thesis, the aim, approach and findings have been entirely different. The most recent was undertaken by the Centre Jean Berard and reviewed the methods of production in Pompeii, (Borgard et al, 2003). The Centre Jean Bérard undertook excavation of properties associated with production, in particular oil, perfume, tanning and textile production, from 2000 to 2002, (Borgard et al, 2003). They state that the textile workshops investigated were those identified by Moeller (1976), and state them to be I viii 19, Vi4, Vi5, VII xiv 17, IX vii 2 and ‘The House of the Queen of England’. No other properties are mentioned as having been identified by Moeller, which suggests that property VII ii 11 was not investigated. Furthermore, Moeller did not list IX vii 2 or ‘The House of the Queen of England’ as dyeing properties. Other than Borgard’s report, there is no other record of any property referred to as ‘The House of the Queen of England’ in Pompeii. It is unfortunate that no map was included in the publication of this work, thus it is not possible to discern which property is referred to.

Watling’s study highlighted the difference in approach between engineering and archaeology from an engineering point of view. She stated that as engineers continued to study an artefact once the method of physical operation (how a person physically stood there and operated it) had been determined, they were able to deduce missing parts and predict further behaviour due to an understanding of the physical processes within the artefact. It was noted that some

The survey later in this work, undertaken in 2002 by R. C. Janaway and the author after construction and use of the first replica, confirmed that there were only 35 vats, not 40.

1 

25

Investigations into the Dyeing Industry in Pompeii archaeologists do endeavour to understand the internal operation of artefacts, but due to a lack of understanding of the principles involved the recording and processing of data may be misleading, sketchy or inaccurate. Furthermore these archaeologists are still in the minority. Coles and Reynolds explored the physical processes and limitations of artefacts from an archaeological point of view through the introduction of the science of experiment to replica reconstruction, (Coles, 1973; Reynolds, 1999), but even they stopped before examining the internal physical processes within the artefact.

determine the operating parameters of the apparatus and critically re-appraised the remains in situ. This allowed further experimental work to determine limitations within the apparatus and the scale of manufacture possible using the apparatus. To determine the scale of manufacturing there needs to be an approach from a differing perspective. There needs to be a merger of archaeological method, to ensure that the work remains in context, and a further investigation using the principles of modern engineering. The conclusions so far on the size and significance of the dyeing industry in Pompeii have been based mainly on theories drawn from studying the remains in situ and theories based on previous literary work and the Roman literature (for example, Pliny’s Natural History). These have been drawn from the structure and size of the vat itself with assumptions toward recipes based on later pre-industrial dye works and the writings of contemporary authors who lived within the Roman world, (Rosetti, 1548; Lagercrantz, 1913, Grierson, 1986; Grieve, 1992). Whilst these provide an insight into the workings of a Roman dye works, they may not be used alone to gauge the size and importance of the industry itself both within Pompeii and as an export industry. This may be demonstrated by the inconsistencies in the perceived importance of the industry.

1.3 The significance of this study Triangulation is the application of varying different approaches, usually from different fields of disciplines, to a problem in order to solve or to better understand it. If a theory withstands scrutiny from one discipline this does not mean that it is valid. If a theory withstands scrutiny from a range of disciplines this suggests that the theory is of greater accuracy. The concept of triangulation is of relevance as this study provides a practical demonstration of the method of application with the benefits and process of triangulation. The combination of investigative methodologies from both archaeology and engineering is a novel approach. Such an approach has been suggested by Wylie, (1989, cited by Bray, 2004), explained using the analogy of a cable: Each strand of independently derived evidence when bound together is mutually supportive and takes a greater weight than a single strand. Wylie was certainly not the first to suggest a combination of methods to test a theory, (Bray, 2004). Suggestions had been made to use alternative methods to critically reassess archaeological conclusions prior to this, such as Clarke’s 1973 call to arms, (Clarke, 1973 cited in Bray, 2004).

1.3.1 To conclude The aim of this investigation is to understand the scale of manufacturing of the dyeing industry in Pompeii through the application of experimental archaeology, the principles of thermodynamics and systems theory. As a combination of the examination of the remains in situ and literature review have not allowed a conclusive understanding of the industry a new approach is required. To determine the scale at which dyed textile may be produced by a dye works it is necessary to reconstruct the relevant part of the dye works and to run an experiment to see how much dyed textile may be produced within the time allowed. The results of this, and the understanding of how the dyeing apparatus operates should be then examined in context, the factors that influence the operation shall be determined and their affects understood. So far no such experiment has been undertaken. The closest reconstruction of a vat of this design was the reconstruction of vats used for cooking aboard the Mary Rose. However, the Mary Rose Trust used their reconstruction for experimental cookery, (Mary Rose Trust Website, 15/11/01). No one has yet reconstructed a vat of a similar size and design for the use of dyeing.

The questions of the scale of manufacturing of the dyeing industry and the operation of the dyeing industry within the context of Pompeii have yet to be conclusively answered. The answers that exist already are varied and highly contentious in the archaeological community. It may even be argued that the approach to each answer is in fact more contentious than the answer itself. The archaeological methods used so far have been insufficient to answer the question. As the approach has been purely archaeological, although varied within the single discipline, the examination of the dyeing industry has not progressed. The approach of this study is different. This study used experimental replicas to

26

Introduction to the dyeing industry of Pompeii

and alteration of each of the apparatus and recording the remains in context.

1.4 Thesis Outline: Chapter One introduces the problem under investigation: the scale of manufacture of the dyeing industry and why there is debate to the industry’s size and significance. It begins to examine the physical structure of dyeing apparatus and the context of each.

Chapter Five contains an ergonomic assessment of the dyeing apparatus, following the matching of Roman skeletal data to a modern population and data set. Chapter Six contains further practical experiments to determine the affects of changes to the design of the apparatus. There is also an application of modern systems theory to the dyeing industry of Pompeii as a whole.

Chapter Two of the thesis reviews the literature that is available, both contemporary and modern, to discern what is already understood of the dyeing industry and the influences of factors affecting the process.

Chapter Seven introduces the application of computer simulation to understand the physical changes in the apparatus through creep arising from changes in temperature and load during the dyeing cycle.

Chapter Three explores the dyeing industry through experimental work, the findings of which shape the investigation into the ergonomic and economic considerations of operating the dyeing apparatus.

Chapter Eight contains a discussion of the results from each part. Conclusions and suggestions for further work are presented.

Chapter Four contains a critical appraisal and survey of the remains in situ in Pompeii, determining the usability

27

Chapter Two

Literature Review 2.1 Introduction When examining the dyeing process it is necessary to place it in the context of textile manufacturing and to understand the factors that affect it. Textile processing is a long and complicated process. It is necessary to determine the material used in the manufacture of textile and the dyes used by the Romans before the process can be fully understood. There are further factors that affect the dyeing process: the apparatus, the inputs to the process that become part of the finished piece and the consumables used that allow the process to take place. These must be understood before a complete picture of the manufacturing process can be constructed. Once the manufacturing process has been understood it is necessary to understand the amount of textile that would have been required. For this it is necessary to not just determine the output but also the number of people that would be using this textile, thereby how far the amount would have met the requirement of the population. Only once the quantity that it is possible to produce, and the requirement of the populace are understood, is it possible to discern the significance of the dyeing industry. 2.2 Textile processing The Roman textile manufacturing process is both long and complicated, as outlined in Figure 2.1. This study concentrates on the role of the dyer in textile production. Due to the nature of the industry, the preceding factors and final product need to be taken into consideration to not only provide a fuller picture of the importance of dyeing, but also its contribution to the process overall. Dyeing apparatus at Pompeii may be identified without ambiguity; therefore this thesis focuses on dyeing. Workshops containing only dyeing apparatus were used only for dyeing. Other premises may have more than one use, for example fullonicae can be used both for fulling new clothing and washing clothing that has already been worn, (Jongman, 1988).

Figure 2.1. The stages of textile manufacture. After Moeller, 1976 and Frayn, 1984

in a dissolved mordant. The mordant, usually a metal salt, then chemically binds to the fabric and upon dyeing also chemically binds with the dye. This results in a fast colour. Without a mordant the dye would leach from the fabric, (Rosetti, 1548; Brunello, 1973; Storey, 1978; Ponting, 1980; Dalby, 1985; Grierson, 1986; Grieve, 1992). To use a mordant dye without a mordant or to use a dye incorrectly would result in a colour that is unfast, that is it would leach through exposure to sunlight or water. The exposure to the wrong temperature or pH

2.3 Dye types Dyes can be classified into two main types: vat dyes and mordant dyes. Vat dyes, such as indigo, do not require a mordant. When a vat dye is applied to a fabric it sticks and remains fast (it doesn’t leach or bleach in water or sunlight). A mordant dye is a dye that requires the fabric to be pre-mordanted, that is treated by soaking 28

Literature Review

would cause a difference in colour, or ruination of the dye or fleece. Dyeing uniformly requires great skill as small changes upset the process and the Roman dyers dyed successfully without modern instrumentation.

limited, (Croom, 2000). It is very unlikely that textiles could be found in Italy, and so Italian textiles have to be determined from other sources, both in terms of the medium of the source and the geographical location of its origin. Assemblages of contemporary textiles discovered in geographically widely located parts of the empire may be used to build an overlapping picture of what technology and materials were available that may be used to determine the textiles likely to be manufactured in the areas for which direct evidence is missing.

2.4 When to dye: Stage at which dyeing takes place The stage of manufacture at which dyeing actually took place is subject to debate. Theoretically this could be as fleece, as yarn, or as fabric. Examination of the warp and weft of surviving Roman textiles show whether a textile was dyed in the cloth rather than dyed in the wool or fleece. If raw wool is present at a site it suggests that the cloth could have been manufactured on site, (Ryder and Gabra-Sanders, 1992). According to the findings from Masada, the cloth had definitely been dyed before it was woven, (Sheffer and Granger-Taylor, 1994). According to Pliny, (cited in Frayn, 1984:142-161) the Romans dyed the fleece before it was spun into yarn. Frayn is in agreement with this text, (Frayn, 1984:142161). Wild is cited as stating that the wool was dyed as fleece, (Wild, cited by Moeller, 1976). For the purposes of this thesis it shall be presumed that the wool was dyed as fleece and that the above figure (Figure 2.1) is accurate.

Two significant assemblages of Roman textiles are those from Masada in Israel (an arid site) and Vindolanda on Hadrian’s Wall (a waterlogged site). These finds are contemporary to Pompeii, and although it may be argued that Vindolanda may not climatically match Pompeii, these areas may be used to examine the technology and styles employed in textile production. The group of textiles from Masada are the largest group of Roman textiles found to date and their preservation state allows the discernment of dyes and materials used in their construction, (Sheffer and Granger-Taylor, 1994). The textiles from Vindolanda form a large group of Roman textile contemporaneous to Pompeii and reflect the technology available in Pompeii, (Taylor, 1983). The difference in preservation techniques allows an overlap of surviving material types.

2.5 Roman textiles The relative expense in the manufacture of textiles (Croom, 2000), led to their reuse ‘until just rags remained, and even these could be utilised to stuff cushions and mattresses,’ (Ellis, 2001). This meant that textile remains when discarded were fragile and so already likely to decay in the archaeological record. Textiles do not survive well in the archaeological record due to their organic nature, (Watkinson and Neal, 1998:65; Harris 1999:8; Baginski, 2001), and textiles that are already damaged are even less likely to survive.

The textiles discovered at Masada (Israel) are of specific importance to this study for two reasons. Firstly, they are the largest group of Roman textiles ever to be discovered. Secondly they have been preserved to an extraordinary degree: the original dyes are still present and the colours may be observed, (Sheffer and GrangerTaylor, 1994). This means that they may be examined to see the type of textile that was most commonly used and the most common dyestuff. Of the 2000 textile fragments discovered at Masada, 122 were analysed. There were three criteria for selection: the piece either had features allowing exploration of the method of manufacture, had unusual features that were of interest or were a typical example of that item. The findings are shown in Table 2.1.

Unfortunately, it is not possible to directly study textiles excavated from Pompeii, or even Roman Italy. This is due to the temperate climate which causes the degradation of organic remains. It is therefore necessary to examine what has survived and where and to build an analogous overlap that covers Pompeii. The textiles that have survived have done so as their decay was arrested. This means that they have been subject to either a hot and arid, desiccated, climate, where the organisms that would otherwise destroy them cannot survive, (Janaway, 1996), or an anaerobic waterlogged environment, where the organisms that would otherwise destroy them have insufficient oxygen. Croom states that textiles may only survive in extremely wet or dry conditions, and that as these exist in Egypt, in a small number in Britain and France, and a few in Denmark in areas outside the Roman Empire, the chance of finding Roman textiles is extremely

From this alone it may be seen that if this were a true representation wool appears to have been the most commonly used material. It may also be noted that wool was used in the greatest variety of textiles, from fine clothes to sacking cloth, (Sheffer and Granger-Taylor, 1994; Frayn, 1984:142-161). Similar studies have also demonstrated the frequent use of wool. For example, of the (approximately) 50,000 textile fragments discovered by Bender Jorgenson and Mannering at Mons Claudianus, of which random sampling was undertaken, the 200 tunics and 45 mantles studied had been manufactured from wool, although linen tunics 29

Investigations into the Dyeing Industry in Pompeii Table 2.1. The Masada textiles. Analysis of 122 textiles, sampled from original 2000 fragments. (After Sheffer and Granger-Taylor, 1994). Number Analysed

2000 textile, of which 122 analysed

All textiles by material

106: Wool 11: Goat 1: Goat/camel 1: Cotton (modern, discounted)

Dyed textiles by material

43: Wool 1: Linen

Colours

Pink, orange, purple, brown, blue, blue-purple, purple, mauve, grey, black

Dyes

Red (madder), blue (indigotin, dyestuff unknown), yellow (unknown)

Mordant

Iron

were known of at this time, (Bender Jorgenson and Mannering, 2001). It has been noted that early Roman textiles contemporary to Pompeii are rare (Bender Jorgenson and Mannering, 2001), so the Masada textiles and the Mons Claudianus textiles are important when examining the materials and manufacturing technology that was available. From the examination of the material used in the manufacture of textiles it is therefore possible to conclude that the most widespread material was sheep’s wool. While it may be argued that wool finds from Vindolanda indicate a use of wool through a climatic need, the wool findings from the desert sites of Masada and Mons Claudianus suggest that wool was geographically widely used in the Roman civilisation.

2.6 Requirements for dyeing There are many technical factors to consider when examining the dyeing process. These include: • The apparatus • The inputs: the fleece, dye, mordant • The consumables that allow the process: water, fuel A failure or inadequacy in any of these factors will limit the outputs of the dyeing process. For example, if there is a supply of fleece and dyestuff but an insufficient supply of water then dyeing can not take place. In addition there are outputs:

Of the dyes that have been found, only madder could be identified from the dye back to the plant, (Koren, 1994). While the presence of indigotin may indicate the use of either woad or indigo, the specific dyeplant may not be determined from this alone. Madder may be identified from the chemicals found within the textiles, (Pers. Comm. Edmonds, 1999). Madder was found present in ‘every’ shade of red, (Sheffer and Granger-Taylor, 1994). It may therefore be assumed that madder was readily available, commonly used and that the dyers of the Roman world were well acquainted with its versatility. Further evidence for the knowledge of madder was the relative purpurin and alizarin contents of the dye. Both of the dyes occur naturally in madder, but in different quantities depending on its treatment and age. The redder shades had a higher content of alizarin and purpler shade contained more purpurin. While the greater amount of purpurin may be due to the use of a different sub-species, it could also have resulted from use of an older plant, or one that has been stored for a greater length of time, (Koren, 1994). This shows a familiarity with the madder plant that along with its shear volume of use demonstrates that it is suitable for this study. According to Moeller, dyers tended to specialise in only one groups of colours, (Moeller, 1976:4-28). Dyers had a detailed knowledge of each plant dye.

• • • •

The dyed fibre, yarn or fabric The used dye liquor (this requires disposal) The exhaust gases (these require disposal) The heat (this will impact the surrounding apparatus and dyers)

Furthermore, each input and consumable must be transported and stored prior to use. Each finished fleece must also be transported and stored. Supply of any of the consumables used in the dyeing process would have limited the output of the apparatus. The consumables include those used in dyeing, such as the material, dye and mordant, and those used in the dyeing process, such as the fuel. The source of each of these, the quality of each, their transport costs and storage would have affected the use and supply of each. 2.7 Consumables used in manufacture 2.7.1 Fleece As has already been shown, the principal material used for textile manufacture in the Roman period was wool. It is therefore necessary to understand the source and quality of the wool, as differing qualities were suitable for 30

Literature Review

different uses and took up dye differently, (Ryder, 1990). Sheep breeding to produce white fleeces and a wool coat began in the Roman period, (Ryder and Stephenson, 1968; Robinson, 1969; Ryder, 1990). By 130AD there were both fine and medium fleeces in the Middle East, (Ryder and Stephenson, 1968). Prior to this the differences in wool in the sheep’s coat had been utilised to create patterns in the textile, (Ryder 1990). The majority of fabric was a natural white or brown, with a small amount of patterning from naturally pigmented or dyed wool. The pigmented wool was a colour, not a background. Breeding for a white fleece developed as there was a need for an increase in the amount of the flock that could be uniformly and brightly dyed, (Ryder 1990). Prior to this the majority of sheep were brown, (Ryder and Stephenson, 1968). Until the Romans reached Britain in 43AD the only sheep in the British Isles were Soays, a factor that should be considered when examining the finds of Vindolanda, (Ryder and Stephenson, 1968). The closest fleece today to a Roman fleece is that of the Shetland, with a fleece of approximately 2kg. The closest sheep today to a British Iron Age sheep is the Soay, (Ryder 1990).

to transport, and dye would have adhered more readily, but wool is easier to spin with the grease still present, (Frayn, 1984:142-161). This may be shown in the cost of the wool in Pompeii: washed wool was cheaper than unwashed and of thirteen accounts entries, only six were washed, (Frayn, 1984:142-161). Any washing of the wool took place before it reached the markets, (Frayn, 1984:142-161). Dyeing could sometimes take place at the farms according to Pliny the Elder, (Frayn, 1984:142-161). Through studies of archaeological assemblages such as the textiles from Masada and Vindolanda it was discovered that wool was the most common material, (Taylor, 1983; Taylor, 1987; Sheffer and Granger-Taylor, 1994). Wool was used in the Roman world for textiles ranging in use and quality from fine garments to sacking, (Sheffer and Granger-Taylor, 1994). This versatility made wool more popular than linen. Pliny states that wool was preferred to linen, as it was softer, (Pliny, Natural History XIX ii). This range in use is due to the qualities of wool. As sheep’s fleece is a mixture of wool and kemp (hair) it is possible to have differing grades of softness and strength. The qualities of the fleece are inherited through continuous variation so is difficult to breed for, (Ryder 1990).

The source of wool would affect the output of the textile industry. Wool would have been imported into Pompeii, but there is a debate as to the actual source. Seneca (cited by Moeller, 1976; Mann, 1994) states that 600 sheep were killed during the earthquake of 62AD in the area around Pompeii. This allows the presumption that although these sheep weren’t kept necessarily for wool; it was possible that wool was gained locally from Pompeii’s hinterland. Frayn points out however that the coastal plains surrounding Pompeii were fertile and so would have been in agricultural demand, (Frayn, 1984:142-161). Moeller also states that the mountainous regions to the south and west of Pompeii could have been used for grazing sheep, (Moeller 1976, cited by Mann). However, to the west of Pompeii is the sea, (Frayn, 1984: Figures 2, 3, and 6), and so this theory does not appear to stand up to close scrutiny. There are another two possible sources for the wool processed in Pompeii. The regions of Samnium and Apulia were possible wool sources, (Frayn, 1984:142-161; Moeller, 1976:4-28; Mann, 1994). However the import from these areas would depend on a large enough industry to warrant the expense, something that Moeller believes existed and that Mann believes did not. Frayn also explains that inter-regional wool trading did not in fact take place, (Frayn, 1984:142-161).

It may be concluded that wool was readily available, but had to be brought into the city and stored. The question remains as to whether this store was outside the city and how often deliveries were made. 2.7.2 Mordants The ‘Papyrus Graecus Holmiensis’, (Lagercrantz, 1913, originally in Greek and translated to German by Lagercrantz) describes dyeing and dye recipes contemporary with the Pompeii era under study. The first part of the work is a description by Lagercrantz of how he happened across the works and the state of the papyrus that they were written upon. The second part is a description of dyeing and dye recipes contemporary with the Pompeii era under study. These include a description of mordants and two recipes resulting in red cloth. Vitriol (believed to be zinc sulphate) and skorpiurus (a leguminous plant) are listed as mordanting agents. In the recipes alum is listed as the mordant of choice. Vitriol is a sulphate of any metal and is so called because of the glassy appearance or lustre. Lagercrantz also states that during mordanting the wool must be left to soak overnight (Lagercrantz, 1913) – further evidence that this was common practice because of the times involved.

The wool would have reached Pompeii in numerous states. Wool was tied as individual fleeces with a cord for transport, (Frayn, 1984:142-161). It is unlikely that the fleece would have been transported any distance unwashed due to the weight. An edict by Diocletian refers to fleece as being washed, (Diocletion, AD 301, in Lewis and Reinhold, 1966). It may have appeared that wool would have been more desirable when washed, but this was not necessarily the case. It was certainly lighter

The only contemporary collections of dyeing recipes were from Egypt. These were the ‘Papyrus Leidensis X’ and the ‘Papyrus Graecus Holmiensis’, (Vogler, 1982) which either Monaghan was unaware of or misreported, as he stated that there were no Greek writings on dyes, (Monaghan, 2001). It is possible that these were written 31

Investigations into the Dyeing Industry in Pompeii in Egypt but were unknown in Greece. These both date from the third century AD, but the recipes are much older. Pliny also reports ancient Egyptian dyeing methods, (Vogler, 1982).1

dyes, red and yellow respectively, woad is a blue vat dye that requires fermentation before use. Madder contains two dyes, purpurin, a purple dye, and alizarin, a red dye, (Koren, 1994). Both of the dyes occur naturally in madder, but in different quantities depending on its treatment and age. A greater amount of purpurin may be due to the use of a different sub-species, an older plant, or one that has been stored for a greater length of time, (Koren, 1994). Prolonged heating and a difference in pH may also cause a differing colour.

It should be noted that there may be inaccuracies in the published pre-industrial recipes. This is because the author and publisher was reliant on breaking the dyer’s trade secrets. Those that were writing the recipes, such as Pliny c79AD or Rosetti in 1548, had usually received a classical education and were detached from the dyeing process, so were not themselves familiar with it.

There is confusion over the use of woad and indigo in the Roman World. The chemical indigotin is the dye from each plant so when it is found it is not possible to discern which dyestuff the dye was from. Indigo originates in India, woad originates in Western Europe. It is therefore assumed that indigotin discovered in Western Europe showed the presence of dyeing with woad, but it is not possible to be absolutely sure, (Koren, 1994; Pers. Comm. Edmonds, 1999).

The use of madder as a colourfast dye requires the use of a mordant, (Ponting, 1980, Greirson, 1986). Ordinarily that is alum. This has been found in the Masada textiles and was in widespread use throughout the Roman world, (Greirson, 1986; Sheffer and Granger-Taylor, 1994; Pers. Comm. Edmonds, 1999). There is evidence of mordanting with alum at Vindolanda, (Taylor, 1983). Most modern dye recipes that use madder use alum as a mordant, (Brunello, 1973; Grierson, 1986; Grieve, 1992). Until the nineteenth century alum’s largest application was in the textile industry as a mordant, (Miller, 2002). Alum is also used in tanning to produce leather that bends and stretches, (Miller, 2002). This means that alum could have had two uses in Pompeii, which raises questions regarding its procurement and any sharing or competition between the dyeing and tanning industries. Alum can form as a natural mineral under volcanic conditions. It solidifies from solution, producing large crystals of characteristic shape, a property that gave alum its commercial value for many centuries (Almond 1975: 11, cited in Miller, 2002). In southern Italy and the Middle East, alum occurs naturally as alunite in ancient volcanic craters and is often associated with hot springs, (Singer 1948: 17, cited in Miller, 2002). This assertion gives rise to the question of whether alum could have occurred naturally locally to Pompeii. However, alum could also have been imported to Pompeii.

Murex is a purple dye obtained from shellfish and was extremely expensive, leaving most dyers to rely on the plant based dyes from which a full colour spectrum may be obtained. Murex is unstable and oxidises from purple to dark blue. Murex only differs from indigotin by the addition of bromine: Murex is di-bromo-indigotin and oxidises to indigotin, with the loss of the bromine molecules. This means that not only was it expensive, but it was unstable and would need re-dyeing and replacement. The instability of dyes causes loss in the archaeological record. ‘Any dyed colour usually disappears from the buried textile. At best the dye is usually seen as a diffuse grey coloration; undyed wool usually has a yellow or brown coloration.’ (Ryder and Gabra-Sanders, 1992). If dye survives it is usually blue (Woolley, 2001) or red. This is because indigotin and alizarin are relatively stable and so may survive in the burial environment if the textile survives. In Northern European sites blue dye is usually thought to be woad and red dye is usually thought to be madder. However, it isn’t possible to identify the dye properly without chemical analysis, (Ryder and Gabra-Sanders, 1992). Of the dyes that may be found, only madder can be identified from the dye back to the plant, Rubia tinctorum, (Koren, 1994). Madder is the most common dye found in the plant record, (Walton-Rogers, 1997). While the presence of indigotin may indicate the use of either woad or indigo, or even degraded murex, the specific dye source may not be determined from this alone. As suggested by the results of Koren (1994), yellow dyes are unstable and so are rarely discovered in the archaeological record. Furthermore, as stated by Pliny, yellow was not likely to be discovered in Roman textiles as the colour was exclusively reserved for bridal gowns, (Pliny, Nat Hist, XXI).

2.7.3 Roman dyes The Roman dyers used madder, woad (possibly indigo), weld and murex to dye cloth or fibres. Madder, woad and weld are plant dyes: madder and weld are mordant The original thesis states “The only contemporary collections of dyeing recipes were from Egypt. These were the ‘Papyrus Leidensis X’ Investigations into the Dyeing Industry in Pompeii and the ‘Papyrus Graecus Holmiensis’, (Vogler, 1982) which either Monaghan was unaware of or misreported, as he stated that there were no Greek writings on dyes, (Monaghan, 2001).” A fuller re-reading of Monaghan’s work notes that Monaghan knew of these sources, but cited BalfourPaul (1998:124) and Forbes (1956) as evidence that they were irrelevant as they weren’t Greek. This study instead cites Lagercrantz (1913) who said they were Greek, from Greece and translated them into modern German. This is explored in other published work (e.g. Hopkins 2018) as an example of how research before the internet may be limited by resources and translations available.

1 

32

Literature Review

logical the ranking of the water supply described by Vitruvius would be, (Vitruvius 8.6.2.; Laurence, 1994:446), Laurence stops short of explicitly stating whether the supply in Pompeii was ranked. The description of ranking was used by Hodge to demonstrate the citing of Vitruvius over the examination of the evidence (Hodge, 1992:281), in particular by Callebut, (Callebut, 1973). Laurence avoids such controversy by not mentioning ranking at all.

2.7.4 To summarise The commonest material used for textile construction was wool, the commonest mordant was alum and the commonest dye was madder (Rubia tinctorum). Textile has been discovered with these three features in common which demonstrates that this may be used as an example of a typical piece. A study of wool dyeing with madder and alum would allow an understanding of the ‘typical’ piece of dyed textile from the Roman world.

It should be noted that although Laurence states that cisterns were still in occasional use, he takes this evidence from Vitruvius and not from the standing remains, (Laurence, 1994:44). As Hodge notes, Vitruvius did not record what happened in reality, instead he recorded his recommendations of what should happen, based on what he observed. This does not mean that his records should be taken as an example of what actually did happen, although, as Hodge notes, there is a tendency for this to happen, (Hodge, 1992:15). Hodge, who specifically examined the water source, supply and storage in Pompeii, states that the cisterns were not used, the aqueduct having replaced them. It should also be noted that Vitruvius warns against the use of cisterns and tanks due to the organisms that were known to grow in stagnant water and instead recommended the use of the aqueduct, a factor unreported by Laurence, (1994).

2.8 Undyed textile It may be noted that in collections of Roman textiles that are contemporary to Pompeii the majority of the textiles, and the majority of each piece, is undyed. Dyed fabric or yarns were used as decorative features rather than as a main background, such as clavi (lines) and gamma patterns in the main fabric, (Bender Jorgenson and Mannering, 2001; Pritchard and Verhecken-Lammens 2001; Cardon, 2001). This may lead to the argument that to determine the amount of dyed textile would still lead to an underestimate of the overall textile produced as not all textile was dyed. However, while there has been a fewer number of soft furnishings found, these tended to be of uniform colour and may be compared with depictions of furnishings which show them to be brightly coloured. Therefore to determine the amount of textiles dyed would still allow the final amount of textiles to be manufactured to be estimated.

Following examination of the evidence, it shall be assumed during this study that the water used in dyeing came via the aqueduct, which flowed over the sedimentary low-lands before reaching Pompeii.

2.9 Process consumables This section will discuss the process consumables, that is the items consumed during the process without which the process cannot take place, but which do not themselves form part of the finished product. In dyeing these include the water in which the dye is dissolved and the fuel used to heat the apparatus.

2.9.2 Water content It has been asserted by Storey that water used in dyeing needs to be ‘hard’ for the dye and mordant to be most effective. Storey even suggests adding chalk to harden the water if necessary, (Storey, 1978:67). It has been asserted that the textile manufacturing industry of other settlements, such as Bradford, took off due to the ‘hard’ nature of the local water, (James, 1990), and so the mineral content of the water was a serious consideration. However, this assertion is misleading as hard water is only required for certain recipes. As stated by Schweppe, ‘It is best to use soft water for dyeing. Only a few dyeing recipes prescribe the use of hard water, eg. Those for dyeing madder on iron mordant to obtain violet shades.’ Schweppe goes on to agree with Storey that to harden the water chalk should be added, (Storey, 1978; Schweppe, 1986). It is therefore possible to assert that it would have been beneficial to the dyers of Pompeii to use water that was soft as any hardening of the water could be done when the recipe required it.

2.9.1. Water supply The access to water was a critical issue to dyeing both in terms of quantity and quality. Laurence and Hodge both state that the aqueduct superseded the wells and cisterns as source of water in Pompeii prior to the AD 79 eruption, (Hodge, 1992:58; Laurence, 1994:44). The aqueduct and fountains were Augustan in date, built to provide better quality drinking water and to imitate the improvements in Rome, (Laurence, 1994; Richardson, 1988). Laurence’s statements regarding water source and supply should be regarded with caution as he approaches the use of evidence so heavily criticised by Hodge, (1992). Laurence states that the water supply was built to a similar design as that described by Vitruvius and uses Vitruvius to state the original source of the water. However, while Laurence states that the water supply is split into three upon entry to the city, as Vitruvius recommends, and states how

Geographically Pompeii should have had ‘soft’ water as Pompeii was constructed on ‘a tongue of lava reaching down to the coast’, (Arthur, 1986). The volcanic rock was 33

Investigations into the Dyeing Industry in Pompeii known of and used locally, (Peacock, 1989). Hard water derives from the water running over sedimentary rock before it reaches its destination. However it has been suggested that the water used in the dyeing process was not actually locally derived, coming instead via an aqueduct from Campagnia, (DeHaan, 1997; Hodge, 1981). This water would have flowed over sedimentary rock and so it is possible that it contained a higher mineral content and so was hard water.

Vitruvius knowingly records the symptoms of illness and death rate of lead workers through lead poisoning. If the lead content in the water had been high enough to be hazardous, it would have been recorded. Lead is of particular interest to the dyeing industry and to this research, but its presence is affected by the ‘hardness’ of the water.

The question of the hardness of the water in Pompeii has further relevance as the dyeing kettles and water pipes were constructed from lead. A high mineral content of the water would have decreased the lead that could leach into the water. As lead is a mordant that brightens colour, for example, when dyeing with madder, an increase in lead would have lead to a decrease in the need for a mordant. It was therefore important to determine the mineral content of the water.

The quality of the water that the dye is dispersed in may affect the outcome of the dyeing process. If there is lead within the water, its colour will be brightened as lead acts as a mordant. This would have reduced the need for the addition of a mordant, and although the chemistry explaining why would not have been understood, it was understood that water quality was different and had an effect, (Rosetti, 1548). There are two ways in which it would have been possible for the lead to get into the cloth in the dyers workshop. Firstly, the vat was itself constructed from lead, (Moeller, 1976; Monteix and Pernot, 2005). Secondly, the water that was piped into and through the city of Pompeii travelled in lead pipes, (Vitruvius, Book VIII; Hodge 1981).

2.9.3 Water quality

The demonstration that Pompeii had hard water and that the water did not contain lead goes hand-in-hand as the evidence of each provides evidence for the other. It is not possible to safely use lead pipes for drinking water if the water is soft water contained in a system that involves storage. To by-pass this problem either the water had to be hard water, had to be continuously flowing or the pipes should not be made from lead (Vitruvius, Book VIII). As the pipes discovered in Pompeii had been manufactured from clay and lead it may be argued the water was hard water, (Hodge, 1981). However, as the pipes were clogged with cinter (Hodge, 1981) and there are no recorded cases of lead poisoning from drinking water (Vitruvius, Book VIII; Hodge, 1981), it may be demonstrated that the water was hard water, thereby by-passing any perceived problems when using lead. The argument that if the dyers may have required hard water to dye madder is unreasonable as it is possible to add chalk as necessary to soft water, but they would still have required soft water for the majority of dyeing.

There has been debate over whether or not there was lead contamination of the piped water and to what extent it occurred. Following modern examples of lead poisoning it is supposed that the Roman populace were in constant danger of lead poisoning as the pipes used were a mixture of earthenware and lead, (Hodge, 1981). This has been stated by a contemporary (Vitruvius, Book VIII) and modern source (Hodge, 1981). However, as Hodge has argued, it may be demonstrated through a variety of methods that the lead content of the water was extremely limited. Firstly, there are no records of dyers becoming ill through lead poisoning, though records exist of lead workers becoming ill and the request to not use lead for water pipes, (Vitruvius, Book VIII). As lead poisoning was recognised it may be supposed that the dyers did not exhibit the symptoms. Secondly, there are some quirks to the system of water supply to Pompeii that mean that it was safe to use lead piping as it was not possible for the lead to contaminate the water. The water supplied to Pompeii came from an aqueduct supplied by a river. This meant that the water was constantly flowing, it was not possible to switch it off (even though there are taps fitted) as the system would back-up and flood the town. This meant that harmful bacteria and other organisms could not accumulate within the water as it could never stagnate. The ever-flowing water was also used to wash the streets as it over-flowed from fountains. While this seems wasteful in modern eyes, the sanitary benefits of this system should not be under-estimated, (Hodge, 1981). However, an additional effect of this constant flow was that there was not time for the lead to leach into the water, (Hodge, 1981).

There are a number of reasons to support the theory that the water was ‘hard’. Firstly, Hodge cites photographs and works written by a contemporary prominent hydrological engineer showing silting on the aqueducts leading into Pompeii. Hodge cites Forbes saying ‘Roman aqueducts suffered from serious deposits and incrustations of carbonate of lime and had to be cleaned out frequently’, (Forbes, 1955 cited in Hodge, 1981:488). The silt was known to be calcium carbonate, (or a close compound), and so the water was extremely hard. Secondly, the rest of Campagnia is over sedimentary rock and so it is reasonable to assume that the water derived from here will be hard, (Frayn, 1984). Thirdly, there were no reported cases of lead poisoning. Whilst this may seem presumptuous that the Roman medical technology would have been able to recognise lead poisoning, the Romans were capable of recognising lead poisoning. 34

Literature Review

The mineral content of the water also prevented lead contamination. The water was ‘hard water’. According to Forbes, cited by Hodge, aqueducts were notorious for silting up due to the hardness of the water. The presence of calcium carbonate in the pipes would have helped to prevent lead poisoning as it coats the pipes and absorbs lead and prevents contact of the pipe with water. This principle is still used in hydrological engineering today, (Forbes, 1955 cited in Hodge, 1981; Pers. Comm. Hopkins (government toxicologist) 1992; Pers. Comm. Hopkins (plumber), 1992). It is therefore possible to conclude that lead contamination was so limited that it did not lead to poisoning and that it is possible that it did not happen at all.

The problem with the use of Vitruvius as a source is that classicists and archaeologists are tempted to cite the works of Vitruvius rather than fully investigate and understand the principles that he described. ‘The temptation, when a problem arose, was always to reach for the volume of Vitruvius, conveniently to hand, rather than go to the site, often inconveniently distant, to look at the actual remains, much less, to cross the quad to that, psychologically at least, even more distant ultima Thule, the Faculty of Engineering’, (Hodge, 1992). Therefore any author who quotes Vitruvius without explaining what they take him to mean should be regarded with caution. An author who recites Vitruvius without explaining what they mean is unlikely to have visited the site to assess the technical points themselves, and is instead citing Vitruvius in the belief that this is how the site actually is, when Vitruvius instead wrote how he thought construction should be.

The amount of lead that may have contaminated the dye liquor is unknown. There are two sources and both would have caused an unknown quantity of contamination. The lead could have leached into the water from either the pipes or the dyeing kettle. It is unknown if it would have been derived from the water or from the vat. The amount leached by the kettle would have been affected by the acidity of the solution. Dye liquor of madder is acidic, (Hopkins, unpublished). The acidity of the solution would affect the colour of the textile. It is unknown how acidic the solutions used in the vats would have been. It is possible that the acidity of the solution would affect the amount of lead that would dissolve into the dye liquor. It is possible that acidity of the solution could have had a multiplying effect: an increase in the acidity would cause an increase in lead contamination thereby increasing the brightness of the fabric as the lead acts as a mordant. The affects of this are still unknown, but as they would affect the process are worthy of further study.

An example of this is Callebut, who cites Vitruvius as his statements are logical, and if constructed would have provided a fully functioning water system. Callebut assumes that as the design is detailed and functional it must also be the design used in Pompeii, when in fact the remains in situ to not resemble the writings of Vitruvius but instead were functional through a differing design, (Callebut, 1973 in Hodge, 1992). The principle in question is the priority given to public fountains over private water sources. In Pompeii the public water supply was given priority over the private, (Hodge, 1992). This would mean that as the dye works (certainly in workshops Vi4 and Vi5) were supplied by public fountains they would have priority too.

The quantity of lead that leached from the kettle during dyeing is unknown. While investigation into the exact pH and the quantity of lead leaching are beyond the scope of this study, this could be investigated through further work.

2.9.5 Fuel Fuel was necessary to the dyeing process, as it was needed for furnaces to heat water for dyeing and premordanting, (Moeller, 1976:29-56; Wild 1970 cited in Mann 1994; Mann, 1994). However there is debate as to whether this fuel was in the form of wood or charcoal. This is an important consideration as it would affect the running time and methodology of the dyeing process depending on which fuel source was used. Charcoal was known to the Romans and used in the domestic environment, (Coghlan, 1977 cited by Tonks, 1998). It had the advantages that it could burn to a high temperature (Horne, 1982) whilst producing little smoke or fumes (Horne, 1982, cited by Tonks, 1998). However, charcoal was brittle, expensive and difficult to transport, (Rostoker and Bronson, 1990, cited by Tonks). Therefore for a small scale industry wood may have been the better choice economically over charcoal. The textile industry in Pompeii would have had to import the fuel source into the city from

2.9.4 Reliability of ancient and modern sources on water According to Hodge there are two authors of the Roman world that predominantly wrote regarding architecture: Frontius and Vitruvius, (Hodge, 1992). Vitruvius wrote the ‘Ten Books of Architecture’ (‘De Architectura’) which included a discussion of the ideal way in which to construct a system of water distribution within a city. It should be noted that although Vitruvius appeared to have written a textbook, it was a statement of how things should be and not how they were. While it was based on the actual structures constructed on the ground it did not necessarily follow them faithfully, and as such should not be taken as a faithful report of Roman building practice, (Plommer, 1973; Hodge, 1992). 35

Investigations into the Dyeing Industry in Pompeii the surrounding hinterland, and additional processing costs could have affected choice.

that give explicit directions to the process. This means that to understand the dyeing process it is necessary to construct a recipe from those that exist in the modern era but which do not use modern industrial equipment.

It is believed that charcoal was more expensive than wood due to the difference in processing requirements. Wood could be imported without processing, (except possibly drying out). Charcoal not only needs drying out, but it also needs processing in a reducing environment in a way that requires additional wood, labour and cost. Following this it would need to be imported into Pompeii, but is fragile and so may not have survived the journey. It is unknown whether the dyers would have been able to secure their own fuel imports or would have had to buy it from another merchant. The charcoal would have been more expensive than the wood. However, the charcoal has many advantages in use over the wood. The main one is that it burns with twice the energy. The calorific value of pine is 18,650 kJ/kg, while for charcoal its 34,750 kJ/kg, (Goodger, 1980:51). This would mean that to produce the same amount of energy through burning, twice as much wood would be needed than charcoal. Charcoal has additional advantages in that it burns with a reduced smoke and is more predictable (wood could be of any thickness and any water content). The dyers would have been mainly concerned with cost and efficiency. Charcoal would have had many advantages in a busy inner-city workshop. Also, if the charcoal was less than double the cost of the wood, as it burned twice as well, it would have been cheaper to buy charcoal instead of wood. If the charcoal cost more than double the wood then it would have been more economical to use wood.

There is a great variance in dyeing recipes, both in times given for each step and the method by which the dyeing takes place. It has already been discovered (Hopkins, unpublished) that the important factor is not to follow each recipe exactly, but to allow the time or temperature required for each chemical change to take place and to ensure that these requirements occur in sequence. Dyeing is sequential and each stage depends on the completion of the stage before. The recipe reported by Storey (1978) allows for pre-mordanting and dyeing to occur within a matter of hours, the minimum time necessary for successful dyeing. Some recipe instructions are dubious and must be taken with the proverbial ‘pinch of salt’. Storey instructs to ‘bring to the boil’ frequently, (1978). However, if wool were to be brought to the boil or introduced into boiling water it would separate and clump. Therefore the water must always be kept below boiling. Also, if madder is brought to the boil, (as opposed to being kept just below it), it releases a yellow colourant thereby spoiling the dyeing outcome, (Storey, 1978). Raising the temperature increases both the amount of dye released from the madder and affects the type of dye released, (alizarin versus purpurin). This stabilises at boiling point. This instruction is in contrast to Rosetti’s instruction that requires the dye liquor to be brought to the boil while the dyer stood over it, stirring. A further advantage to Storey’s recipe is that only madder, alum, water and a fleece are required. Each of these is simple to obtain and safe to use.

Pool, raised a further economic point, (Pool, 1997). Charcoal is highly desirable as a household fuel as it is efficient and does not produce much smoke. This would mean that the dyers, (or in Pool’s case the potters) would be in constant competition with the householder for charcoal. This could raise the price artificially. It could reduce the incoming charcoal supply. Metalworking also requires charcoal, as it is possible to create a fire with greater heat with charcoal than it is with wood. This would mean that the metalworkers and householders would already be competing for charcoal, which could raise the price, leaving the dyers unable to afford it. This could reduce the price of wood as demand would be lower. It would be of benefit to the dyers if they were able to use wood instead of charcoal.

2.10.1The Recipe used in this study Recipe for mordanting It is required that 250g of alum per kg of wool should be dissolved in a vat of water. The quantity of water should be 4-5 gallons per lb of wool. (1lb= 0.454 kg. Therefore 10 gallons of water is needed for 1kg). Therefore for a 2kg fleece 20 gallons (90 litres) of water, 500g of alum and 1kg of madder would be required. Recipe for mordanting: 1.

Fleece (2kg), alum (500g), water (90 litres) placed into vat. 2. The water brought to the boil and simmered for 30 minutes. 3. The wool should be left in the vat to cool. 4. The wool should be kept damp until required. This should be at least 45 minutes, but ideally overnight.

2.10 Recipe directions There are no complete surviving Roman or Greek recipes for dyeing, (Monaghan, 2001). There are fragments of recipes, such as those in the ‘Papyrus Leidensis X’ and the ‘Papyrus Graecus Holmiensis’, (Lagercrantz, 1913; Vogler, 1982), which contain the consumables used in dyeing, but there are no recipes 36

Literature Review

Recipe for dyeing with madder:

made easier by the need to leave the wool to cool in the dye liquor before it could be removed. This allowed the further binding of the dye and prevented the fleece from matting.

1.

Hard water is required. If the water is soft chalk needs to be added. 2. Madder: 500g per kg of wool, (therefore for 2kg fleece, need 1kg of madder). Bring madder to the boil and then boil for 20 minutes. 3. Add the wool (still damp from mordanting). Simmer for 1 hour. 4. Allow to cool in the dye liquor. 5. Rinse several times. 6. Wash in ‘soap’ bath to brighten colour. (Soap was unspecified in recipe. The Romans did not possess ‘soap’ as understood in modern terms and so this part of the recipe is irrelevant to this study).

2.12 Quantity of textile When understanding the amount of textile that Pompeii was capable of producing, it is necessary to study the clothing and soft furnishings of the period. When the quantity of cloth needed per garment and how many garments were needed by the population as a whole may be defined, it is possible to place the quantity of textile manufactured in context. The textile may have been required for personal use (such as for clothing) or for furnishing (such as curtains or cushions for a household) or for a business, religious or government purposes, such as the decoration of a public building. Textile required for personal use would differ with status, for example, as a slave would require a small amount (possibly a single tunic) whereas someone economically better off would have required numerous garments. As textiles have not survived in great number in Pompeii, contemporary sites of a similar climate must be studied and used as analogy in order to determine the quantities used in Pompeii.

Times for recipe Table 2.2. Recipe for mordanting wool with alum, from Storey, 1978. Activity in Mordanting

Time (minutes)

Dissolve alum in water

10

Put wool in vat

5

Bring to the Boil

10

Boil

30

Cool

60 (1 hour)

Decant

10

Total time

125 (2:05 hours)

There are numerous sources of information that may allow an understanding of the requirement of textile. Surviving textiles allow a study of the technology available, the methods of manufacture, the shapes and colour of garments. Depictions (either in stone or through paintings) also allow an understanding of the technology available and the colours used. They also allow an understanding of how each garment was worn. There are four different types of evidence source that may be used to determine the type of article manufactured and its physical make-up and use. These are:

Leave for at least 45 minutes, preferably overnight. Do not allow to dry out. The fleece should still be damp for dyeing. Table 2.3. Recipe for dyeing wool with madder, from Storey, 1978. Activity in Dyeing

Time (in minutes)

Bring madder and water to boil

10

Boil

20

Add wool

5

Simmer

60 (1 hour)

Cool

60 (1 hour)

Rinse

10

Wash

10

Total time

175 (2:55 hours)

• Archaeological textile, the physical remains of the textile, from which it is possible to discern the shape, colour, materials used and the process of manufacture of the textile. • Depiction, where a drawing, painting or sculpture allows the discernment of shape, colour and use of a textile or an examination of the manufacturing process. • Text sources, such as those of Pliny, who wrote contemporaneously with the manufacture describing the process and the use of the end results. • Archaeological remains, such as the workshops in which the textile was manufactured.

2.11 Removing the water It was noted in the preliminary survey undertaken by Janaway and Robinson, (1994 unpublished) that some dyeing apparatus had drainage taps plumbed into the kettle walls near the base. However, the majority did not. It is therefore supposed that the vats were emptied using a bucket or a siphon. This task would have been

Through studying each of these sources and combining the evidence discovered it is possible to piece together the method by which textiles were manufactured, 37

Investigations into the Dyeing Industry in Pompeii the materials used and the final outcome to the manufacture. Once an understanding has been formed of how the manufacturing occurred it is also possible to determine the requirements for manufacture, the time taken to manufacture specific articles and the throughput of the textile industry.

The gown was probably white and yellow as it aided composition, and so would not be representative of the garments really manufactured. Popular paintings were executed without composition or the clash of colours in mind. They are far more likely to be a depiction of people in the clothes that they were wearing, for example Pompeii VII.3.30, tablinum, which depicts two men and a boy in calf-length tunic and cloaks at a bread stall. It may be argued that there could have been limitations on the palate, for example Pliny stated that some of the pigments were so expensive that wealthy sponsorship was needed (Pliny, Natural History, XXXV xxvi), but the detail of some of these pictures suggest that they are reasonably accurate.

2.12.1 Roman garments The most common form of clothing was a tunic. A basic Roman tunic was constructed from two rectangular pieces of cloth sewn at the edges and shoulder leaving holes for the head and arms, (Crowfoot, 1989). The length was from the shoulder to knee and sleeves were formed by the overlap down the arms by the shape of the rectangle. This basic design has been discovered throughout the Empire, both geographically and in time, with examples including ‘The Orator’ statue of 80-130 BC (Granger-Taylor, 1982); A depiction in ‘The House of the Vettii in Pompeii’; and tunic from The Cave of Letters in Israel, 130 AD. The fabric type and amount was the same in each of these with only the stitching techniques changing over time. The fabric was woven to shape thereby avoiding wasting material, a technique copied in the manufacture of Byzantine textiles, (GrangerTaylor, 1991), and Islamic textiles of Schweinfurth’s 1886 collection from Fajum Oasis in Egypt, dating between the 4th and 13th centuries AD (Linscheid, 2001).

Furthermore, in the Hellenistic style there was a tendency to depict the Gods and heroes in inaccurate clothing, single pieces of cloth artistically draped but entirely impractical, yet attendants in functioning garments drawn from life. (From depictions including for example ‘Europa and the Bull’ in Carolis, 2001). Croom states that when studying clothing and fabric depiction it is necessary to remember that the artist may depict the clothing as more ‘floaty’ than it would have been in real life – it may be depicted as taking up a greater volume of fabric than in actual fact had been the case, (Croom, 2000). Also the people depicted may be in stylised clothing or clothing that was from another era – similar to the Victorian depiction of an idealised Medieval, (Croom, 2000).

2.12.2 Pictorial representation of Roman dress Textiles and clothing that have been depicted allow some understanding of the style and manufacture of clothing, (Granger-Taylor, 1982). Through combining information gained through studying depictions and the finds of actual textiles, it is possible to build up a picture of clothing manufacture and use during the Roman period, (Granger-Taylor, 1982), and the dyes used in their manufacture. However, it is necessary to understand the artistic background of Pompeii when studying the depictions.

Colours used in depiction must be considered. Pliny had stated that while reds and blues were used for dyeing, yellows were not as this colour was associated solely with the bridal dress, yet in murals yellow is included in the colours worn, (Pliny, Natural History, XXI xxii). This may have been due to the internal decoration of a house containing reds, yellows, whites and blacks while blue and green were reserved for external renditions that suggested a natural or extended garden, (Ciarallo and Carolis, 1999). This would be detrimental to the accuracy of internal depiction if undertaken in the Hellenistic style. The Italian styles of portraits (the realistic rather than composed ones) show an expansive palate. There are reds, brown, and for renditions of the countryside yellows, blues and greens. These may be seen in Lararium, such as that said to depict Vesuvius before the eruption, (Ciarello and DeCarolis, 1999). This demonstrates that had the colours been used to dye clothes seen everyday, the colours were available to depict this. The tendency to portray clothing as dark and natural colours, even when this did not aid composition, suggests that these were the colours used in life. It has also been noted how accurate these depictions are. Even when the person was supposedly untrained they were still depicting items, scenes and people accurately.

According to Carolis there were two schools of painting in Pompeii – the Hellenistic and the Italian, (Carolis, 2001). The Hellenistic is the educated artistic style that worried over composition and used the Myths as its subject, the Italian was the everyday scrawlings done by the people of Pompeii to represent the world around them – for example it is the Italian style that is found by the Lararium as people illustrated their requests. The Italian style paintings are also called ‘popular pictures’. There is a sharp contrast in the compositions of the two styles, (Carolis, 2001) which leads to doubt in the accuracy of the Hellenistic style. For example in the depiction of Andromeda being rescued from the monster by Perseus (Pompeii VI.9.6, Casa dei Dioscuri) she is depicted as wearing a yellow and white gown. 38

Literature Review

population in time, this was just one moment, and so can not be representative of the population as a whole, (Wallace-Hadrill, 1994).

2.12.3 Problems with depiction Ryder has studied depictions of sheep. These are of interest in determining the breeds available to the Romans and the amount of wool each sheep would provide. There are few direct intentional depictions of sheep and of these few match each other, (Ryder, 1987). Depictions do appear to be a source used throughout the literature to determine what items were like from the Roman world, (although probably with varying degrees of success). They are used to study clothing (for example by Croom 2000) and by Ryder (1987) to study what sheep looked like. However, it must be remembered that the viewer is at the mercy of what the artist thought was worth recording and how it was recorded. If the artist wanted a chunky sheep to aid the composition the sheep would be drawn chunky even if in life it was not. If the artist didn’t want to draw a sheep then the sheep was omitted from the picture. Depictions have to be taken on trust to an extent and with the proverbial ‘pinch of salt’.

2.14 Intangible evidence As is the nature of archaeology, while the remains of the dye vats and properties may be studied in situ, and the literary evidence reviewed, the actual conditions that the dye workers operated in are unknown. The physical conditions of the workplace and the social conditions of the working day and social structure are intangible and may only be inferred from other findings. This presents a problem in the understanding of the operation of the vats, as the intangible factors would have influenced its operation. Emperor Severus Alexander complained that those who visited him smelled dreadfully of rotting fish, (as they wore imperial purple), and that movement in clothes with gold thread was difficult, (Croom, 2000:23). Freb claims that the report that dyers smelled (although there is debate as to whether this was of rotting fish from murex of imperial purple or from urine), was ‘proof for the pitiable condition of a dyer in early times’, (cited in Vogler, 1982). Small-scale indigo dyeing can still be found in remote areas, where it is still a labour intensive and complicated procedure to prepare it, (Balfour-Paul, 1999). It is still known as ‘the really unpleasant job’ as the people who undertake it smell and turn blue once they have finished, and the indigo has the reputation of causing illness, (Balfour-Paul, 1999).

2.13 Population To calculate how much textiles would need to be produced to supply the population the actual size of the population would need to be established. Storey has done this using figures and methods based on pre-industrial manufacturing and commercial towns, (Storey, 1997). He has arrived at a figure of 11,132 persons inhabiting Pompeii. The mechanism used to calculate this figure may have merit, but quoting a figure with this level of precision shows numerical naivety and begins to cast doubt on the accuracy. The figure does fit with others that have been calculated however. Jashemski has reported figures from 6,400 by Russell and 30,000 by Cary and Scullard, (Jashemski, 1979 cited in Storey, 1997; Russell, 1977 cited in Storey 1997; Cary and Scullard, 1975, cited in Storey, 1997), while Jongman has estimated 8000 and WallaceHadrill has stated 12,000, later revising the figure to 20,000, (Jongman, 1988; Wallace-Hadrill 1994 cited in Storey, 1997, Wallace-Hadrill in live interview, 2006). Storey has cited other studies that appear to confirm his findings of approximately 11,000. Following the evidence presented by Storey and other authors the figure of 12,000 shall be used for this study.

There is debate over the working conditions of a dyer. It appears due to the the nature of the industry that dyeing was an unpleasant occupation with a high incidence of illness and little respect. The physics behind the operation of the apparatus appear to suggest that dyers certainly had to be physically fit and were working to their limit. However this is not supported economically. To train a dyer would have taken years and each would have been expensive and time consuming to replace. It is therefore probable that they were relatively well treated. What is most likely to have occurred is what has been suggested by Mouritsen, (2001) in his study of the economic and social classes of Pompeii. Dyeing was seen as a socially low occupation, yet one that required skill and years of training. It is most likely to have been undertaken or at least overseen by the skilled underclass of freedmen and slaved who ‘still had the stigma of servitude’. Each slave or freed dyer was valuable due to their training, yet socially low due to their birth and status. A slave was freed under contract and kept in their original place of work as they were too valuable to loose and too dangerous economically to be a rival. Each

It may be noted that Pompeii was a holiday resort, with a transient population, (Storey, 1997), and an ex-colonia, with a retired population, and so demographic statistics regarding cause of death may have been skewed. It may also be noted that increased migration occurred following the AD 62 earthquake and the smaller tremors that led up to the final eruption of Vesuvius. As noted by Wallace-Hadrill, a population is constantly changing and that while the events of Pompeii froze the 39

Investigations into the Dyeing Industry in Pompeii would command new respect from their new status and each would continue in their role. This would mean that the trained dyers themselves were well treated and had some say over their work. Those that worked beneath them were in a more pitiable condition. It is probable that beneath the respected and/or freed overseers, were the dyers who were not yet trained (either because they were new or they were juveniles).

having been part of it. There are social conventions that were unrecorded and so have been lost over time. This is similar to Clarke’s concept of endemic theories – those that may only be understood by those aware of the internal processes within the culture. (Pers. Comm Taylor, 2002, citing Clarke 1978).

It is unfortunate when studying production in the Roman world that it is not possible to see it in process. Not only would it answer many questions about the working conditions and health of the workers, but it would also allow archaeologists to understand the social constraints and the social ‘norms’ that the people laboured under. However, it is possible to study accounts from the time and accounts from other time periods to construct an idea of what it would have been like. From this it is possible to assess what conditions would have been very much in the minds of the workers, but which we 2000 years later would be unable to even guess at. However there are parts that it is not possible to even contemplate as the ideas were based in a society that it is not possible to understand without

Having examined the evidence it may be concluded that the most commonly dyed material was wool and this was dyed ‘in the fleece’ before processing to yarn or fabric. The modern equivalent is a 2kg Shetland fleece. The most common dye used was madder and most common mordant used was alum. As no contemporary recipe survives, a modern equivalent has been constructed from an amalgamation of pre-industrial recipes. The water used in the dyeing process was not locally derived and so was ‘hard’ water. The water also did not contain lead. The fuel used to heat the dyeing apparatus was wood, although the species of wood is still unknown. The quantity of textile produced had to adequately meet the requirements of a population of 12,000 people.

2.15 Summary

40

Experimental Replica

Chapter Three

Experimental Replica in a purely theoretical understanding of the apparatus would not be relevant. However, this archaeological approach would result in an incomplete understanding of the apparatus, especially as there would not be the foundation of knowledge to allow an exploration of any conflicting factors. The theoretical engineering approach would also give an incomplete understanding. If the engineering approach were purely theoretical there would be no understanding of the method of operation of the apparatus. Furthermore, a purely theoretical understanding of the dyeing apparatus would not allow for any factors that became apparent during the construction or operation of the dye vat.

3.1 Introduction A theoretical understanding of the dyeing industry already exists. However, there are problems with this understanding due to the methods by which it was developed. The conclusions to date have been based mainly on theories drawn from studying the remains in situ and previous literary work, and Roman literature, (Pliny’s Natural History, cited by Edmonds, 1999; Pliny the Younger, cited by Frayn, 1984). These had been drawn from the structure and size of the vats themselves with recipe assumptions based on later preindustrial dye works and the writings of contemporary authors who lived within the Roman world who may not have been acquainted with dyeing. Although these studies have provided an insight into the operation of a Roman dye works, they should not be used alone to gauge the size and importance of the industry. Conclusions based purely on the theoretical findings are further complicated when the approach of the previous studies is considered. To further understand the dyeing apparatus and the capabilities of the industry there is a need to move away from a theoretical approach to a more practical and experimental approach. To understand how a dyeing apparatus could be operated and the parameters of this operation would require the construction and use of a replica. Following this it is possible to use the new understanding of the apparatus to define further questions and to question previous assumptions.

It was decided that the only way in which to test the theoretical understanding of the dyeing industry would be to build a full scale replica of the parts that influenced the manufactured outcome and determine how they operated. This required the construction and use of a full replica of a dyeing apparatus. A series of dye runs could then be undertaken to establish the limitations involved and the effects of each contributing factor. This information could then be used to calculate the volume of textile that it was possible to produce. This would be beneficial to both disciplines of archaeology and engineering: a replica would allow the test of archaeological theoretical assumptions and place the theoretical findings of engineering into context through an understanding of the operation of the dyeing apparatus. ‘Experimental archaeology’ is the scientific discipline that explores the design and function of artefacts and remains through their reconstruction and use. Any replica should fit the criteria of ‘experimental archaeology’ to be judged a valid part of archaeology. The construction of a full replica is within the umbrella of experimental archaeology, (Coles, 1973; Mathieu, 1999; Reynolds, 1999).

3.2 A differing approach Previously there had been no investigation into the implications of other factors, such as the type, quantity or availability of fuel used or the heat transfer through the apparatus during its use. The previous studies used the approach of archaeology. These additional factors would be seen by archaeologists as not falling in their own scope of research and also irrelevant to the question in hand, namely how the dyeing apparatus operated. An archaeological interpretation of ‘how’ means the actual physical operation of a dyeing apparatus by the operator, how to cause the apparatus to function and cause the production of the tangible end products. This is subtly different from the engineering definition of ‘how’ which means the internal physical science within the apparatus itself, how the fuel would heat the apparatus and the losses and movement within the system of the apparatus. These were seen by archaeologists as ‘engineering’ considerations, and so

3.3 Experimental archaeology It has been noted that although experimental archaeology has been in existence (if not as a branch of archaeology in its own right, then certainly as an implemented idea) from about 1860, (Coles, 1973:14), extremely little has been written about it in comparison to other branches of the subject, (Mathieu, 2001). The use of experimental archaeology to compliment mainstream archaeology, and allow a greater 41

Investigations into the Dyeing Industry in Pompeii understanding, was demonstrated by Crabtree, (1977). ‘Since most forms of Stone tool can not be observed in actual use, the most straightforward and simple way to determine their original function is to make copies and then attempt to employ them in a variety of ways. While absolute proof will always be lacking…’ (Yellen, 1977 reviewing Crabtree, 1977). This provides an analogy that may be applied to other experimental work. Crabtree used trial and error in his study, demonstrating that it is possible to consider more than one theory before starting an experiment. He concludes that ‘As often happens in experimental work of this kind, the conclusions may extend well beyond the problem as it was first conceived, and lead to additional questions.’ Crabtree concluded that functional experiments are hampered by a lack of knowledge in using the tools, and that once the knowledge has been lost it can not be regained, (Crabtree, 1977).

incomplete artefact the decision must be made whether to allow the artefact to be usable or to reconstruct entirely faithfully to the discovered artefact. This is alluded to by Dixon who states that although the building of the Telleborg reconstruction (a replica Iron Age meeting hall) was considered successful ‘it is, after all, still standing’ it was discovered afterwards that it was not accurate, (Dixon, 1976: 63). It has been noted that it is easier to determine what is missing from some replicas than it is others. For example, it is easier to determine what is missing from a ship, which would have been constructed to function as a ship, than a ditch, which could have had a multitude of purposes, (Crumlin-Pedersen, 1999). An experiment should be constructed in controlled conditions so that only one variable is changed at a time, (Crumlin-Pedersen, 1999). Mathieu states that it is better to construct a controllable environment than an entirely authentic one, as this allows each variable that affects the apparatus to be understood and altered in turn rather than guessing how the artefact works from studying it in the midst of many different and unquantifiable variables, (Mathieu, 2001).

‘Experimental archaeology … seeks to test, evaluate, and explicate method, technique, assumptions, hypotheses, and theories at any and all levels of archaeological research,’ (Speth, 1977). ‘Experimental’ suggests that the study should be set up as a standard scientific experiment: there should be a null hypothesis, an aim, a method, and it should be repeatable, (Mathieu, 2001; Crumlin-Pedersen, 1999; Coles, 1973). However the point of the experiment in experimental archaeology is to determine a null hypothesis by proving whether something is actually possible, (Coles, 1973). This usually occurs as the artefact has been discovered incomplete or the method of use has been lost over time, (Coles, 1973). Artefacts do not usually come with surviving instructions, (Crumlin-Pedersen, 1999). Even when there are depictions, such as those in Pompeii, and contemporary written accounts, such as Pliny’s Natural History, the process by which an artefact was used may still not be completely described. It may be argued that experimental archaeology is not conclusive – it does not actually prove that an artefact was used in a particular way, (Coles, 1973). Most experiments are negative – they demonstrate what cannot be done but do not prove what can – they may only suggest that it is possible, (Speth, 1977). Experimental archaeology is used to answer questions that cannot be answered through the application of theory alone, (after Lawson, 1999).

The complexity of a project may hinder its standing as an ‘experiment’. For example, Roar Ege, a replica Medieval ship, may be judged unrepeatable due to its complexity, (after Crumlin-Pedersen, 1999). An experiment should be repeatable, but a replica may be too expensive or time-consuming to repeat. However, as it is the time and money involved that precludes replication, not the actual replica itself, it still stands as an experiment. This may be demonstrated through the repeat in the construction and use of Ra I (repeated as Ra II) by Heyerdahl, (1971). Heyerdahl’s work also demonstrates the change in recent years between ‘experiment’ and ‘experience’ in archaeology: Ra I and Ra II were true experiments, Cabot’s Matthew (a replica of Cabot’s 1497 ship that sailed from Bristol to Canada) was a commercial venture to recreate the ‘experience’ in a safe and controlled manner. The Matthew not only set sail with radar and GPS, but also had an engine on board. It is rare in the literature to find an article that considers technical aspects of textile production through experimental work, rather than the elements of dye colour or fashion. What appears to be the only example of a technical article was written by Haynes in 1975. It is refreshing to find such an article, but it highlights the lack of others. A study of the relationship between indigotin and murex was undertaken by Edmonds at the Chiltern Open Air Museum, (Edmonds, 2003). While this work was detailed, it examined the chemistry and physical principles of the dyeing process rather than the function of apparatus.

The difficulties of experimental archaeology may be summed up by Hans Ole Hansen, (1959, cited in Stone and Planel, 1999). Ole Hansen states that even when a plan of a building exists it still does not tell how the building as a whole would have looked or functioned. He feels that if a person contemporary to the building were to see his efforts in its reconstruction they would find it hilarious. He highlights the problem of reconstructing an artefact when elements of it are missing. When constructing an experimental replica based on an 42

Experimental Replica

There has been exploration of which artefacts work and which do not, and which materials are the best for certain purposes. But there has been no deeper understanding of why a specific artefact works better than another of a different material or the physical processes of industrial apparatus. The closest to this so far is by Cotterell and Kamminga (1990) who used archaeological artefacts to demonstrate and explain various physics principles. But this was a fairly random approach, citing artefacts from a variety of periods to illustrate the theory, rather than an in-depth study of each artefact for a fuller understanding of each one. Speth just reported the physics behind the creation of stone tools and did not attempt to explain the terms to the archaeology public, or place the tools tested into a cultural context, thereby making the results useless within archaeology, (Speth, 1977). The majority of reconstructive archaeologists have reported the results of the experiments and then not explained the science behind what happened, (for example Coles, 1973). Cotterell and Kamminga, (1990), attempted to explore and explain the physics behind a selection of archaeological artefacts, but chose such a sporadic collection that it was not possible to learn and understand the principles to apply them to other artefacts and findings.

3.4 Constructing the apparatus There are two reasons to construct a replica apparatus: • To understand the building process • To determine the operating parameters under which the apparatus functioned To fully understand both the process of building and the design of the apparatus it is necessary to construct an accurate replica of a dyeing apparatus that has been discovered in Pompeii. The design must be accurate to allow an understanding of function. The apparatus chosen was Vat 5 in Property VII ii 11. This was an unflued vat of a relatively small size, originally identified by Moeller (1976) and re-examined and recorded by Janaway and Robinson, (1994, unpublished). As there is no record of dyeing apparatus having had lids, a preliminary dye run was undertaken to test the design and determine whether a lid was necessary. As lead and lime mortar, the original materials used in constructing the Roman apparatus, are hazardous it was decided that modern materials should be substituted in their place. However, there is an argument that this devalues the reconstruction as merely ‘experience’ archaeology and not an ‘experiment’, (Reynolds, 1999). The aim of this experiment is to understand the function and operating parameters of the apparatus. This shall be determined through the design of the apparatus. If the design is accurate the apparatus is valid as a replica of an apparatus from Pompeii. If the modern substitutes match the physical properties of each of the original materials the design is valid. The superficial aesthetic qualities of the materials are unimportant as they would not impinge on the function or the design.

The formulation of a null hypothesis before experiment, rather than the reconstruction of an artefact purely to understand its operation may be detrimental to archaeology as a whole. Each piece of experimental archaeology that has been undertaken as an experiment has been carried out with a definite, although narrow, goal in mind. This was beneficial to each project and it fulfils the ‘scientific’ aspect – there was a definite aim and the aim was fulfilled. However, while each project may be beneficial to those involved with the project, it can shed little light on similar projects. This has been highlighted through the work of Cotterell and Kamminga, (1990), as the questions they have asked of each artefact are so specific that the wider principles that govern their operation may not be expanded and overlapped to allow a greater understanding overall.

It was demonstrated through calculations by Watling that the specific heat capacity of the substituted materials had to be considered when choosing representative substitutes for the construction of the replica apparatus as this would affect how the heat transferred through the apparatus, (Watling, 2004; Hopkins et al, 2005). To ignore this factor would result in a differing quantity of fuel to be used during the dye run, causing an inaccurate understanding of the fuel requirement.

This present study strives to redress this imbalance by not only determining the output of the dyeing apparatus through both physical use of a replica and the application of theoretical considerations, but also by developing an understanding of the physical principles involved, such as heat transfer through the apparatus. The combination of these approaches shall be used to determine how the apparatus operated within a system, the inputs and losses from the system and the upper and lower limits of the maximum possible output. As each vat in Pompeii is different it would be advantageous if there were some rules that each conformed to, as the only other alternative to determine the output of each of them would be to reconstruct each one in controlled conditions (a method that would be unfeasible as there are 40 vats altogether).

It was noted that stainless steel and lead have such similar values for thermoconductivity, 27 and 35, respectively, (Callister, 2000) and specific heat capacity 0.460 KJ/(kg.K) W/(M.K) and 0.128 KJ/(kg.K) W/(M.K) respectively, (Callister, 2000), that they may be used interchangeably to understand how heat transfers through a structure, (see Figure 3.1). The ceramic surround proved more problematic as there was a range of values for the materials examined. However, as the apparatus were constructed from a combination 43

Investigations into the Dyeing Industry in Pompeii

Figure 3.1. Effect of different variables on fuel consumption, after Watling, 2004.

of rubble and mortar it was decided to substitute the values for brick, as this was a middling value of the range (see Figure 3.1). It was therefore decided to construct the replica apparatus from brick, mortar and stainless steel. This substitution resulted in the reconstruction being a physical representation of the original Roman apparatus both by design and heat-relating properties. This meant that calibration with the experimental work was possible. Watling concluded that the materials used in construction had the most significant effect on the quantity of fuel required. The volume of water used within the dye run had the second most significant effect. Watling’s findings were published as part of Hopkins et. al. (2005). The article by Hopkins et. al. is presented in Appendix Two.

a determination of what constitutes an apparatus, an understanding of the design and the context of the remains in situ. To mimic the dye run it was not necessary to actually dye the fleece. The runs were used to mimic the operation of the Pompeii vat, to measure quantity produced and not the actual quality. The runs were used to see the time it would take to heat and cool the vat and its contents, and the amount of fuel used each time. Prior to the dye runs it was also unproven whether a lid was necessary, or was used for retaining heat, as lids have never been found in context with dye vats in the archaeological record. It was also unknown whether bellows or a similar apparatus were necessary for ventilation of the fire.

3.5 Experiment One: Preliminary experimentation

The recipe used had been determined through preliminary laboratory experiments. Prior to this experiment using the replica dyeing apparatus a preliminary experiment had been undertaken using the replica to determine whether a lid was necessary. It was discovered that without a lid the loss of water from the apparatus was so great that had a fleece been present in the preliminary experimentation it would have been ruined. It was determined theoretically that a wooden lid of 20mm thickness halved the heat lost during the

Before the apparatus may be understood completely there is a need to explore the three factors involved in its operation: The dyeing process itself, including the consumables, chemistry and time; the heat transfer within the apparatus, specifically the fuel requirements and heat loss within the system; the user-apparatus interface, specifically the ergonomic considerations of how a dyer was able to operate the apparatus. An understanding of these factors will allow 44

Experimental Replica

experiment, (Watling, 2004). It may be supposed that the lid had been constructed from an organic material and so had not survived Vesuvius’ eruption or that older lids were eventually burnt when they began to break down. It was concluded that a lid had been used by Roman dyers and so a lid was used during the replica dyeing experiments.

the apparatus, an amount that would have led to the ruination of the fleece, (see Figure 3.2). Plywood was chosen for the lid material as it was of a known thickness and with known properties. If the actual material used to manufacture the lid becomes known, the properties of the material may be substituted appropriately into the equation to determine the heat loss from the dyeing apparatus.

3.6 Experiment Two

The replica was then used to determine the operating parameters of a dyeing apparatus in Pompeii.

Following the determination that a lid had been part of the dyeing apparatus, a lid was included. This was made from 20mm thick plywood. It was calculated by Watling that it would halve the heat lost by the apparatus and significantly reduce the water lost through evaporation, (Watling, 2004).

3.6.1 Hypothesis • The remains in Pompeii when replicated with the addition of a metal kettle form a functioning dyeing apparatus. • It is not possible to run a dyeing cycle more than once in 24 hours.

It was determined through a preliminary dye run that without the lid that 14 litres of water was lost from Calculations from Watling, 2004: Xl Rcond = kl

Rcond Xl kl

Rtotal = Rcond X Rconv Rtotal Rcond Rconv

= Conduction resistance = Thickness of lid = Thermal conductivity of lid

= Total resistance to heat loss = Conduction resistance = Total resistance to heat loss

Qwith lid = A x ΔT Qwith lid Rtotal A ΔT Rtotal

= Rate of heat transfer with a lid = Area of top of vat = Temperature difference between water and ambient air = Total resistance to heat loss

Qwithout lid = A x ΔT Qwithout lid Rconv A ΔT Rconv

= Rate of heat transfer with a lid = Area of top of vat = Temperature difference between water and ambient air = Total resistance to heat loss

Rcond = Xl kl

= 0.02 m = 0.17 W/mK = 0.1176

Xl kl Rcond

Rtotal = Rcond + Rconv Rtotal Rtotal

= 0.1176 + 0.0619 = 0.1795

Qwith lid = A x ΔT A Rtotal ΔT Rtotal Qwith lid

= 0.2376 m2 = 75oC = 0.1795 = 99.3 W

Qwithout lid = A x ΔT A Rconv ΔT Rconv Qwithout lid

= 0.2376 m2 = 75oC = 0.01619 = 287.9 W 45

Investigations into the Dyeing Industry in Pompeii Figure 3.2. Dyeing apparatus in use without a lid. 14 litres of water was lost during this dye run. (Thomas Halliwell looks into the apparatus.)

• Cooling the fleece in the dye liquor would be the limiting factor. • The quantity of fuel required should not exceed 5kg. • Charcoal may be used to heat the vat. • External influences such as weather could influence operating times and fuel use. 3.6.2 Apparatus Replica kettle (stainless steel) Replica brazier 9kg pine off-cuts Lid, 20mm thick wood (x2) Thermometer Stopwatch Sheep fleece, 2kg Shetland 90 litres of water 3.6.3 Diagram

Figure 3.3. Front view of apparatus. Kettle was constructed from 2mm thickness stainless steel. Original Roman kettle was 5mm thickness lead. Flange on both modern replica and Roman original was 0.05m

46

Experimental Replica

Figure 3.4. Aerial view of dyeing apparatus. Firebox is directly below kettle so they have the same diameter.

mortar, (the external surface was plastered with mortar). The kettle was constructed from stainless steel. This differed from the original Roman construction methods (lime mortar and rubble were used for the brazier and a lead kettle was used instead of a stainless steel one) but the conductive properties of the materials were the same (Watling, 2004) and they were of the same size so the design remained the same. The replica was therefore a true replica. The kettle was lifted into place. Ninety litres of water was placed into the kettle. A small fire was lit underneath using pine off-cuts and fed with off-cuts throughout the Figure 3.5. Photograph of dyeing apparatus, with lids, during use. experiment. The lids were placed directly on to the kettle top. The 3.6.4 Method water was brought to the boil and simmered for 20 minutes. The fleece was then added and simmered for A replica Roman brazier was constructed, replicating the one hour. Following this the fire was extinguished and design of apparatus V in property VII ii 11. The brazier the whole apparatus allowed to cool naturally. The fleece was 0.8m tall and 0.65m diameter. The kettle was 0.55m was allowed to cool in the dye liquor and then removed. tall with a diameter of 0.55cm. The firebox was 0.2m The use of a wool fleece and water and the timing of the tall with an opening 0.2m2. The brazier was constructed dye run allowed a full simulation of the dyeing process from modern brick and mortar and rendered with required for dyeing fleece with alum and madder.

47

Investigations into the Dyeing Industry in Pompeii 3.6.5 Results Table 3.1 Fuel used during the experimental dye runs. Ambient temperature 20 - 30oC Dyeing run One

Dyeing run Two

Dyeing run Three

Time (min)

Temp ( C)

Fuel (g)

Time (min)

Temp ( C)

Fuel (g)

Time (min)

Temp (oC)

Fuel (g)

00

13

1250

00

13

125

00

12

2480

o

o

15

14

30

13

2400

15

18

2075

30

30

45

14

2000

30

38

1075

45

51

60

26

45

53

1300

60

72

1075

75

32

60

72

75

78

1250

90

45

75

82

90

95

105

62

90

98

105

98

120

74

105

97

120

98

130

100

120

98

135

96

150

100

135

98

150

93

165

96

150

97

165

88

180

90

Total:

7510

2200

Total:

6725

Total:

130g ash

2025

600

8255 125g ash

Average fuel consumed: 7496.6g

the kettle was 126,195 kJ. As the calorific value of pine is 18,650 kJ/kg, this means that theoretically 7.99kg of pine would be needed to heat the dyeing apparatus (Watling, 2004), which is comparable to the actual quantities used.

3.6.6 Discussion With the addition of a metal kettle, the replica functioned as a dyeing apparatus. It was determined that the water within the vat took between 90 and 120 minutes to heat, but that external variables such as weather could lengthen the time taken or reduce the amount of fuel required. If the ambient temperature was relatively high (30°C), less fuel was required. If the ambient temperature was low (20°C), the water took longer to heat.

Charcoal could not be used to heat the apparatus as ventilation was insufficient. There were numerous attempts to use charcoal in the apparatus, but it would not sustain combustion. Even when a wood fire was lit and the charcoal then subsequently added the charcoal would cause the fire to go out. This was due to the calorific differences: charcoal has nearly twice the calorific value of pine, 34,750 kJ/kg instead of 18,650 kJ/kg (Goodger, 1980:51), and so requires substantially more oxygen to sustain combustion.

The process has a cycle time of 24 hours. The dyeing apparatus took at least 90 minutes to heat, a further hour to simmer while the dyeing took place, and over four hours for the dyeing liquor and vat to cool naturally to below 40oC. When placed in context this would allow for the process to be completed only once in the Roman working day of eight hours and so only once each twenty-four hour period.

3.6.7 Conclusions from practical experiment The experiment demonstrated that the remains identified in Pompeii as dyeing apparatus that correspond to this design may operate as dyeing apparatus with the addition of a metal kettle. There is sufficient space and airflow to allow the complete combustion of solid fuel (although there is insufficient airflow for the complete combustion of charcoal). There is also sufficient space for the free movement of fleece within the kettle, if the kettle is reconstructed according to the specifications allowed for by the brazier.

The apparatus was 20% efficient. The quantity of fuel required to heat the water in the dyeing apparatus and hold it at simmering temperature was five times the quantity required to heat the body of water and hold it at simmering temperature if it had not been contained within the apparatus, as calculated by Watling, (2004). If the specific heat capacity of water is 4.18 kJ/(kg K) (Çengel and Boles, 1998) and there are 90 litres of water to heat from 20°C to 95°C the energy required is 28,130 kJ (Watling, 2004). However, the energy lost to the environment while the water was heating and the quantity needed to heat the water contained within

Processing a fleece in the apparatus would have taken at least eight hours. Heating the vat, dissolving the 48

Experimental Replica

madder and simmering the fleece would have taken at least four hours and at least a further four hours would have been required to let the fleece cool in the dye liquor. As the Roman day was only eight hours in length it may be presumed that the fleece was left to cool overnight and the vat emptied and cleaned the following day. This would allow one cycle of dyeing every 24 hours.

•

It was possible to make conclusions about the operation of the dyeing apparatus following use of the replica. It was noted during the preliminary fieldwork (Janaway and Robinson, 1994, unpublished) that some of the dyeing apparatus had drains. These were in the form of taps built into the metal at the base of the vat wall that drained through a hole in the surround. However, each of these taps had a small diameter, usually less than 15mm. Any drain from the dyeing apparatus would have become blocked by the wool if the dyeing had not taken place in the thread or cloth phases of manufacture. If madder had moved freely within the water it would have become matted into the fleece, resulting in uneven dyeing, a point first noticed in preliminary experiments (Hopkins, 1998, unpublished) that was confirmed through use of this equipment. It is possible that dyeing did not take place in the fleece or that the madder was contained in a netting bag.

•

•

It was noted that while it was not possible to remove fuel to lower the temperature of the water, the Roman dyers would have been able to remove fuel and place it into an adjacent dyeing apparatus. This would have allowed them to better regulate the temperature. 3.6.8 Summary A number of new conclusions may be drawn form these experiments. • The Pompeii kettles could not have been lifted into place or removed whilst full. It was stated in the literature that the kettle was ‘emptied’ and refilled (Frayn, 1984; Wild, 1970), and it has always been assumed by those reading this that they were lifted and tipped to empty. However, the original kettles were made from lead. The kettle in this experiment was stainless steel and so much lighter, yet it took two people to lift it in and out of the brazier while empty and lifting it while full could not be done to any great degree, let alone for manoeuvre. The kettle was emptied using buckets leaving the wool to cool in situ, as this could not be removed until it had cooled naturally due to the danger of matting and to mimic the original recipe. • The steps built on to the sides of the larger Pompeii vats were probably necessary for stirring

•

•

49

and bailing out. Two persons of about Roman average height (1.64m and 1.53m) (Niblett, 2000:38) could manoeuvre the reconstructed vat fairly easily and fill and empty it. However, this vat was a small one. The Pompeii vats were left to cool overnight. First, the cooler night-time temperature would have aided this; secondly, following the results of the dye runs, it may be seen that it takes at least four hours, even when excess water has been removed. It would have made no sense to waste half of the working day waiting for it to cool. Cleaning the vat after scouring the fleece only took a relatively short amount of time. It took five minutes with a scouring sponge after the kettle had cooled and been removed. However it was necessary to place the vat on its side and climb into it to reach the bottom whilst draining the water used in cleaning. This also proves that a person of small stature could fit into the vat without any difficulty, as this was the easiest way to clean it. How long the kettle would have taken to clean after being used for dyeing is still unknown. The Romans did not use charcoal to heat their vats, if the brazier was of this design as the ventilation was inadequate. Wood was used for heating. This burned without any assistance, demonstrating that ventilation is more than adequate. If anything, the wood fire needed monitoring to ensure it did not spread. The fire box was the right size for wood and too large for charcoal. It may be argued that charcoal may have been used, and the replica may be inaccurate, but although the materials are modern, the design itself was copied faithfully from a Pompeii vat that was intact enough to allow the data to be collected. Therefore an inability to burn charcoal in this replica vat is a reflection of an inability to burn charcoal in the Pompeii vat. The persons attempting to burn the charcoal in the replica were experienced with outdoor activities including archaeological fire-lighting experience. Therefore inexperience of the people concerned is also unlikely to have been a factor in the failure of the charcoal burning. Vats of this design could have been used for scouring, but for the sake of cleanliness and to save labour it is probable that one set was used for scouring and another for dyeing. Adding cold water and bailing out the hot water worked in cooling the vat without matting the fleece. This decreased the time it took for the vat to cool, but it still took at least four hours each time.

Investigations into the Dyeing Industry in Pompeii • It would have been possible to use the fire for other uses without detracting from its heating of the vat, for example, cooking. The temperature at the mouth of the fire box was 70°C when in operation. It is unlikely that if the Romans had a series of small contained fires they would have lit a separate one in a different room to cook over. • Repeated heating and cooling of the vat compromises the structural integrity of the masonry surround. After two weeks the cracks began to show in the mortar and render. It is not known how long the surround would have lasted and whether any repairs needed would have been only superficial, or whether its integrity would have been completely compromised. • Following these experiments it has been possible to discount some theories, such as the use of charcoal, and to gain a better understanding of the dyeing process including the times that each procedure would have taken. Following use of this replica it is possible to see inaccuracies that have arisen in the literature due to the previous lack of practical work.

higher calorific value and so would heat the vat faster. Also it burned without smoke, so there would be no soot to stain the newly dyed cloth, (Horne, 1982). However, charcoal was more expensive than wood due to the further processing it required, more difficult to obtain, and difficult to transport as it is brittle, (Rostoker and Bronson, 1990). In terms of the amount to be transported wood appears to have been the cheaper option. Despite this it appears from the calorific values that charcoal may have been the better option economically, unless the wood was less than half of the price of charcoal: the charcoal has double the calorific value and so would have to be twice as expensive before it was less economical to use. However, during these experimental runs it was discovered that the ventilation in the vat was wrong for the efficient burning of charcoal. There was insufficient oxygen flow for complete combustion. The charcoal lit, but then went out. The amount that could be placed in the fuel box was limited as piling it up extinguished the charcoal that was below. This left the rest of the box empty. The entrance to the fuel box was certainly hotter than with wood, it was impossible to place more charcoal on by hand, yet the vat still took seven hours to heat. In smithying there is a forced draft. Here there is not and no design has been made for one. Dyers would have been in competition for charcoal against the blacksmiths and householders. This may have driven the price of charcoal up or wood down, resulting in their actively seeking to use wood. There is modern ethnographic evidence to support this as in modern Mexico the potters use palm leaves which have very low calorific value but which are exceptionally cheap due to there being no need for them in relation to the vast demand for charcoal, (Pool, 1997).

3.7 Discussion 3.7.1 Reconstruction There are additional factors that the practical implications of the replica dye run may impact upon. Theoretically some variables, such as fuel type, were dependent not just on practical restrictions, but also on economic ones, and therefore would have needed more than a practical reconstruction to discount them. This would have been more difficult to prove as the main source of contemporary economic theory is Diocletian’s AD301 price fixing edict, which firstly may not have mentioned everything that was relevant and secondly was written two centuries after the volcanic eruption. (Diocletion, AD 301, in Lewis and Reinhold, 1966). However, it would have been useful in giving a general ratio of costs, such as between wood and charcoal. In the event however it was not necessary to refer back to this documentation.

The reconstructed vat was used for scouring the fleece. Vats of this design could have been used for scouring the fleece before dyeing. However, it would probably have been better to use one set of vats for scouring and another for dyeing due to staining. By the end of the second week, cracks had appeared through the rendering in line with the bricks underneath. This appears to have been due to the heat. This may have presented a problem had the dye runs continued, but poses the question of how often the braziers and vats had to be either repaired or replaced. It appears that patching over the cracks may have been a common practice in Pompeii. It may also be seen that the paving slabs that the vat was built on are collapsing inwards in the centre. This is also probably due to the heat.

Pine off-cuts were used for fuel during the experiments. This was because there was a sufficient supply of pine. At present the type of fuel used in the dye works is unknown, but it would be possible to convert the energy required to heat a dyeing apparatus from pine to the fuel actually used using the relative calorific values, when the fuel type is determined.

3.7.3 Relative energy of wood and charcoal

3.7.2 The use of wood or charcoal

Fire has been defined as ‘High-temperature selfsustaining oxidation’, (Rossotti, 1993). Once a fire is lit it has to stay lit. There needs to be a continuous supply of air, fuel and heat – the moment one of these

According to the theories of thermodynamics (Pers. Comm. Seale, 2002; Çengel and Boles, 1998), it would be in the interest of the Romans to use charcoal as it had the 50

Experimental Replica

there was sufficient oxygen to allow combustion of wood, but not to allow self-sustaining of a charcoal fire. The charcoal in the brazier failed to continue burning, even once lit, because it could not reach a high enough temperature to generate a sustaining influx of air, (Pers. Comm. Seale, 2004).

ceases, fire itself ceases. Once lit a fire must maintain a high enough temperature to generate a sustaining influx of air to allow combustion to continue. Size, temperature and flame all depend on fuel, air and the exact physical conditions present. Carbon combines with oxygen to produce either carbon monoxide (CO) or carbon dioxide (CO2), (Rossotti, 1993). It is not possible to give a full mathematical description of burning as it is not possible to predict the mixing of fuel vapour and air, (Rossotti, 1993). It involves both diffusion and convection, (Rossotti, 1993). Convection is the most important form of heat loss from solid fuel fires as the heat rises. ‘When the diameter of the fire is larger then about 1/3 of a metre, radiation of heat becomes increasingly important, although heat loss by convection continues to occur’, (Rossotti, 1993). The fire in the dyeing apparatus remained between 0.250.5m in diameter throughout the experiments. Some of the energy released must transfer back to the fuel to sustain combustion. Three logs burn better than one as two of them radiate to the other one. Exhaust gases must escape to allow fresh supplies of fuel and oxygen to be sucked in, but these must be allowed to heat up before heat is lost through the chimney (through the loss of the exhaust gases), (Rossotti, 1993).

Figure 3.6 demonstrates why wood is self-sustaining, but charcoal is not. The moment that it is lit, the amount of energy required to sustain combustion increases, as the input of energy increases, the quantity of oxygen required to utilise this energy within the fuel also increases. If the oxygen cannot be brought in at the rate needed then the energy within the fuel is not released and the fire will go out. A forced draft helps this process as it increases the amount of oxygen that is present. Depending on the chemical make-up of the fuel a differing amount of oxygen is required. Charcoal has twice the amount of energy per unit mass than wood does. This means that it needs a significantly larger amount of oxygen immediately following the point of ignition to release the next amount of fuel (the next part to be burnt after ignition) for burning. If it fails to do this the flame will go out. In the dyeing apparatus there was insufficient oxygen to allow the charcoal to sustain a flame, even once it had lit.

Charcoal ignites at 300oC. Tinder ignites at 120oC. The tinder has an increased surface area and access to oxygen. The charcoal has more energy per unit mass. The combination of these factors means that wood will burn a lot more easily than charcoal, as less oxygen is needed. ‘Campfires are… difficult to ignite, and, particularly in the initial stages, the sluggish combustion may not provide enough draught to suck in sufficient oxygen’, (Rossotti, 1993). This was demonstrated practically during the experiments as

In a reduced atmosphere, an atmosphere with little oxygen, there is enough heat to pyrolyse the wood products, and they burn to charcoal, but there is insufficient gas to allow complete combustion. Charcoal may smoulder for days with insufficient oxygen, as the heat is intense enough to support it, but will not support flames. Coniferous wood is more flammable than deciduous, (Rossotti, 1993). Confined fires burn more intensely than open ones. A fire needs a draught, but not too much. Fuel should not be consumed too fast otherwise the fire will get hotter than needed. Gusts of wind can sweep hot gases away and deprive a small fire of heat, thereby causing it to go out. It is possible to protect the fire from a wind through a windshield, or from all draughts by a circular enclosure. Convection allows the escape of waste gases, but also the loss of heat. Cold air can get swept into the fire, (Rossotti, 1993). Wood requires a limited amount of air to continue combustion (without flaming), (Rossotti, 1993).

Figure 3.6. Diagram showing ignition by input of energy. From Rossotti, 1993.

51

If a flue is present the change in density of the warm air (the density between the air at the base when compared to the air at the top of the

Investigations into the Dyeing Industry in Pompeii flue) allows it to rise up the flue, (Fullick, 1994; Çengel and Boles, 1998). Therefore the draw is greater in a fire assisted by a flue. While the air travels up the flue it loses heat to the sides and so indirectly heats the vat. While the only thing driving the movement of the air is the heat generated by the fire, it should be noted that the combustion may only be sustained if the movement draws sufficient new air. A flue allows the increased movement of air as the exhaust air may be released in a greater volume thereby increasing draw.

unpleasant conditions. Even if the workers were able to leave the room and breathe freely the atmosphere within the room does not support their working, even if they were healthy, young and acclimatised to the extremely high temperature. 3.8 Further work Following the practical experiments using the reconstructed replica it was discovered that there were still unanswered questions. It is possible to discern how a single dye vat operates, for a given definition of ‘how’ (the archaeological definition), but it was not possible from the single replica alone to discern how a vat operated in context with other vats and within a dye works. Also, it was not possible to discern the difference made by a slight change in design. The replicated vat had been chosen as it appeared to have been unaltered, had been thoroughly recorded and stood alone.

Charcoal burns more cleanly than wood, with more control, (Rossotti, 1993:97). Smoke is a noxious byproduct that must be expelled. If it has an organic origin the solid part of the smoke will be soot, carbon particles caused by incomplete oxidation, (Rossotti, 1993:149). If a dye works were to be lighted at night, there would be the question of how this would be done. Fumes from the light source would build up and the soot would affect the fleeces and dyes. Lighting at night would be a source of contamination for the scoured or dyed fleeces. Also, as the dyeing cycle is 24 hours to allow the fleece to cool in the dye liquor to allow the dye to stick and the colour to become fast, it is advantageous to dye the fleece in the day and leave it overnight to cool.

When Moeller’s original survey had been undertaken, the knowledge of dyeing was purely theoretical. The later survey, used in this study, was undertaken with a view to recording each vat in detail and with a background of practical experience in dyeing, but even during this survey there was no specific understanding of how the apparatus actually operated. This would mean that if a feature was missing that was necessary for the operation of the dye vat, or if a feature had been added or amended, there would be no way to discern this from the survey that had already been undertaken. It was discovered on use of a replica that ventilation within the vat could influence the combustion of the fuel and the method of using the vat. It was also noted how labour intensive using the apparatus was and how much lifting and storage of related articles would be required. It was noted that a small modification in design could have greatly altered the influence of any of these factors on the dyeing process. As these factors could have influenced the amount produced it was important to review what was understood about each of these vats and their contextual surroundings. It was therefore necessary, following the experiments performed to understand how the apparatus operated, to undertake fieldwork in Pompeii to accurately survey the apparatus remains in situ. This would allow an understanding of the apparatus within the context of dyeing. It would also mean that any subsequent replicas and experiments would be more accurate. The survey of the dyeing apparatus in situ is discussed in Chapter Four.

Fire is a positive feedback loop, (Rossotti, 1993:181). Hot air and fumes rise in a plume and collect at the highest point possible in a room because they are less dense than air. This spreads into a layer under the ceiling and flows back into the room. Carbon monoxide results from incomplete combustion due to insufficient oxygen and has a devastating effect. Carbon monoxide has a greater affinity than oxygen to haemoglobin, combining 200 times faster than oxygen and being slower to release. Exposure to CO results in weakness, dizziness, and impairment of decision making and consciousness during the time of exposure, but can lead to the development of chronic psychological and memory problems and disturbance to the victim’s gait following the CO poisoning. Chronic exposure gives flulike symptoms. While the majority of people recover following exposure, recovery may take a year, (ToxBase, 30-6-07). Smoke irritates the lungs and increases their susceptibility to toxins contained within it. Toxic smoke inhaled below the lethal level stupefies victims making sensible decisions impossible, (Rossotti, 1993:183). Given these figures and the size and ventilation of dye works it is looking increasingly likely that dyers had a limited life expectancy or at least worked in extremely

52

Chapter Four

Review of Remains in situ 4.1 Fieldwork in Pompeii

4.2 Gazetteer of the dye vats discovered in Pompeii to date

The throughput of dyeing apparatus contributes most to the final output of the dye works and the size and scale of the industry. Therefore a full survey of the dyeing apparatus was undertaken and the apparatus viewed in the context of the workshop to gain an understanding of their design. To date, Moeller had carried out the most comprehensive survey of dye vats in Pompeii, (Moeller, 1976). He had identified apparatus and properties that he believed to be dye works. However using the data from Moeller’s survey was problematic because it was based on his own theoretical assumptions as to what a dye works would require rather than applying criteria gained through a practical understanding of how a dyeing apparatus may work. Also, while Moeller’s survey was comprehensive in that he surveyed a geographically wide area, it was not detailed as to the design of the apparatus, meaning that it was not possible to reconstruct an apparatus and understand its design purely from his survey.

This project examines the quantity of textile that may be processed through various dyeing apparatus. The Gazetteer documents the measurements of each of the dye vats while the plans document the placement of each vat within the properties and the location of the properties within the city. The gazetteer is divided into two parts, the first section documenting the evolution of the definition of a dye vat and the second part documenting each vat in detail. Moeller’s work provided a foundation to the understanding of the dye vats, both in their identification and their capacity of production, (Moeller, 1976). He estimated the size and influence of the dyeing industry (although he failed to list the criteria that he used). Moeller has since been heavily criticised. In the review of his work (Wild, 1977) he was accused of making assertions that were beyond the extent that the evidence could have taken him. But his identification of the dye vats included here was not criticised. In Jongman’s work (1988), Moeller was again criticised, but the identification of the dye vats was again unquestioned. In each subsequent work the identification of the dye vats have not been called into question, (for example, Mann, 1994, Hopkins, 2002). The main criticism levelled at Moeller was his economic and operative assumptions. The identification of the vats themselves has never been questioned.

Given the ambiguities resulting from Moeller’s survey it was necessary, as part of this thesis, to reexamine the assumptions that had been used in his identification of dye works and dyeing apparatus. This fieldwork was the first to take consistent and accurate measurements of the dyeing apparatus and workshops (through photographs, measurement and the construction of plans). The recording forms showing the criteria and method for recording may be viewed in Appendix Three. This was also the first time that each of the apparatus was viewed in context and with a working knowledge of the vats gained through replica experimental work, allowing the assessment of whether each vat could have been used in the way that Moeller suggested. It was discovered during the reassessment of each of the apparatus during the fieldwork that Moeller had not been entirely accurate. Furthermore, the data collected during the fieldwork allowed further models and replicas to be constructed resulting in a greater understanding of the operation of the dyeing apparatus. This understanding together with the data collected allowed the construction of a computer simulation to further examine the properties of the materials used and the limitation these had on the output.

In the current investigation the dye vats were examined in a new way. Previously the vats have been examined by persons familiar with dyeing and textile production, through the application of theory to the remains in situ. This present work incorporated a reassessment and re-measuring of the remains in situ with full scale reconstruction to test the assumptions that had been tested through theory – whether or not these structures could have operated as dye vats. The assessment evaluated each aspect of the physical process of dyeing and whether the apparatus could have been used for that task. The physical use of the apparatus was examined, not just in the processing of the fleece but the additional tasks such as cleaning the apparatus after use. The limitations that these tasks placed on the time required or that the physical

53

Investigations into the Dyeing Industry in Pompeii properties of the vat placed on the ability to perform these activities were considered. This assessment was more conclusive and definitive – if the structures could not have operated as dye vats or another part of the dyeing process then they should not form part of the understanding of Pompeii’s dyeing industry.

understood a further set of attributes were defined, and these added to the original list. Attributes of a dye vat following the operation of a replica: • The operator must be able to reach the base of the water container when it is in place. Steps may be necessary. • The volume of the area or kettle capable of holding the liquid must be great enough to allow the free movement of fleeces. A fleece requires 90 litres. • There must be adequate ventilation to allow the complete combustion of the fuel. This may be achieved through the location of the dye vat or the addition of a flue.

4.2.1 Defining the dye vats During the summer of 1994 preliminary fieldwork had been undertaken in Pompeii by Janaway and Robinson, (Janaway and Robinson, unpublished). They undertook a survey of the dyeing remains identified by Moeller, (1976). Although they had a working knowledge of dyeing procedure they had not seen a replica dye vat in operation. This meant that although they measured and photographed the standing remains with accuracy they were not able to appreciate the remains in context. The recording, while accurate, reflects this.

In addition to this a list of attributes has been defined for the dye properties: • The property must be close enough to a source of water to allow the filling of the vat. Vats take a minimum of 90 litres, so the source must be capable of delivering a substantial amount. • There must be a means at the property of disposing of the water following dyeing. • There must be a storage area at the property for consumables such as fleece or fuel. The size of this storage area is still under investigation.

It was noted that each time a ‘dye vat’ was identified, whether the authors of the study were aware of it or not there had to be a criteria that was being used for this identification, (Robinson, pers comm. 2005). Without this the ‘dye vat’ that had been identified by Moeller but which was discounted in the 2002 fieldwork following closer examination would not have been excluded. The question was then posed: what were the criteria, and how had each dye vat been constructed.

There is the additional criterion of context. Dye vats are usually found located together. This may be due to the nature of dyeing – to operate a dye vat is labour intensive at different times during the process and also requires large quantities of consumables. Therefore dye vats are usually located together to save time, effort, transport and consumables. A solitary container that meets all of the other criteria may have had another purpose.

Following closer examination of the vats, a set of attributes (after Clarke, 1978) was established that each dye vat could be seen to possess: • The ability to hold a volume of liquid or a kettle that could hold a volume of liquid. • The ability to heat the liquid, the heat emanating from below the liquid. • A masonry structure large enough and strong enough to safely hold the liquid and the heat source, without the spillage of either.

When assessing the dye vats for inclusion in the gazetteer it was necessary to evaluate each one, and the properties that they were contained within, according to these criteria. All of the dye vats that were included have been shown to match the criteria and be capable of being used as dye vats. Each dye vat was examined in the field and assessed according to the list of attributes obtained through theoretical study of the operation, the study of the engineering principles regarding airflow, and the experimental work. Only vats that could satisfy all of these criteria were included.

This set of attributes are called the polythetic entities of an artefact. They are the physical attributes that an article may be seen to have, and the possession of a multiple number that coincides with the attributes of another artefact leads to the suggestion that both artefacts may have served the same purpose. Polythetic entities may be gauged by an external viewer of the artefacts as they are physical attributes, (Clarke, 1978). Endemic entities are the intangible attributes that are assigned an artefact by the culture that produced or used the artefact. These may not be discerned through viewing by an outsider to a culture and so become lost over time, (Clarke, 1978).

It should be noted that attributes in the field may be difficult to define due to the degradation of the artefact under examination, (after Clarke, 1978). This was taken into consideration, and where artefacts had degraded or been restored to the point that an attribute was not recognisable or had been substantially altered this was noted and included in their assessment.

During the experimental work a replica dye vat was constructed. As the operation of the vat became 54

Review of Remains in situ

a gazetteer recording each of the dyeing remains and placing it in context within the property and in relation to each of the other apparatus, are included at the end of this chapter on pages 63-95.

4.2.2 Extent of excavation in Pompeii. The extent of excavation within Pompeii may affect the final conclusions of any study of the city. Only two thirds of the city has been excavated to date (Laurence, 1994) and as part of the city is below the modern town of Pompei it may never be excavated at all. Therefore it must be remembered that any conclusions about the size and scale of the industry may be incomplete as further evidence may as yet be undiscovered. However, the geographical size of Pompeii is known and the majority is excavated, including large structures such as the amphitheatre. Therefore there should not be undue concern about the reliability of conclusions based only on the excavated areas.

4.3.1 Summary of dye vats Each of the vats in the properties designated by Moeller was examined individually first-hand during the fieldwork undertaken at the start of this study. The following is a brief summary of the conclusions. A full account of each vat is presented in the second section of the gazetteer. The seventh property was discounted and is discussed separately below. Property I viii 19

Excavation is ongoing in the city. Previously discovered properties are still the subject of excavation. Borgard et al (2003) document the excavation of dye works Vi4 and Vi5. This work appears to be the most current work published on the excavation of dye works. The article is of interest to this study for a number of reasons. Borgard et al (2003) state that they followed the work of Moeller to guide their choice of which properties to study as dye works. This shows that again Moeller (1976) is taken as accurate without question. But the authors choose different properties to the ones defined by Moeller without properly defining their reasons. The article as a whole documents the architectural sequence of each building excavated, but does not examine the production of the industry so is not directly relevant to the current study. There are some statements made regarding the production rates of the dyeing industry, but these are generalised and presented with insufficient evidence to support them.

These structures are regarded as some of the most complete dye vats in Pompeii. They retain enough of the brazier structure to allow an understanding of design and height, and the original kettles. It was determined through a combination of examination in situ and replication of the design that these structures when complete could each have operated as dye vats. Properties V I 4 and V I 5 The structures within these properties have been heavily restored. It has been determined that their size and design would have resulted in their being suitable as dyeing apparatus, although research is continuing into the limitations on construction materials that the larger vats may have posed. It appears that the restorer was unaware of the operation of the vats, having blocked flues and similar design features necessary for their operation. However the presence and design of these features may be determined through an understanding of the operating methods.

Thus far six dye works have been discovered in Pompeii. Their distribution and relation to other buildings and the routes within the city has been studied in detail by Laurence (1994) and Robinson (1999) amongst others. Robinson wrote as a reaction to Raper (1977) who stated that there was no structure to the distribution of the dye works within Pompeii. Robinson also undertook a study of which social classes participated in the dyeing industry and related his findings to the location of the dye works. He related this work to the continuing debate about the economic organisation of Pompeii and the place of the dyeing industry within this. His approach and conclusion are documented in Appendix One. Laurence’s study (1994), has already been commented on within the ‘Strengths of weaknesses’ section of the ‘debate so far’ section of Chapter One.

Property VII xiv 17 The structures contained within this property show signs of heavy restoration. There are also extensive signs of robbing and restoration that masks this. Some of the structures contain the original metal kettles (known to have been in place before restoration as they have been cemented into place in a damaged state). However the structures themselves appear to follow the same basic design, including shared flues, and so the parts missing could be extrapolated from the more complete structures. It was determined that all of structures could have originally been used as dye vats – vat five was a comparable size and design to the replicated vat from property VII ii 11 and so was used to provide the replica of the flue.

4.3 The Survey and Gazetteer To fully understand the standing remains of each apparatus and place each within its context, it was decided that a full reappraisal of each of the remains in situ should be undertaken. The results of this survey,

Property VII ii 11 The structures contained within this property do not appear to have been restored. Their size and design has 55

Investigations into the Dyeing Industry in Pompeii been shown to be appropriate for a dyeing apparatus – vat five from this property was reconstructed as the replica.

and experimental work using the replica dyeing apparatus. It was concluded that six properties could have operated as dye works, but that the seventh (located in I xii 4) could not.

Property IX iii 2

The ‘vat’ contained in the property proved a test to the criteria. Figure 4.1 is a photograph of the apparatus in situ. It was concluded that this was not a vat because:

The structures contained within this property could not have been used as dyeing apparatus in their current state. While it would be possible to heat and empty them it would not be possible to clean them. This gave rise to questions as to how these could have been used as dye vats. However it was discovered on closer examination that the vats had been heavily restored. Their original heights were still discernible and it was determined that if the structures had been this original height they could have been used as dye vats.

• It was located singly within the property. Dye vats are usually located together or with other textile processing equipment. Furthermore the vat was located at in the middle of the property, suggesting the need for all-round access, when all other vats are located at the edge of the room, suggesting such access would not ordinarily be required. • The room has no ventilation and the firebox is extremely small. It is doubtful whether complete combustion would have been achieved. • Whilst the brazier is large, the space taken to allow the construction of the firebox has resulted a much reduced space for the kettle. This means that the brazier can only hold a kettle that is 15cm deep. This would not provide sufficient room for the movement of a fleece during dyeing.

Having examined the dye vats in situ and determined their operation through use of a replica it has been concluded that each of the structures included in this study may have been used as a dye vat. 4.3.2 The discounted ‘dye vat’ There were originally seven dye workshops identified by Moeller, (Moeller, 1976; Laurence, 1994). These were re-evaluated during the fieldwork undertaken in 2002 (Hopkins et al, 2005) using the list of attributes defined using a theoretical understanding of the dyeing process

It was concluded that this is not a dye vat and the property is not a dye works.

Figure 4.1: The vat in the seventh property identified by Moeller as a dye works. Note that the depth of the firebox prevents the inclusion of a kettle that could hold a fleece within the brazier. Source: Author

56

Review of Remains in situ

to Pompeii as a whole and determine the source of water most convenient to each dye works. It is supposed that cisterns and wells were not used to supply the dye works of Pompeii and that instead the aqueduct was used. The wells and cisterns were abandoned after completion of the aqueduct in the Augustan era, (Hodge, 1992; Laurence, 1994, see Chapter Two). Also, it may be reasoned that as the dye works would have required a substantial quantity of water from a reliable source it was most likely that they utilised the public source of water rather than a temporary, smaller source. However, to test whether this is a practical idea, as after all the Romans were pragmatic and would have relied on the cheapest and easiest practical solution, it is necessary to determine where the sources of water were in relation to the dye works.

4.3.3 Flued vats It was discovered during the survey that there were two basic designs of dyeing apparatus: the flued and unflued vat. The flued vats were of the same design as the unflued, the only difference being the presence of a flue. The apparatus in property VII xiv 17 and in property VII ii 11 were of virtually the same size and design, the only difference being that those in VII xiv 17 had flues. It was noted during the survey that the flued vats existed in properties that were sheltered or enclosed. It has been supposed from this that as the flue would have required extra labour, materials, time and technique to build that it was necessary. It has been deduced that there was a need for a flue for each of these vats as the ventilation was different on each of them. It has therefore been deemed necessary to understand the difference that a flue makes to a vat, an answer that may only be understood through further experiment. Therefore it has been decided to add a flue to the original replica vat to allow comparison between the two combustion processes. To add this flue will cause the replica vat to cease to be a replica of vat 5 in property VII ii 11 and instead become a replica of vat 4 in Property VII xiv 17. The replica will still be accurate as the height, width, size and volume of the two braziers are the same, as is the volume of the two kettles. However, there is a discrepancy between the two braziers which has caused a problem with the reconstruction. This is discussed in detail in Chapter Five, Section 5.1.

Figure 4.2 shows a map of Pompeii, (after Laurence, 1994), illustrating the location of each dyeing workshop and the location of each public fountain. It may be noted that each dye works was within 50 metres of a working public fountain, a fact confirmed during the 2002 survey of the dyeing remains. Furthermore, Laurence is inaccurate in his assertion that each public fountain was at the junction of roads, (Laurence, 1994:44). As may be seen in Figure 4.3 two dye works (Vi4 and Vi5) were actually located adjacent to a public fountain (not shown on the map). These dye works have been discovered to have had plumbing below the floor level, (Borgard et al, 2003), but the purpose of this, whether to supply or drain water, is unknown. It is possible that Laurence only recorded larger fountains, those that comprised of both a fountainhead and a trough.

4.3.4 Vats and steps When examining the dyeing apparatus in Pompeii it was discovered that a number had either a step or rubble that could have been a step attached at the side. There was the assumption made that if a step was present, or appeared to have been part of the original construction, a step was required when using the apparatus. However, when this assumption was tested ergonomically it was discovered that there was no correlation between the presence of a step or the remains of a step and the need for a step. The vats that required a step had a step and vats that did not require steps did not have steps, but there were short vats that did not require steps that had steps. Furthermore, these surplus steps were impractical to use. The presence of an impractical step was not due to the reduction in height of the dyeing apparatus as the original height could be gauged against the adjacent wall. The implication of these structures will be discussed in Chapter Five.

The disposal of water would have been an issue in the dye works as each dyeing apparatus contained between 90 - 900 litres of dye liquor which required disposal at the end of each dye run. It is possible that this was simply disposed of in the street. This is not as farfetched as it sounds. As the method of water supply in Pompeii was an aqueduct, the fountains had each been designed to ‘overflow’ as the water could never be shut off, (Hodge, 1992:335). While this may appear wasteful to modern eyes, it served the purpose of constantly cleaning the streets, (Hodge, 1992:335). The height of the kerbstones and stepping stones indicate the need for a raised walkway as the occasional rainstorms and constant flow of water mean that at times the water may reach 0.5m deep along the roads, (as witnessed by the author during the 2002 survey). If the fountains contributed to this constant flow it would be possible to dispose of dye liquor in the street as it would be washed away.

4.4 Water supply to dye works To understand the source of water used by each dye works it was necessary to examine the supply of water 57

Investigations into the Dyeing Industry in Pompeii

Figure 4.2. Map showing the location of all fountains and dye works in Pompeii, after Laurence, 1994.

Figure 4.3. Fountain outside dye work Vi4. Vi5 is adjacent (to the right). It may be supposed that the dye works Vi4 and Vi5 obtained water from this fountain. This fountain is not shown on Laurence’s map, (Laurence, 1994).

58

Review of Remains in situ

this flow of water. As may be seen by the map in Figure 4.2 and Figures 4.5 and 4.7, properties I viii 19 and VII xiv 17 abutted enclosed backstreets in which flowing water would not have excited comment, Vi4 and Vi5 abutted a road with free-flowing water from a fountain and property VII ii 11 and IX iii 2 abutted a main road. The enclosed backstreets contained a central groove in which water would have collected. It is therefore probably that not only was dye liquor disposed of into the street, but that the waste water created during dyeing was also disposed of in the same way. The water would have been diluted and disposed of through addition to the waste water from the fountains in the main streets. Figure 4.4. Slice cut in doorstep, property I viii 19, allowing drainage into sidestreet. Horizontal measure: 0.2m

4.5 Discrepancy between the digital map and the aerial photograph of Pompeii Following the survey of remains in situ undertaken in 2002 the overall dimensions of each workshop were examined. However, a problem has arisen in this examination. Discrepancies exist between the aerial

Circle highlighting the groove in the step shown in Figure 4.4

Figure 4.5. Photograph showing groove cut into doorstep and the backstreet into which water drained, property I viii 19. Vertical measure: 1m. Horizontal measure: 0.2m.

The dyeing process would have led to excess water leaving the apparatus and wetting the floor. This may have occurred during dyeing, filling of the apparatus, removing the fleece or emptying of the apparatus. There is further evidence that the water simply ran over the floors and out into the street: Figures 4.4, 4.5 and 4.6 show the grooves cut into the doorsteps to allow

Figure 4.6. Grooves left-right cut into doorstep of Vi4 to allow drainage. (Single groove top-bottom was for the edge of the door shutter).

59

Investigations into the Dyeing Industry in Pompeii

Figure 4.7. The door in property VII xiv 17 leading from the dye workshop to an enclosed backstreet. The dye works was contained in a backroom away from the rest of the property and main entrance. Vertical measure: 1m. Horizontal measure: 0.2m.

photograph and the survey (which agree) and the digital map of Pompeii. The digital map was reported to be based on the aerial photograph and so should be the most accurate map to date. All photographic detail should appear on a mapsheet (Burnside, 1979:68; Wilson, 2000:225), as there is sufficient information for this to happen and to not transcribe all of the data would lead to inaccuracy. However, it may be demonstrated that in this map of Pompeii, this is not what has happened.

be undertaken from the air. The extent of remains in Pompeii is advantageous, the relationship between each part of the city and the city itself may be viewed in a broader geographical context. This is further aided by the extent of standing remains: it is possible to discern each feature easily on a photograph and to relate the internal features within buildings to each building’s wider context, (Brophy and Cowley, 2005:123). However, a photograph would still not allow an understanding of the multi-layered structures and any aerial image may become warped during processing, (Burnside, 1979:68).

It is believed that the detailed plans of Pompeii made prior to the digital map were made by a Dutch team c1970s, at least 20 years prior to the 2002 survey. It is believed that this plan was to allow the mapping of all standing remains in Pompeii, although it should be noted that as late as the 1990s it was stated that no complete survey of all the standing structures had been made, (Fulford and Wallace-Hadrill, 1998). This was attributed to the length of time that such a survey would have taken, and so in 250 years of excavation no one had managed to obtain sufficient resources to attempt such a survey. It should be noted that while no complete survey has been undertaken, the majority of excavated Pompeii has been surveyed minutely during each excavation, so there only remains a few areas in need of survey and the tying together of existing plans. Following the advent of aerial photography, satellite and digital mapping it was deemed that a visual geographical survey may now be redundant – a fully comprehensive and detailed survey could

In the light of this new technology and the accuracy of the aerial photograph taken in 2004, it is somewhat surprising that the plans that accompany the photograph are inaccurate. Warping would be expected, but structural differences are surprising. On closer examination it may be seen that the plans drawn in 2004 are closer to the earlier plans drawn in the 1970s than the photograph itself. When examining the dyeing workshops alone it may be seen that there are inaccuracies between the survey results, photograph and the digital map. Below are a section of the aerial photograph and the corresponding part of the plan. Arrow (a) on the photograph denotes a chink in the wall of property VII ii 11. The chink in the wall denotes a door frame or other gap in the wall, albeit not a large one, and so should have been recorded on the plans as such. A door does exist on previous plans, 60

Review of Remains in situ

but on the digital plan of 2004 no door is depicted. However, the plan depicts a door in an adjoining wall (arrow (a) on the digital plan) that, as can be seen on the aerial photograph and was confirmed by archaeologists during the present survey, does not exist. In effect, the cartographer who drew the digital map has moved the door. Arrow (b) on the photograph denotes the room containing the dye vats in property IX iii 2. This room is depicted on both plans previous to 2002 and the plan published in 2004 as containing a quarter-circular construction in front of the vats (arrow (b) on digital plan, below). As may be seen in this photo and was viewed during the survey during the current study and later fieldwork (Robinson, 2003 unpublished), no such construction exists. It should be noted that although it is possible that remains of such a structure could have been found, the archaeologists who visited the property in 2002 and 2003 did not see anything like this structure, (during the present survey of 2002 and later fieldwork, Robinson, 2003 unpublished). The floor of the property was flat and covered with gravel and there was no indication of a structure like the one indicated. Furthermore, as may be seen on the digital plan in Figure 4.8, the vat shapes are approximate. An attempt to use the digital plan to gauge the shape of structures would lead to inaccuracy. For example, vat 9 of property VII ii 11 is shown to have two equally sized spaces for dye kettles, when in fact there was only one space (which was considerably larger than shown). The confusion is that the Fig 4.8. Top: Section of Aerial photograph showing Properties VII ii 11 and XII plan also shows the columns that the iii 2. Bottom: Section of Digital plan showing Properties VII ii 11 and XII iii 2. Photograph: Soprintendenza Archeologia di Pompei. 2004. Unpublished. kettle surrounds are built into, and so to allow a complete kettle space Archaeological remains change over time. They to be shown the cartographer has reduced the kettle degrade, break down, and sometimes they are amended size. While this may allow a closer accuracy in shape, or conserved so that their overall appearance changes. it detracts from accuracy in size, which in this study This has happened in Pompeii. In the plans drawn would lead to misleading findings. Each of the dye prior to 2004 a semi-circular structure may be seen vats was photographed and recorded in situ during the in property XII iii 12. It was not there during the survey of 2002. Separate, and more accurate, recordings 2002 survey and does not appear on the 2004 aerial have been made and may be found in the gazetteer in photograph. This is not to say that it has never existed this volume. 61

Investigations into the Dyeing Industry in Pompeii – it is unlikely that those drawing the plans prior to 2004 invented this structure. What is more likely is that the structure was not large and that over time (since the original plans that included it were drawn) it has broken down and its remains become obscured so that nothing of it remains today. This has resulted in its absence from the records of the 2002 survey and from the 2004 aerial photograph. However, the plans drawn in 2004 feature this structure. This means that they have either copied it from previous plans or invented it. If they had not copied it then evidence of its existence would have been lost. But here they have placed the structure on the plan in such a way as to give the assumption that it is still there. This is what has led to inaccuracies. Furthermore, it is possible that if lost structures are copied from previous plans they may be copied inaccurately. Unless there is some indication of what is currently on the ground, what has been lost and what has been supposed to be there, three very distinct states for a structure to be in, there will be inaccuracies in reconstruction, theoretical or physical.

entirely accurate map and plan of each building already existed. It should be noted that the plans used in this study are based on what was observed and measured in situ during the survey of 2002, as the relation between each dyeing apparatus and its context within the dyeing properties were recorded, so the apparent inaccuracies are not a cause for concern. However, the existence of these discrepancies mean that any use of the digital plan should be made with consideration of the photograph and survey results. 4.6 Bowing

It is hoped that this study will avoid some of these inaccuracies as each vat was measured and recorded in situ. However, while each of the rooms containing a dye vat was measured accurately not all of the rooms within each building were measured as it was believed that an

During the visual examination of the dye vats it was noted that the majority of the intact remaining kettles exhibited significant bowing of their bases. The internal parts of the kettle in vat six in property VII xiv 17 appeared to have been stretched (Figure 4.9) while the base was bowed (Figure 4.10). The kettles in vat five of I viii 19 appeared to have bowed. Vat five of property VII xiv 17 appeared to have bowed and the base contained concentric circular ridges (Figure 4.11). The amount of bowing differed between the kettles and it was noted that the kettle of vat nine in property VII xiv 17 exhibited no bowing at all (Figure 4.12). The significance of this bowing will be discussed in Chapter Seven.

Figure 4.9. Vat six in property VII xiv 17 showing apparent bowing of the base. Vertical measure: 1m. Horizontal measure: 0.2m.

Figure 4.10. Vat six in property VII xiv 17 showing apparent internal stretching. Vertical measure: 1m. Horizontal measure: 0.2m.

62

Review of Remains in situ

contemporaneously. There were assertions that a ladder had been used to reach the upper floor and that the cocciopesto (crushed earthenware shards, in this case from amphorae) had originally contained alum, but evidence to support either of these claims was not presented.

Figure 4.11. Vat five in property VII xiv 17 showing concentric circles on base. Vertical measure: 1m. Horizontal measure: 0.2m.

It was discovered that in property Vi5 the equipment had been constructed contemporaneously. The property had contained plumbing below the floor level, which was strange as there was a fountain that could provide water next to the property. It is possible that the fountain provided a catalyst for the construction of the internal plumbing of the property. It is unknown whether the plumbing acted to supply or dispose of water. There was a rectangular tub in a backroom and an upper floor.

The significance of ‘The House of the Queen of England’, and its absence from Moeller’s original list, has already been discussed in Chapter One. It is possible that this was due to mis-recording or mis-interpretation. It is unfortunate that this has occurred as it is reported that the property contains lead-lined vats which are plumbed in, that the property’s plumbing was constructed to be used in conjunction with the dyeing apparatus, that the dyeing apparatus were designed to overflow and that the overflow both flowed over the floor and was channelled, (Borgard et al, 2003). These assertions are numerous and would require investigation, but Figure 4.12. Vat nine in property VII xiv 17. The base is not bowed. Measure: 0.2m unfortunately the location of the property can not be ascertained from 4.7 Recent excavation the evidence presented by Borgard (2003). Furthermore, there is no other record of a property entitled ‘The House As already stated in Chapter One, properties Vi4 and of the Queen of England’ so it is not possible to infer a Vi5 were excavated in 2000, (Borgard et al, 2003). location from any other source. It was discovered that property Vi4 had originally had an upper floor. The lower floor, the one entered from the street, had originally been divided to maintain the structural integrity of the building following the earthquake of 62 AD. The upper room would have allowed extra storage and ventilation. It was discovered through examination of the foundations that the dyeing equipment had not been constructed

4.8 Further work Following the survey of the remains there is now an accurate recording through measurement, written description and photographs of each apparatus, the context of that apparatus with each other, and the context of the apparatus within the property. This is the most accurate and comprehensive survey to date. 63

Investigations into the Dyeing Industry in Pompeii The knowledge of dyeing gained through the experimental work meant that the remains could be reviewed in the context of the requirements of dyeing, leading to an understanding of the nature of the apparatus and awareness of form. There was also an understanding of how the qualities and nature of the Roman dyer may have influenced the process. However, due to this new understanding, it was noted during the examination of

the standing remains there were structural features, such as steps, that did not appear to assist the dyer in using the apparatus. To fully understand the reason for these features and to determine the usability of the dyeing apparatus, it was decided that an understanding of the ergonomic considerations of the design and use of the dyeing apparatus should be undertaken. This examination is discussed in Chapter Five.

► The gazetteer is replicated here as it was first recorded, faithfully transcribed from the original notes, to maintain the integrity of the data collected at the time. This includes the development of new ideas in ‘further observations’ as the survey progressed. “Gravel fairy” is a shorthand reference to the phenomenon of the over-zealous application of gravel to a property, particularly one in which other artefacts have been stripped out, which has caused the burial of the ground and bases of remaining artefacts, so that it is not possible to accurately record details such as their height. 64

Full Gazetteer of Dyeing Apparatus in Pompeii Property:

I viii 19

Vat:

1 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete? Restored?

86cm 1.15m wall - firebox ?Approx 50cm above bottom edge of lintel 98cm 38cm 42cm 22cm ?unexcavated At least 51cm No No No

Flue? Present? restored? incomplete Absent? ? 3 pipes present, 2 lead, 1 ceramic. Unknown use. Lead to outside. Ceramic and lead to next room. Step? Absent. Metal kettle? Details: Yes, Shattered. It's fallen through. Depth at least 51cm, circumference 1.12m. Possibly 1.63m Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Rubble (stone, brick, tile) 60%. Mortar 40%. Random. Rubble mortared together, then rendered. 10 x 8 x 4cm Smooth, no inclusions. Not obviously as lime. No. 22cm, but wider at wall and ends. Carved to fit across gap, not neatly, to fit. Completely curved/warped. 74cm length, 17cm width, height 24cm 90o from wall N/A

Restoration/repair. Ancient repair?

N/A Doesn't appear to be.

Further Observations:

Map of Pompeii showing location of workshop 1 viii 19 highlighted in red.

65

Plan of I viii 19 showing the location of vat 1 highlighted in red. Plan taken from 2004 digital plan of Pompeii.

Investigations into the Dyeing Industry in Pompeii

Property:

1 viii 19

Vat:

2 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

84cm+ 123cm (brickwall to lintel) vat 54cm ? piD 2.64cm 34cm 38cm 12-36cm Unknown through warping and collapse ? 54cm+ see diagram yes, except for final height

Restored?

None observed

Flue? Present? Restored? incomplete Absent? None seen at vat. Lead pipe? Seen going into next room from centre of vat Step? Yes. X3 Metal kettle? Details: Circumference 2.64cm. However it's warped and top is missing. Illustration Photograph of vat 1 vii 19, vat 2 Construction Material 70% vs 30% mortar. Mostly stone "no bricks". Rubble, rendered. Amphorae shards, approx 20cm2 Brick/Stone % Construct' Technique Steps added later, or at least rendered seperately? Steps obviously rendered as part of vat 3. Vats 1+2 look as if they match, 3 added/at same time as a trio. Look all same. Dimensions 13 x 7 x 10cm Mortar Type Wall uniform width No. Curved sides, straight back and less curved, rendered front. Firebox Lintel Stone, carved rough to fit lintel dimensions ? As one half (approx) is rendered, carries on 7cm after firebox. 16x16x47cm from end of lintel to end of firebox. Firebox Orientation 90o to wall Step Material same - rubble and rendered Restoration/repair. Ancient repair?

? Believe not More additions or detection of order of build vats 1+2 rendered together

Further Observations: pipe is pipe and not flue

Map of Pompeii showing location of workshop 1 viii 19 highlighted in red.

66

Plan of I viii 19 showing the location of vat 2 highlighted in red. Plan taken from 2004 digital plan of Pompeii

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

1 viii 19

Vat:

3 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

99cm 1.14m 63cm 74cm 36cm 40cm 21cm Varies ? 63cm see vat(?) 2 Virtually intact kettle and brazier

Restored? No, but lots of addition Present? Restored? incomplete Absent? No, absent Step? 3x on one side Metal kettle? Details: 74cm wide, 63cm tall. 5cm flange.Has it slipped? Vat flush with inside was it resting on pot shards? Illustration

Flue?

Construction Material Construct' Technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Rubble and mortar Rubble and mortar, then rendered. Square pillars at front edge. 12 x 12 x 13cm

Restoration/repair. Ancient repair?

Not apparent Not apparent

No. Curves at sides, thicker at back. Square at front large stone, carved to shape 20 x 53cm 90o to wall, into room. Rubble and mortar

Further Observations: Doesn't appear to have dropped the kettle: meant to be flush. Vat appears purply, similar to madder staining on lead Flue but no pipe. Firebox observable. Interior is horseshoe shaped. Firebox empty, inc no pillar. Reused carved limestone (?) . Vat 3 and 4 were two sheets welded

Map of Pompeii showing location of workshop 1 viii 19 highlighted in red.

67

Plan of I viii 19 showing the location of vat 3 highlighted in red. Plan taken from 2004 digital plan of Pompeii.

Investigations into the Dyeing Industry in Pompeii

Property:

1 viii 19

Vat:

5 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht

1.14m 1.43m 75cm 97cm 41cm 41cm 20cm 5 pillars of small bricks with gaps between. Built into side. Kettle Depth 75cm Step Dimensions 43cm x 41cm wide x 21cm tall Complete? Virtually. One of the best preserved. Restored? No Flue? Present? Restored? incomplete Absent? Present. Two. One to open air, one above tank. Both have access to open. Step? Present, one to rear has small step next to it. One on front was marble. Front is odd size: doubt an adult could stand and lean over without scaulding. Metal kettle? Details: Present. 5cm flange. Drain. Fits into brazier. Flush with brazier. 75cm tall, 97cm wide. Base is bulging. Whether its meant to is unknown.

Illustration Construction Material 70% stone, 30% mortar. Occasional tiles. Construct' technique Stone and mortar, with lines of tile. Rendered. Built overlapping tank edge. (not all of width) and incorporating pillar. Firebox and base if firebox lined and below floor (slope in). Gritty interior to rendering. Rendered top. Vat fits snug. Dimensions 10 x 10 x 5cm lots of tile and thin bricks. Mortar Type Wall uniform width Yes, as are the supporting pillars. Firebox Lintel 15cm deep, 23cm tall, 54cm wide lintel dimensions Curved block of stone. Smooth, some curvy bits. Firebox Orientation Toward pillar/left edge of door. Doesn't appear to allign with anything. Appears orientated for convenience. Step Material Back: Stone slab. Front: Built up with marble top. Restoration/repair. No Ancient repair? ? Believe not. Further Observations: Built in stages: steps and ledges added last to fit in. Then all rendered. Step built with vat as marble is set into wall. Bulge built after and takes in volume. Drain flows onto bulge. Flange: 7cm. Where wall curves to suppot lintel in a square is used

Map of Pompeii showing location of workshop 1 viii 19 highlighted in red. 68

Plan of I viii 19, vat 5 highlighted in red. Plan taken from Pompeii 2004 digital map.

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

IX iii 2

Vat:

1, 2, 3

Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

1.33m, 1.30m, 1.35cm 21cm, 31cm, 38cm 31cm, 52cm, 58cm 20cm 33cm, 57cm, 48cm Yes

Restored? Yes, mortar is two different colours. Mortar is difficult to see on other vats, but clear here. Flue? Present? Restored? incomplete Absent? Left: Close-up of vat 3 Yes, on all. Step? Yes, 3 steps up side of vat three. Also two Scales: steps set into the front of vats. 1m and 10cm Metal kettle? Details: No. Ridge present. Too tall to . reach to ridge from brazier top. Illustration

Above: Vats 1, 2, 3.

Construction Material Stone. Brick is internal support. Construct' technique Built as a single unit. Rerendered after initial construction, possibly in modern era by someone unaccustomed to vat design and operation. Render present on inside. Not seen render on other vats. Dimensions 10 x 10 x 20cm (ish) Mortar Type Render in the side has no inclusions. Top render on outside has gritty black inclusions and is lighter than bottom render which has white inclusions. Wall uniform width No. Firebox Lintel Stone lintel dimensions 20 x 50 x 17cm, 67 x 19 x 17cm, 67 x 20 x 23cm Firebox Orientation Into room. Step Material Brick for vats 1-2, Brick for vats 2-3. Steps at end of vats are stone. Restoration/repair.

The restoration is modern as had these heights been original the vats would have been unusable.

Ancient repair? Further Observations: Three vats built as a single unit. Can't reach bottom of vat 3 by 22cm if starting at the top step. New render may have been put over old render, not just used to build things above it. Vat 2: flue very likely, but inside bricks have dropped off/been robbed. Render on outside of vat 2, not on others. When all altered details are removed the vats are shorter and usable. Plan of property IX iii 2 showing the vats highlighted in green. They are 1, 2, 3 from top to bottom.

Map of Pompeii, workshop IX iii 2 highlighted in red. 69

Investigations into the Dyeing Industry in Pompeii

Property:

Vi4

Vat:

Vat 4 is on the right, built into wall. Vat 5 is central to the picture Scale is 1m.

4 Max Ext Height 73cm Max Ext Diameter 97cm Max Int Height 73cm Max Int Diameter 97cm Firebox Opening Ht 42cm Firebox Open' Wdth 48cm Wall thickness Varies kettle Support Ht ?Appears to have been stripped Kettle Depth out. Step Dimensions None. Arguable that steps weren't needed although reaching the far side would have been difficult. Room for one in area (a) Complete? Nearly. Heavily restored, but features discernable. Restored? yes,New concrete all over. Flue? Present? Restored? incomplete Absent? Yes, up back wall. Unknown destination once over vat However, traces of old render suggest it continued up Step? Restored, but appears truly absent. Nowhere to put one. possibly (a) on plan. Digital plan and aerial photo, fig 4.8, p.59. Metal kettle? Details: No Illustration

Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

All stone, except for tile lintel Rubble with mortar (inch) made to fit into corner 20 x 13 x ? Cm depth x3 Original, new, concrete over top (new) No. Varies. See diagram Triangular to fit with opening. 7cm thick. 59cm wide Tricky. Wedged in to fit rather than pointing in a particular direction Isn't one

Restoration/repair. Ancient repair?

Yes

Further Observations:

Map of Pompeii showing the location of workshop Vi4, highlighted in red

70

Plan of workshops Vi4 and Vi5. Workshop Vi4 is shown on the left. Vat 4 is highlighted in red.

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

Vi4

Vat:

Vat 5 is central to the picture. The scale is 1m.

5 Max Ext Height 66/61, but lintel juts 6cm into 61 Max Ext Diameter 89cm Max Int Height 77cm Max Int Diameter 66/61, but lintel juts 6cm into 61 Firebox Opening Ht 20cm Firebox Open' Wdth 38cm Wall thickness 13cm kettle Support Ht 32cm Kettle Depth 52+cm Step Dimensions X Complete? Appears to be nearly complete, but it has been restored. Restored? Yes. Don't know if its meant to be attached to next vat Flue? Present? Restored? incomplete Absent? Restoration. All the walls are smooth, but its been restored. Step? Restoration. Wouldn't need one. Metal kettle? Details: No. Remains of shelf 52cm below top. Illustration

Construction Material All stone. Brick and tile shelf. Construct' technique Rubble with mortar. Rems of tile/brick wall for shelf. Unclear if this was free standing. Dimensions 10 x 5 x ? cm Depth Mortar Type 2, poss 3, different sorts. Wall uniform width Yes, but its been restored. Not uniform in bit where it joins vat 5. Firebox Lintel One huge piece of stone, uncarved. lintel dimensions 46cm long, 25cm wide Firebox Orientation Odd, Not in line with vat 5. Step Material X. Step not necessary, and there doesn't appear to be one. Restoration/repair. Ancient repair?

lots. X

Further Observations: Differences in height are more to do with possible growth and gravel shift than Romans.

Map of Pompeii showing the location of workshop Vi4, highlighted in red

71

Plan of workshops Vi4 and Vi5. Workshop Vi4 is shown on the left. Vat 5 is highlighted in red.

Investigations into the Dyeing Industry in Pompeii

Property:

Vi4

Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material Restoration/repair. Ancient repair?

Vat:

"6"

Max Ext Height 105cm Max Ext Diameter 141cm Max Int Height 122 less gunk on the inside Max Int Diameter 120cm Firebox Opening Ht 30cm Firebox Open' Wdth 57cm Wall thickness Not uniform 19cm kettle Support Ht Tricky. 43cm. Halfway up several start looking newer. 23cm definite. Some look OK and are 43 cm. Kettle Depth 80cm Step Dimensions No actual step but there is a heap of step-like rubble. 15cm deep. Complete? Yes, but it is restored and large. Restored? Yes, lots. Very smooth and level. Flue? Present? Restored? incomplete Absent? Restored and incomplete. 3 places where flues could have been. Step? Restored and incomplete. 16cm gap between hand and vat bottom.Need step. Metal kettle? Details: Restored. Ridge 17cm. unrestored 12cm.

Stone rubble and mortar 18 x 12 x ? Cm Depth. 3x. Original, new and concrete finish on top. 12 cm at back, 9cm at front, 19cm on its own, 28cm to vat 7. Huge slab. Flat both sides. Odd shape. 11cm thick 90o into room. ? Rubble with a rock poking out. Yes!

Further Observations: Vat 8 ridge: 8cm out.

Map of Pompeii showing the location of workshop Vi4, highlighted in red.

72

Plan of workshops Vi4 and Vi5. Workshop Vi4 is shown on the left. Vat 6 is highlighted in red.

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

Vi4

Vat:

Vat 7 is central to the photograph. Scale is 1m.

Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

7 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter

108cm 105cm 108cm Due to moss growth 74cm Max is 80cm, but 74cm is vat width possibility? Firebox Opening Ht 29cm Firebox Open' Wdth 30cm Wall thickness 12. Varies kettle Support Ht 31? Definitely 27+ Kettle Depth 78+cm Step Dimensions See diagram Complete? Yes, but heavily restored. Pillar depths are presumed to be accurate, although could be 4cm lower Restored? Yes.

Flue? Present? Restored? incomplete Absent? Restored and incomplete. Line for flue. Stone in upper half. Q is how much of stone is original. Step? Present and restored. Still functional and necessary. Needs a step of 15+cm if full pillar is original. Metal kettle? Details: No Illustration Rubble with mortar. Rubbish and mortar. Large +squarish at front, smaller for interia. 31 x 14 x ? Sq, 10 x 15 x ?depth At least 3 new. Probably the original under there somewhere… No. Very smooth, but changes at each angle/direction. 17cm thick at least. 56cm wide, 46cm tall. 90o into room. Large stone, with rubble, with tile top.

Restoration/repair. Yes Ancient repair? ? It's been restored, but it's not possible to tell when. Further Observations: ? Depth, prob cubed or brick shaped. Rendering with this entire property is unknown, except where stated, due to restoration and zealousness of render. Flue is possible, but unknown if there is one because of restoration.

Map of Pompeii showing the location of workshop Vi4, highlighted in red.

73

Plan of workshops Vi4 and Vi5. Workshop Vi4 is shown on the left. Vat 7 is highlighted in red.

Investigations into the Dyeing Industry in Pompeii

Property:

Vi4

Vat:

8 Max Ext Height 113cm Max Ext Diameter 155cm Max Int Height 121cm Max Int Diameter 120cm Firebox Opening Ht 38cm Firebox Open' Wdth 42cm Wall thickness 14-16cm kettle Support Ht 43cm Kettle Depth c81cm Step Dimensions See Vat 7 Complete? Sort of. Heavily rendered, but features discernable.

Vat 8 is to the right of the photograph.

Restored? Lots. Smooth, blockages.

Scale is 1m.

Flue?

Present? Restored? incomplete Absent? Yes, rendered but workable. Step? yes, functional and necessary. Metal kettle? Details: No Illustration

Construction Material Stone with brick pillars. Construct' technique Rubble with mortar inbetween. Lots of small pieces with lots of mortar. Expected the bits to be larger than this. Dimensions Brick in corner/door filling. 17 x 7 x ?cm depth. Mortar Type 3 Sorts on the pillars alone. Heavily restored, but details still discernable. Wall uniform width No, but it's predictable. Smooth. Firebox Lintel Large reused stone slab. lintel dimensions 17cm thick. 27cm tall, 71cm wide. Firebox Orientation 90o into room. Step Material Rubble with tile top. Restoration/repair. Yes Ancient repair? ? Further Observations: Note - brazier is larger on the outside than the inside. This may be attributed to the gravel fairy. None of the brick pillars on any of the vats are "pillars" They're small brick walls, a combination of two building techniques. Number and location of pillars on sketches is reasonably accurate. Vat 7 pillars at least 7cm wide into vat. Gap between the pillars is difficult to measure as they are broken/rendered. However there are a lot of them and nothing is going to block them. Therefore the updraft ventilation would have been good.

Plan of Vi4 (left) and Vi5 (right). Vat 8 is highlighted in red.

Map of Pompeii, workshop Vi4 highlighted in red. 74

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

Vi5

Vat:

Vat 1 is to the right of the photograph. Firebox is not visible in this photograph. Flue is concealed within corner with door. Scale is 1m.

1 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht

1.11m 1.19m Descends to 1.23m 95cm (98cm to lintel) 39cm 39cm Varies 60 or 67cm 60 definite, 67 mortared Kettle Depth c60cm Step Dimensions No Complete? Nearly. Rendered across ridge. Restored? Heavily restored, but flue and ridge are still apparent Flue? Present? Restored? incomplete Absent? Present, restored, obvious and in the corner. Don't know if it continued up wall . No evidence that it did, and smoke could have driffed out of door. Step? Incomplete, restored. Can reach sides . without one, but not far side of vat. One was needed and there is space for it Metal kettle? Details: No

Illustration Construction Material Rubble. Brick wall for ridge Construct' technique Rubble and mortared. Then restored. Lots of mortar. Unclear how much because it covers the original. Dimensions 11 x 18cm Mortar Type x2 with concrete on top. Wall uniform width No, but it's smooth and predictable Firebox Lintel One large block of stone. Smooth on outside, possible rendering on inside it sticks out. lintel dimensions 59 x 19 cm Firebox Orientation 90o away from street. Parallel to wall.Toward other vats Step Material No step Restoration/repair. Ancient repair?

Yes, lots

Further Observations: 60cm from top of vat to base of render, 53cm to top of render. Possible removal of step. Could measure width of shutter. The significant bit is that the vat is built over it and the length of the walls is known.

Map of Pompeii showing the location of workshop Vi5, highlighted in red. 75

Workshop Vi4 (left) and Vi5 (right). Location of Vat 1 shown in red.

Investigations into the Dyeing Industry in Pompeii

Property:

Vi5

Vat:

Vat 2 is central in this photograph. Scale is 1m.

2 Max Ext Height 99cm Max Ext Diameter Max Int Height ? Max Int Diameter 50cm Firebox Opening Ht 47cm Firebox Open' Wdth 47cm Wall thickness Varies kettle Support Ht Don't know. Shelf sticks out 91cm down. Gives an absolute max of depth. Kettle Depth less than 91cm Step Dimensions Aren't any. Probably not needed. Can reach down to 31cm off floor and 23cm off shelf (it's on the opposite side). Complete? Shaft and firebox. All other details have Restored? been restored. Flue? Present? Restored? incomplete Absent? Not present, possibly due to incompleteness prior to restoration. Step? Not present, possibly due to incompleteness prior to restoration Metal kettle? Details: No Illustration

Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Occassional Brick Rubble, mortar between. 10 x 15 x rendered on inside so can't tell. A lot appears to have inclusions. Fades into each other. No, see diagram. Smooth and Predictable.

Restoration/repair. Ancient repair?

Yes

48cm wide, 27cm tall, 13cm thick 45o into room. Opposite vat 1's firebox. Strange. ? No step.

Further Observations: All rubble in any property is stone unless otherwise stated. The entirity of this property has a swept out feel: no rubble etc. It appears that shabby things were removed and what was restored. The step may have looked shabby.

Map of Pompeii showing the location of workshop Vi5, highlighted in red.

76

Workshop Vi4 (left) and Vi5 (right). Location of vat 2 shown in red.

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

Vi5

Vat:

Vat 3 is central in this photograph. Scale is 1m.

3 Max Ext Height 99cm Max Ext Diameter 1.72m Max Int Height 99cm Max Int Diameter 1.05m Firebox Opening Ht 47cm Firebox Open' Wdth 50cm Wall thickness Varies kettle Support Ht 31cm Kettle Depth 68+cm Step Dimensions See Diagram Complete? Yes and no. Obviously restored appears restorer tried to follow the original plan. Restored? Steps heavily restored Flue?

Present?

Restored? incomplete Absent?

Step? Yes, restored, but looks authentic. Functional. Although getting to the back would be a hassle whatever happened. Metal kettle? Details: No Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material Restoration/repair. Ancient repair?

All. One small piece of brick. Rubble with mortar (inch). Large bricks even on inside. 12 x 25 cm x3 Evidence of Roman render on inside. No. Smooth and predictable. Big bit of stone. Carved vaguely to shape. Rough front and back, uneven all other dimensions. 65cm wide, 26cm tall, 19cm thick. Strange. Not actually in any direction. 10o off 90o from wall. Rubble with neat 2 layers of tile on each. Yes. A lot.

Further Observations: Something pipelike sticking out of back of vat. Either pipe or odd masonry. See Photograph

Map of Pompeii showing the location of workshop Vi5, highlighted in red.

77

Workshop Vi4 (left) and Vi5 (right). Location of vat 3 shown in red.

Investigations into the Dyeing Industry in Pompeii

Property:

Vi5

Vat:

4 Max Ext Height 1.02m Max Ext Diameter 1.09m Max Int Height 1.07m Max Int Diameter 85cm x 85cm Firebox Opening Ht 37cm Firebox Open' Wdth 34cm Wall thickness varies kettle Support Ht 38cm Kettle Depth 69+cm Step Dimensions See Vat 3 Complete? Sort of. Heavily restored, with flue in place. Restored? Yes. The impression is that everything is mortar lined. It's all smooth. The back is brickwall lined. It's unusual as the rest isn't (except in patches) and no other vat has brick except for the entirely brick one Flue? Present? Restored? incomplete Absent? Yes, Potential for 2nd flue (a line of mortar up the wall). Step? Yes, but heavily restored. Still functions, although not always necessary. Metal kettle? Details: No

Illustration Construction Material Heavily restored. Stone original. 50%, mainly at the back, but the top layers everywhere are brick. Can still reach ridges, it appears restoration may be accurate. Construct' technique Rubble with brick plinths for vat support. Restored. Dimensions 10 x 6 x ? Cm depth. Mortar Type A lot of different types from different eras. Restoration (1920s), Roman, etc. Wall uniform width No, but its smooth and predictable. Only ununiform so it'll fit into wall. Firebox Lintel Huge rock originally. When bottom dropped off it got replaced with 2 steel/iron rods and lots of render. lintel dimensions 10 x 62cm a possibility. Entirely rendered. On inside its 53cm wide 36cm tall Firebox Orientation 90o into room. Step Material See Vat 3 Restoration/repair. Lots. Rendered everywhere. Ancient repair? Further Observations: Were there 2 flues? Evidence of rendering in a line at the side on this one (and the bac back being rebuilt with brick) but there's already a preserved (and restored) flue at the back. Restoration includes iron bars across lintel (one front,one back). Restoration of 3, 4 and 1 appears to have been done by someone who had taken note of what was there and allowed for the possibility of the holes being part of the operating method, rather than someone who smooth everything down and had a clear-out to make it look pretty.

Workshop Vi4 (left) and Vi5 (right). Location of vat 4 shown in red.

Map of Pompeii, workshop Vi5 shown in red. 78

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII ii 11

Vat:

"1"

Max Ext Height 1.08m Max Ext Diameter 1.48m Max Int Height ? Ivy covered Max Int Diameter c1.06m Firebox Opening Ht 39cm Firebox Open' Wdth 43cm Wall thickness 25cm kettle Support Ht ? Ivy covered Kettle Depth ? Height presumed from lintel Step Dimensions See above Complete? Nearly, Shape and size discernable Restored? Not obvious. Believed not.

Flue? Present? Restored? incomplete Absent? Not present. Truly absent. Step? Yes, eroded, at least 3 steps. Flight between vats 1 and 2 to back wall. Metal kettle? Details: None present. Impossible to investigate fully due to aggressive foliage. Illustration Construction Material Stone with mortar. 80% stone, 20% mortar. Substantial rubble. Construct' technique Rubble with mortar. No obvious bricks. Tiles on the steps (capping). Dimensions 12 x 20cm. Depth into wall unknown. Mortar Type ? Grit inclusions Wall uniform width Appears to be. Firebox Lintel One stone, carved and supported by two stones with mortar lintel dimensions 60cm wide, 20cm tall, 16cm thick 90o to wall. Into room. Firebox Orientation Step Material Rubble with mortar. Capped with tile. Restoration/repair. No Ancient repair? No Further Observations: All of the vats were examined once the ivy had been cleared from the walls. The ivy was a particularly aggressive variety with barbs. The length of time it had been in place meant that it was now an integal part of the structure and that any attempt to remove it would have resulted in greater damage than leaving it in place. Therefore details of these vats were not entirely discernable. Vats 1 and 2 are lopsided. They have a smooth curve one side round to steps and a definite corner on the other. Vats 1 and 2 appear to be a matching pair. The whole development is post-earthquake - the house is too high status and the vats too awkwardly situated for the vats to have been original. There is mortar on the lintel which suggests rendering and the firebox was originally at a different height. The wall appears to be of a uniform width, but if is difficult to be exact.

Map of Pompeii showing the location of workshop VII ii 11, highlighted in red.

Plan of workshop VII ii 11, showing the location of vat 1, highlighted in red. 79

Investigations into the Dyeing Industry in Pompeii

Property:

VII ii 11

Vat:

"2"

Max Ext Height 1.14m Max Ext Diameter 1.43m Max Int Height ? Max Int Diameter 1.06m Firebox Opening Ht 37cm Firebox Open' Wdth 55cm Wall thickness 17-22-26cm kettle Support Ht ? Kettle Depth ? Step Dimensions See vat 1 Complete? No, but dimensions are noticable Restored? None observed Flue?

Present? Restored? incomplete Absent? No, truly absent. Step? Flight of 3 between vats 1 and 2. Metal kettle? Details: Not present. Illustration

Construction Material Stone. No bricks 90%. 10% mortar. Construct' technique Rubble with lots of mortar. Evidence of render stuck to front. Dimensions 20 x 10cm. Vats 1 and 2 depths unknown, but doesn't go through whole of wall depth. Mortar Type Gritty inclusions. Used for rendering. Wall uniform width No, but variation is slight. 17-22cm. Thick at back wall. The width of the side would depend on what has dropped off. Firebox Lintel Large stone. Carved, ridge on it. Reused? Crack in centre. lintel dimensions At least 8cm thick. Full measurement blocked by rubble and vegetation. Cm 37 tall, 82cm long. Firebox Orientation 90o to back wall, straight into room. Step Material Rubble with tile capping. Restoration/repair. Not apparent Ancient repair? Not apparent Further Observations: Vat 1 is built into wall, Vat 2 has its own wall. (Vat 1 saves on materials?) All of the other vats are built against the wall but have their own back walls. Vat 1 is strange.

Map of Pompeii showing the location of workshop VII ii 11, highlighted in red.

Plan of workshop VII ii 11, showing the location of vat 2, highlighted in red.

80

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII ii 11

Vat:

"3"

Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete? Restored?

78cm 31+30+88cm = 1.69m ? Impossible to see ridge brasure full of rubble 96cm 50cm 40cm 19-22-58cm. Uniform except for corner. ? At least 60cm down to the rubble 16cm height Virtually ?

Flue?

Present? Restored? incomplete Absent? Not present, Absent. Step? Unknown, full unexcavated Metal kettle? Details: No Illustration

Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Stone. Inch of mortar between each one. Rubble and mortar. No apparent render. 10 x 15 x 8cm Grey, no inclusions No One long stone. Looks like offcut from something else. 53cm long, 28cm thick, 9cm deep. 60o Not facing into room. Is facing out of corner. Looks like its meant to line up with Vats 5 and 4 as it turns the corner, but it doesn't. Rubble.

Restoration/repair. Ancient repair?

? ?

Further Observations: Squidged in corner to match Vats 5 and 4, fit in last minute. Hole is not circular, vat in it larger than 5 and 4, but there's still room. Vat 4 has tile in to make up layers. Step is part of vat 4 and abuts vat 3. Vat 3 looks like it cut through the step. It's an awkward shape. Can't have been a flush fitting vat.

Map of Pompeii showing the lcaotion of workshop VII ii 11, highlighted in red.

Plan of workshop VII ii 11, showing the location of vat 3, highlighted in red. 81

Investigations into the Dyeing Industry in Pompeii

Property:

VII ii 11

Vat:

"4"

Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete? Nearly

82cm 1.11m 56cm 76cm 40cm 37cm 17cm ? Base is full of rubble. 56? ?

Restored? Concrete on top. Unknown vintage. Doesn't match. Flue? Present? Restored? incomplete Absent? None observed. Truly absent. Step? One each side. Purpose made step on Vat 5 and 4. Squidged step between vats 4 and 3. Metal kettle? Details: No Illustration Construction Material Stone - large pieces. 90% stone. Mortar sandwiched between. (Other vats are mortar with stone in). Construct' technique No evidence of render ? Dimensions Rubble. Stones don't go all the way through. Mortar Type Sparingly used. Wall uniform width Yes, on three sides. Back is infilled with rubble to meet wall. Firebox Lintel Solid stone. Possibly cut to shape, but not carefully. Flat bottom, uneven sides, smooth front. lintel dimensions 16cm thick, 60cm wide, 42cm tall. Firebox Orientation 90o into room. Step Material Rubble. Same construction. Restoration/repair. Ancient repair?

? There's concrete round the top with different inclusions. ? There's concrete round the top with different inclusions.

Further Observations: Middle of 3 "matching" vats. Skewed to form the beginning of curve. "Fitted kitchen" look with vats fitted against wall.

Map of Pompeii showing the location of workshop VII ii 11, highlighted in red.

Plan of workshop VII ii 11, showing the location of vat 4, highlighted in red. 82

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII ii 11

Vat:

5 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

80cm 90cm 50cm to inner surface 55cm to outer. 37cm 37cm 20cm ?Appears to be absent ? 37cm Appears to be nearly complete

Restored? No. However structure was obscured in ivy (now cleared). Flue? Present? Restored? incomplete Absent? None observed. Step? Yes. Flush with vats 4 and 5. Wide enough. Metal kettle? Details: No support. Ivy and rubble removed, interior lacks evidence. Illustration Construction Material Construct' technique Tightly packed rubble, mortar between. Virtually a cube. Evidence for rendering. Unknown if lintel was rendered. Dimensions 9 x 13 x 9cm Mortar Type Black grit inclusions. Wall uniform width Yes. Firebox Lintel Possibly carved to size. Possibly reused. Very neat. lintel dimensions 60 x 28 x 11cm thick. Thicker on inside. Flat on outside with curve at base. Firebox Orientation 90o from wall, straight into room. Step Material Rubble from mortar. Covered in mud and ivy. Unclear view. Restoration/repair. Ancient repair?

Patch of mortar of unknown vintage. Mortar around the top looks different to the rest.

Further Observations: Vats 1 and 2 meant to be rounded. These are designed to be square. 1 and 2 are a matching set. 3,4 and 5 are also a matching set. No flue present. This vat was the reconstructed one.

Map of Pompeii showing the location of workshop VII ii 11, highlighted in red.

Plan of workshop VII ii 11, showing the location of vat 5, highlighted in red.

83

Investigations into the Dyeing Industry in Pompeii

Property:

VII ii 11

Vat:

"6"

Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete? Very nearly

Scales: 1m and 10cm

81cm 1.15m 54cm 86cm 25cm 42cm 21cm ? 54+cm 16cm tall, 26cm wide, 1.15m from wall.

Restored? ?Covered in ivy. No obvious evidence when cleared. Flue? Present? Restored? incomplete Absent? Not present, truly absent Step? present, unexcavated, bad repair probable. Metal kettle? Details: No Illustration

Construction Material 95% Stone. 5% mortar. Neat small blocks of rubble on inside. Stone with very occasional brick. (Only one on front wall). Construct' technique Rubble with thin amount of mortar sandwiching. Large rocks on outside, small inside. Ledge for vat built up with bricks Dimensions 31 x 9 x ? cm Doesn't go through the internal wall. Mortar Type No inclusions. Very light grey. Wall uniform width Apparently so, (except at corners). Firebox Lintel One large oddly shaped stone. Cracks in it. Just wider than firebox opening. lintel dimensions 24cm thick. 54cm long. 37cm tall. 30cm thick at top on inside - juts out. Firebox Orientation 90o to wall into room. Step Material ? Overgrown and unexcavated. Restoration/repair. ? Ancient repair? ? Further Observations: The ledge for the vat was built up with bricks like the original reconstruction design. Are all of the large and nice pieces of stone what is left over from the earthquake? Lots of buildings became dye-works after the earthquake. People leaving? Or building becomes unliveable to some and people move out? Whatever is constructed from the rubble is always smooth on the inside (laid so the squares turn to form a curve). The brick supports are not in pillars - they are one continual curve supporting to support the vat.

Plan of workshop VII ii 11, showing the location of vat 6, highlighted in red.

Map of Pompeii, workshop VII ii 11 shown in red. 84

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII ii 11

Vat:

"7"

Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete? Virtually

77cm 85cm 55cm 55cm 31cm 36cm 22cm Approx 21cm 55cm (ish) See vat "6"

Restored? No

Scales: 1m and 10cm

Present? Restored? incomplete Absent? No. Truly absent. Step? Yes. See vat "6". Dishevelled. Metal kettle? Details: No Flue?

Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Stone - Mortar on inside to cause curve. Inch between on outside. Rubble and mortar. No render evidence. 10cm3 or inside 19 x 10 x 11 cm. Mid grey. No Big stone block. 31cm tall, 48cm wide, 18cm thick. 90o into the room. ?

Restoration/repair. Ancient repair?

? ?

Further Observations:

Map of Pompeii showing the location of workshop VII ii 11, highlighted in red.

Plan of workshop VII ii 11, showing the location of vat 7, highlighted in red.

85

Investigations into the Dyeing Industry in Pompeii

Property:

VII ii 11

Scales: 1m and 10cm

Vat:

"8"

Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness

69cm 1.30m 38cm 99cm 26cm 43cm 22 at back, 23 at bulge, Rest is c16cm. kettle Support Ht ? Contains rubble. 35+cm. Kettle Depth At least 38cm. Step Dimensions 67 x 25 x 27cm tall Complete? Nearly. Appears top is missing to unknown level. Restored? Not apparent Present? Restored? incomplete Absent? No absent Step? Yes. X2. One large step between Vats "8" and "9" with rubble/mortar back. One stone step. Possibly unitentional. Metal kettle? Details: Illustration Flue?

Construction Material Brick with a stone lintel and stone lintel supports. Corner by pillar has stone. Construct' technique 8cm Vat ridge. Brick ridge - uniform. Brick outer ring, mostly uniform, stone in it on pillar side and toward pillar. Dimensions 3cm wide. Various lengths. 9cm on inside. 13cm outside 7cm wide. Mortar Type Light grey. Limited inclusions. Packing in bulge unknown. Wall uniform width No. Bulge in front by built step. Back corner built up toward pillar (in a point). (Rubble). Firebox Lintel One large flat stone. lintel dimensions 14cm thick. 64cm wide. Firebox Orientation 90o into room. Step Material Stone for one step. Bricks for other. Back of other is rubble. Looks like back was added as a brace support. Restoration/repair. No. Not obviously. There's some concrete/asphalt like stuff on the top of the Ancient repair? bricks. It appears this isn't original. Looks like soil baked hard. Bricks survive higher at the back therefore front isn't original height. Further Observations: It's an art work! The whole vat is constructed from little slim bricks.

Map of Pompeii showing the location of workshop VII ii 11, highlighetd in red.

Plan of workshop VII ii 11, showing the location of vat 8, highlighted in red.

86

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII ii 11

Vat:

Scales: 1m and 10cm

"9"

Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht

73cm 1.45m c55cm 94cm 41cm 59cm 30cm c20cm, at least 18cm. Unexcavated middle. Kettle Depth 55+cm Step Dimensions See Vat "8" Complete? No, lintel is highest point, but it isn't the true height. Restored? Not apparently Present? Restored? incomplete Absent? No. Truly absent. Step? Yes. One between vat 8 and 9. Flush to both. Metal kettle? Details: No Illustration

Flue?

Construction Material Inside: brick 3/4 inch mortar between. Outside: 95% squarish stone rubble, 5% mortar. Construct' technique Stone rubble with mortar. Flat front. Brick lined (as in a brick inner wall). Brick supporting ridge. Dimensions 6 x 20 x 17cm is stone. 10cm long ish. Brick 3.5 or 35? Cm thick. Varying lengths and thicknesses. Brick and tile. Inside wall has random shapes. Mortar Type Very light grey, very tiny grit inclusions. Nearly. As close as probably makes no odds. Wall uniform width Firebox Lintel Large stone. Lotsa holes in it. lintel dimensions 29cm thick. 76cm wide. Firebox Orientation 90o into room. (Away from pillar). Step Material See vat "8". Rubble and mortar. Restoration/repair. Ancient repair?

? ?

Further Observations: Vats 6, 7 and 9 have flat fronts.

Map of Pompeii showing the location of workshop VII ii 11, highlighted in red.

Plan of workshop VII ii 11, showing the location of vat 9, highlighted in red.

87

Investigations into the Dyeing Industry in Pompeii

Property:

VII xiv 17 Vat:

1 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

95cm 85cm 68cm 35cm 35cm 28-16cm 38cm 50cm Yes

Restored? Present? Restored? incomplete Absent? Present? Step? At front. 10cm above current floor. Metal kettle? Details: No

Flue?

Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Stone brick - tile used to level lintel and for kettle support. Stone mortar - a number of well squared off block.

Restoration/repair. Ancient repair?

Not apparent

No Limestone 42 x 18 cm Into room Black and white specked metamorphic rock.

Further Observations: Evidence of Opus Signinum render on wall and outside surface of vat. Flue probably shared with vat 10 on other side of wall.

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 1 highlighted in red, tables and tanks shown in blue. 88

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII xiv 17 Vat:

5 (new 2) Max Ext Height 1.11m Max Ext Diameter 1.38m Max Int Height 1.10m Max Int Diameter 1.17m Firebox Opening Ht 46cm Firebox Open' Wdth 40cm Wall thickness 18.12.9 kettle Support Ht Kettle Depth 65+cm Step Dimensions ?But it's possible to reach base. Complete? Sort of. Complete enough for measurements, but obviously "restored". Restored? Abundant evidence. Difficult to distinguish between original and recent mortar. Neat around the top. Flue? Present? Restored? incomplete Absent? Yes. Strange design. It feeds up from the fire but stops at the vat. Is it a pipe down? Step? Possible. Rubble from back wall looks promising. Metal kettle? Details: No Illustration

Construction Material Stone. Brick wall for ridge. Construct' technique Stone rubble with mortar inch. 2 stone supports for stone lintel. Ridge is brick wall built up. Dimensions Stone 9 x 12 x ? cm Depth. Mortar Type No inclusions. Light grey. Wall uniform width Sort of. Change is smooth and predictable. Firebox Lintel Big rock slab. lintel dimensions Firebox Orientation Step Material Restoration/repair. Ancient repair? Further Observations:

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 2 highlighted in red, tables and tanks shown in blue.

This was part of an argument with the classicists who felt that all the kettles were cemented into place 89 their reconstructions into place, but the warping because this one was, with Borgard’s team cementing showed it was modern cement. I’ve published a fuller version elsewhere.

Investigations into the Dyeing Industry in Pompeii

Property:

VII xiv 17 Vat:

3 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

1.22m ?1.18m 1.22m 1m 49+cm 40cm 9cm 42cm 76cm ? Front missing

Restored? Yes. Sone stones have been repaired/faced. Present? Restored? incomplete Absent? Present, broken. Step? ?Pile of rubble, no step evidence at back. Metal kettle? Details: No

Flue?

Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material Restoration/repair. Ancient repair?

Stone with brick shelf (wall up to it). Rubble and mortar. Brick shelf built up seperately. Gap left for flue. 10 x 10 x 15cm stone Light mortar for patching. Patched over by darker. Original is lighter still, with white pebble inclusions. Yes, but little remains. 9cm None. It has been robbed. ? At least 40cm wide or it'll drop off. 90o into room. ? Yes. Flue blocked. Tiles rendered?

Further Observations: The design of the flue may be decerned from vat 4. Obvious that the flues have been blocked by modern restoration. All brick over entire vat is 3.5cm height x various.

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 3 highlighted in red, tables and tanks shown in blue. 90

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII xiv 17 Vat:

4 (new 7) Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

1.14m 1.19cm 1.19m 99cm 43cm 46cm 8cm 43cm ?68cm ? Chimney, yes. Front, no.

Restored? Yes. Shelf and mortar differing widths. Present? Restored? incomplete Absent? Yes. Very informative. Step? ?Rubble at back with marble slab on (small) Metal kettle? Details: No

Flue?

Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Stone, brick shelf. Rubble and mortar. Heavy restoration. Stone: 10 x 12cm. Tile/brick: various 3.5cm x 19cm 3x No. Original width unknown. Appears uniform, but isn't. None remaining 52+cm oterwise it would drop off. 90o into middle of room. ?

Restoration/repair. Ancient repair?

Yes. 3x mortar. Wall in wrong place. ?

Further Observations: Brick shelves were all built from the ground.

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 4 highlighted in red, tables and tanks shown in blue.

91

Investigations into the Dyeing Industry in Pompeii

Property:

VII xiv 17 Vat:

5 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

86cm 89cm 43+cm 60cm 42cm 38cm 43+cm ? Nearly. So is the vat.

Restored? Yes, neat top, 3x mortar. Present? Restored? incomplete Absent? Yes, under bucket Step? ? Metal kettle? Details: Well preserved. 90o bucket shape. Not complete, but circular. 60cm wide. Illustration

Flue?

Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material Restoration/repair. Ancient repair?

Stone. Brick shelf Rubble blocks with mortar. Brick shelf. 3x. With patching and smoothing layer on top. Yes, but restored. One large (cracked) stone, with mortar splodge (unknown age), cracks unrepaired. 51 x 26 x 11cm 90o to back wall into middle of room. ? Pile of rubble at back. Yes. Walls smooth, straight and too low. Vat cemented into place in its warped state. ?

Further Observations: Vat is flush to side at one side and 9cm away at other therefore actual gap unknown.

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 5 highlighted in red, tables and tanks shown in blue.

The warping in the metal kettle occurred before it was cemented into place. Warping occurred during the eruption of Vesuvius. Cementing had occurred during modern restoration. The kettles had not been cemented when in use in antiquity. Figure 4.11 shows this more clearly and the author has explored this in further publications. Borgard’s team cemented their reconstructed kettle into place, causing an inaccurate understanding of how the dyeing apparatus operates.

92

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII xiv 17 Vat:

6 Max Ext Height 88cm Max Ext Diameter 1.67m Max Int Height ? Max Int Diameter 1.09m Firebox Opening Ht 40cm Firebox Open' Wdth 40cm Wall thickness 47cm at entrance kettle Support Ht ?Tends to sit on the floor Kettle Depth Geometry used to determine this Step Dimensions ?Rubble at back shows possibility. Must be a step cos of height. Complete? Very nearly, the only missing part is flange. Also firebox is full of debris. Restored? Yes. Vat is cemented into place in parts that are missing. Flue? Present? Restored? incomplete Absent? Step?

?Incompleteness. Debris everywhere. Step looks likely because of depth. Metal kettle? Details: Odd shape. Conical. 6cm flange. Flat bottom. 1.09m deep not including flange. Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Brick front. Stone everything else. Rubble walls. Brick entrance way and front. 9 x 9 x depth cm = stone 4x Not apparently Two. Stone. Front cracked, back has crack and given way. Together 43cm thick. Back one approx 17cm. 49cm wide. 8cm deep. 45o out of the corner of the room. Fits between other vats. ?

Restoration/repair. Ancient repair?

Yes No

Further Observations: Brick front has new mortar - but is it new? There are two lintels, one in front of the other. The front one is cracked, but the back one has collapsed. It is difficult to discern inside the firebox is filled with debris. the stones are of unknown depth as unless it continues through the wall it is impossible to determine its size.

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 6 highlighted in red, tables and tanks shown in blue.

93

Investigations into the Dyeing Industry in Pompeii

Property:

VII xiv 17 Vat:

7 Max Ext Height 51cm Max Ext Diameter 1.15m Max Int Height 51+cm Max Int Diameter 68cm Firebox Opening Ht ? Firebox Open' Wdth 43+cm Wall thickness 22cm in total ?wall kettle Support Ht 43cm Kettle Depth ? Step Dimensions ? Complete? Steps at ridge, the rest has been rendered.

Restored? yes. Evidence of wall has been removed. Vat 7 is the horse-shoe shaped structure to the right. It appears that vat 7 was Flue? Present? Restored? incomplete Absent? originally of the same design as vat 8 Yes at base. Blocked at shelf height. (to the left), but after robbing and restoration Step? Restoration and incompleteness. has been left in this state. Metal kettle? Details: No Scales: 1m and 10cm

Illustration

Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Brick shelf. Stone everything else. Lot of modern mortar. Stone with brick wall shelf. Extensive restoration. Unknown due to render. 10 x 12 x ?cm. Brick 3.5cm x various. x2. Old and new. ?currently is.

Restoration/repair. Ancient repair?

Extensive Unknown. Believed not.

At least 38cm width or it would have dropped off. 90o more or less, slightly off from into the centre of the room. ? Built up dirt at back of vat.

Further Observations:

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 7 highlighted in red, tables and tanks shown in blue.

94

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII xiv 17 Vat:

Vat 8 is the structure on the left. It appears that vats 7 and 8 were built to the same design, possibly as a pair.

8 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

1m 1.2m 1.07m 95cm 53cm 44cm 14-8cm 39cm 63cm N/A Yes

Restored? Yes Present? Restored? incomplete Absent? Yes Step? Truly absent. Metal kettle? Details: No

Flue?

Illustration Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Rubble with mortar. Rubble and mortar -brick for shelf - brick for lintel support Stone, brick shelf Black and white speckled No Stone with black specks. 27cm tall, 55cm wide. Middle of the room. N/A

Restoration/repair. Ancient repair?

Yes, render, cement capping (?)

Further Observations: Exit to flue blocked by restoration to vat 9. This pair of vats 8/9 is similar to vats 4/5.

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 8 highlighted in red, tables and tanks shown in blue.

95

Investigations into the Dyeing Industry in Pompeii

Property:

VII xiv 17 Vat:

9 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete?

82cm 1.03m front - back ? 51cm 51cm 32cm 22-38cm 43cm 43cm N/A Yes

Restored? Some render Present? Restored? incomplete Absent? Not present due to restoration. Step? Truly absent Metal kettle? Details: Yes. Largely complete. 4cm lip. Lead. Some red staining. Illustration

Flue?

Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Bulk is rubble, rubble mortar, tile let into top surface, largely whole. Rubble and stone block in mortar

Restoration/repair. Ancient repair?

Some.

Largely covered by render No Limestone 50cm wide, 14cm tall Away from back wall - see diagram N/A

Further Observations: Complete. Bottom of kettle on level with bottom of lintel. It is not possible to identify a flue but it is possible that it had a joint flue and chimney with vat 8 in a similar 4/5.

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 9 highlighted in red, tables and tanks shown in blue.

96

Full Gazetteer of Dyeing Apparatus in Pompeii

Property:

VII xiv 17 Vat:

10 Max Ext Height Max Ext Diameter Max Int Height Max Int Diameter Firebox Opening Ht Firebox Open' Wdth Wall thickness kettle Support Ht Kettle Depth Step Dimensions Complete? Restored?

Scales: 1m and 10cm

1.06m 45cm 1.10m 94cm 55cm 48cm 23-30cm 50cm 60cm 50cm tall(?) Yes Yes, some render.

Flue? Present? Restored? incomplete Absent? Yes. Exit through wall. Unclear - vat backs on to vat 1. Present, brick and tile. Step? Metal kettle? Details: No Illustration

Construction Material Construct' technique Dimensions Mortar Type Wall uniform width Firebox Lintel lintel dimensions Firebox Orientation Step Material

Rubble and mortar construction, brick for steps and kettle support.

Restoration/repair. Ancient repair?

Some repointing.

No. Metamorphic (?) stone. 70-30cm Into centre of room. Brick

Further Observations:

Map of Pompeii showing the location of workshop VII xiv 17, highlighted in red.

Plan of VII xiv 17, vat 10 highlighted in red, tables and tanks shown in blue.

97

Chapter Five

Application of Ergonomics to Apparatus and Skeletal data loose fibres these would need to be retrieved from the emptied vat. There would be a need to reach the bottom of the vat for the cleaning out the final excess water. Following scouring it was necessary to scrub the grease from the sides of the vat. Following all of these practical considerations it was deduced that a dyer must be able to reach the bottom of each vat for it to be physically possible for each vat to be used.

5.1 Ergonomics of a dyeing apparatus There are two factors that determine how an apparatus may be operated. One is the apparatus itself, the design and the processes that take place within it. The other is its design in relation to the operator – how the processes are affected by the operator’s ability to use the apparatus. The speed and accuracy of a process may be affected by the operator’s experience and ability but a further factor is the size and reach of the operator, (Galer, 1987). Constructing a full-scale replica allowed the design of the dyeing apparatus to be understood and the physical limitations of the process determined. Furthermore an ergonomic assessment of the dyeing apparatus could only be undertaken following such a reconstruction.

It was demonstrated that it was possible for the modern operatives to reach the base of the replica apparatus and so it was possible to argue that a Roman operative would also have been capable of cleaning the base of the vat. However, one criticism made during the work already presented by Hopkins (2002) and Hopkins et. al. (2005) was that the modern individuals who were building and operating the dye vat were not representative of the Roman individuals who had used the dyeing apparatus in antiquity, (Pers. Comm. Jones, 2002). The problem is that a modern person will differ in size, build and abilities from the Roman operator of the apparatus 2000 years ago through natural drift over time. This means that it is possible that the modern person may be larger or smaller than the Roman and so incorrect assumptions about the apparatus may be made, based on size. It was assumed that the modern individuals were taller and better nourished than the original dye workers, (Pers. Comm. Jones, 2002). The slave Roman may have had the genotype to allow a comparable height but had not had the diet to achieve it, or the genotype of the population as a whole may only allow a shorter final height. If it is to be supposed that the persons working on the vat had to reach the bottom, it is obvious that if modern persons were taller they would be able to reach further than the Romans. A further point may be added to this in that the operatives of the modern replica would have had more sedentary lifestyles and would not be accustomed to the physical process of dyeing. It was supposed that it was not possible to make any physical assumptions about the Roman dyeing apparatus through the use of a modern replica as their foundation would be invalid, (Pers. Comm. Jones, 2002). Until the actual height of ‘a Roman’ was understood, this would give a false understanding as to limitations of the vats as it would be assumed that vats that were in fact unworkable were still in the limits of possible. It has been seen that many of the vats in Pompeii have been ‘restored’ and are now taller than they would originally have been. This has given a false result in the kettle size for these vats and a greater output of finished wool. However it had not

Whether an apparatus could have actually been used by someone of average Roman height would help determine how an apparatus could be used and whether the apparatus in question had been altered. To determine this it was necessary to understand the average height and build of a Roman dyer. It was then necessary to determine a method by which the height, build and ergonomic abilities could be applied to the replica apparatus through the use of a modern equivalent. 5.2 The height of the average Roman Following the initial operation of the dyeing apparatus it is believed that it was cleaned following every dyerun in situ, as it is not possible to lift it either to empty or to clean it. The Roman lead kettle would have been heavier than the stainless steel kettle used in the experiments. The replica kettle weighed 10kg, the lead kettle would have weighed 36kg, if the dimensions were the same other than the thickness (2mm for stainless steel, 5mm for lead) and the relative densities were 7800 kg/m3 and 13,100 kg/m3, (Cengel and Boles, 1998; Callister, 2000). The lead kettles would have been too heavy to lift, especially when containing water. Some kettles were plumbed in and so impossible to lift. Furthermore, the kettle was constructed from lead, a malleable metal, and lifting would have shortened the lifespan. Therefore it was concluded that if the vat could not be lifted out to be cleaned, then the cleaning implement would have to have reached the bottom of the vat. As dyestuffs stain the surface of the dye vessel, this would need to be scrubbed by hand (not just with a lengthy cleaning implement). If the wool had lost 98

Application of Ergonomics to Apparatus and Skeletal data

until now been possible to assess which vats were too tall to be used and how much extra height had been added. It was therefore necessary to assess how much (if at all) taller the modern person was than the average Roman, and whether the average Roman could in fact have reached the bottom of each vat. The inability to apply findings gained through the use of a modern replica by a modern population would have meant invalidating any findings about the operation of the original dyeing apparatus. It was therefore necessary to find a method by which to validate the use of modern individuals. This may be done by determining a link between the Roman and modern data through which it is possible to prove the abilities of either – if it may be demonstrated that one group are capable of something, it can be assumed that the other group is too.

5.4 Skeletal evidence from Herculaneum The skeletal evidence from Herculaneum has survived both in greater quantity and quality than the skeletal evidence of Pompeii. Therefore Herculaneum will be used to provide evidence. The skeletal evidence from Herculaneum has been examined. 162 individual skeletons were discovered, of which 96 were recognisable as individuals and studied. Of these, 55 were male and 41 were female, (Capasso, 2001: 923). In studying the heights only adult skeletons were examined, (Capasso, 2001: 923-4). Heights were divided by sex of the skeletons. The maximum and minimum heights for each sex were determined as was the median height over all, (Capasso, 2001: 926). It may be seen that the maximum male height was 174.9cm (5ft8.8) while the minimum was 148.5cm (4ft10.5). The maximum female height was 159.5cm (5ft2) while the minimum female height was 140.0cm (4ft7). However, the average heights allow a greater comparison with other data sets. The average height of a Herculaneum male was 163.8cm, while in the rest of Campagnia it was 164.0cm. The average height of a Herculaneum woman was 151.7cm, while in the rest of Campagnia it was 152.6cm.

5.3 Ergonomics It should be noted that it may be assumed that the dyeing apparatus are of a size that allows operation by individuals of Roman size. If the dyeing apparatus could not have been operated by these individuals then it would have been amended to allow this. The final amendment would have been what survived into the archaeological record. For the use of an apparatus by the Roman population and the modern population to be comparable a link must be made between them.

Ciarallo and Carolis give differing averages and ranges of heights (Ciarallo and Carolis, 1999). They use data from Pompeii taken from all skeletons found in public buildings, streets and villas in Pompeii (Ciarello and Carolis, 1999: 51), rather than the Herculaneum data which appears to have come mainly from the boat house skeletons, (Capasso, 2001). They state that the average Roman man in Pompeii was between 157.48cm and 167.64cm tall, (5ft2 and 5ft6 tall, 62-66 inches), with an average of 163.83cm (5ft4.5m, 64.5 inches). The average Roman woman in Pompeii was between 147.32cm and 157.48cm (4ft10 and 5ft2 tall, 58 – 62 inches), with an average of 153.67cm (5ft0.5, 60.5 inches tall). This is comparable to the heights stated by Capasso, (2001).

The data of the skeletal remains of the Roman population of Pompeii and Herculaneum is a known and finite data set. However, while the physical sizes of these individuals is known, their ergonomic limitations (such as reach) are not, and so alone these measurements can not be used to prove if they were capable of using the apparatus. When ergonomic data is collected it must come from a living, moving population. The data collected from the Roman population was collected while they were in skeletal form. While ergonomic data from a modern population may be applied to the Roman population it must come from a population of a comparable size and ability and closely resemble the original Roman data set.

It should be noted that conversion to cm are taken from the original texts and that this has resulted in some discrepancies between the conversions.

When a modern skeletal and ergonomic data set has been discovered the physical limitations of these individuals may be studied and compared. However, it is extremely unlikely that the modern population would have attempted to use a replica Roman dyeing apparatus. It is therefore necessary to find individuals that fit this data set and to allow them to use the dyeing apparatus. Only when a ‘real, live’ person of the same size and physical limits as a Roman dyer has been demonstrated to be capable of using the apparatus can any assumptions from this operation be applied to the Roman dyers.

Unfortunately, Capasso has not published the individual bone measurements that he used to calculate the individual skeleton statures. This means that it is not possible to check his calculations or use the lengths to calculate anything else such as the reach of the skeleton in life. Capasso cites his equation for height being taken from Trotter and Gleser (1958) without stating the equation itself. Biological anthropological equations used for height give a range with a correction. Capasso just states a final number. It is therefore not possible to work the equation backwards to determine the lengths of each bone. 99

Investigations into the Dyeing Industry in Pompeii Once the heights of the population of Herculaneum and Pompeii have been determined it is necessary to find a modern dataset that matches. By a coincidence of nature and history, such a data set exists in the modern era and has been examined and recorded in detail. It so happens that this data matches the heights of the population of the United States between 1900 and 1970, (Murrell, 1957; Dreyfuss, 1966). During this time three social changes occurred: women were increasing in number in the workforce, the motor car became widely available and diets changed. The change in diet lead to a sudden increase in height, or at least the expression of a greater range of genotypic heights, in both men and women. The availability of the motor car, combined with an increase in height, led to a sudden need to understand the ranges and abilities of stature of a suddenly increasingly divergent society. This means that there are data sets from numerous studies undertaken at this time by sources as diverse as car manufacturers and the armed forces. The data from these studies is a record of the change of a society and allows the understanding of a previous society that were of this physical size.

a b c d e

Maximum distance from centre of body to base of vat 0.66 m Body width 17 inches 0.43m Height to waist / Elbow height 1.01m Height to shoulder 1.32m Height 1.638m

Once the heights of the chosen modern Figure 5.1 The body measurements of a person who is 1.638m tall, population have been understood it is after Murrell, 1957. necessary to find modern individuals who can both match these heights and can operate the dyeing the use of the apparatus by these two individuals may be apparatus. In the original Archaeological study it was applied to the use of the apparatus by the original Roman determined that finding individuals who could match population. the original Roman data set and who were capable of operating the dyeing apparatus would be sufficient. The two individuals who constructed and operated the dyeing apparatus matched the American ergonomic The individuals working on the original construction and data. This allows a link with a data set that matches operation of the replica ranged in height between 153.67cm the original Roman set and the application of these and 187.96cm (5ft0.5 and 6ft2). When these heights are dimensions to the dyeing apparatus. compared with the ranges of the Roman heights it is possible to argue that the modern era population contained It was discovered following a visit to Pompeii that the larger individuals and that a comparison between the two author would have been capable of using each of the dye would lead to inaccuracies. However, when the averages vats except three, which were found to have been altered. were examined and compared with the two builders of All of the vats were usable if at their unaltered levels. the apparatus vat, it is shown that this argument would be inaccurate. The average height of a Herculaneum The equation used by Trotter and Gleser (1958) has yet male was 163.8cm, while in the rest of Campagnia it was to be surpassed in accuracy and is still used in forensic 164.0cm. The average height of a Herculaneum woman investigation, but even they state that it was not entirely was 151.7cm, while in the rest of Campagnia it was accurate when used without a knowledge of the subject 152.6cm. The height of the author is 163.83cm – within in life that is under investigation in death. As the range 0.03cm of the average Herculaneum male. The height of they give, at the most accurate, is +/- 4.59cm, (Trotter the co-builder is 153.67cm – within 1.07cm of the average and Gleser, 1958, cited in Bass, 1995:32), it may be noted Roman female. These two individuals not only operated that the skeletal measurements of the Roman skeletons the dyeing apparatus but actually constructed it. It is and those involved in the reconstruction are comparable therefore possible to argue that findings made through within the boundaries allowed in modern forensics. 100

Application of Ergonomics to Apparatus and Skeletal data

will stand. Further criticisms may be levelled against the reported skeletal conclusions. Ciarallo and Carolis use skeletal data from each part of Pompeii but do not list the individuals so it is not possible to check on inclusion, (Ciarallo and Carolis, 1999). Capasso states the sex of juveniles, (Capasso, 2001). Unless this is gained through DNA analysis this is notoriously inaccurate.

5.5 Build A further argument was that the build and physical fitness of the Roman dyers would differ from the individuals of the modern population, (Pers. Comm. Jones, 2002). It may be argued that the modern individuals had a better diet but a more sedentary lifestyle than the original dyers. However, the modern builders and operators of the apparatus were physically fit, both with a healthy BMI, both being keen sports women. Neither had a sedentary lifestyle and both were capable of lifting and reaching everything necessary for the dyeing process. It may be argued that the original dyers would have also been relatively fit as they would have been replaced by fitter dyers if they were not.

It may be noted that individuals who were 187.96cm (6ft2) also had no difficulty in operating the dyeing apparatus. 5.7 Difference of approach The difference of approach between archaeologists and engineers may be demonstrated when using the findings of Capasso (2001) and Ciarallo and Carolis (1999). Both sets of authors publish their findings, but do not publish the data, equations or working of each of their calculations. They instead publish the results and their interpretation of what this result means. This means that it is not possible to check their calculations or to use their data to answer other questions.

It may be noted that Ciarallo and Carolis described the build of the average Herculaneum resident, (Ciarallo and Carolis, 1999). However, they did not detail how they had calculated the build. If they had calculated the build according to the average weight for the BMI for each height, to use this figure to estimate the weight of each individual would result in a circular argument.

This problem has arisen when examining the skeletons of Pompeii and Herculaneum for an ergonomic study in engineering and the application of the skeletal data to the use of the dyeing apparatus. To demonstrate that the apparatus could have been used by the Roman population and the validity of findings determined through the use of the apparatus by a modern population it is necessary to demonstrate the continuity between these populations. For this it is necessary to determine the heights and reaches of the Roman population and a modern population, to determine that these match and can be used interchangeably and to apply the physical limitations of these individuals to the dyeing apparatus through the use of the apparatus by individuals from the modern data set.

5.6 Criticism of data Criticism may be levelled at Capasso’s use of skeletal data, (Capasso, 2001). Two individuals who were over 1.80m (6ft) were not included in the calculation of average height in Pompeii. These individuals were not included as they were not discovered at the boat houses. However, as the calculation is supposed to be representative of the entire population of Pompeii, this would mean the average is incorrect by up to 2 cm. However, it was noted that one of these individuals was wearing legionary uniform and the other was a gladiator in the Gladiator school. The Roman army never stationed recruits near home, so it may be presumed that the soldier was not local. It may also be argued that the gladiator was also not local. As they were both of known professions it may be presumed that they were not dyers and so would not have used the dyeing apparatus. Furthermore, as legionaries were unable to marry and the gladiator was locked away they were unlikely to have had a family locally.1 No other person in Pompeii had grown to this stature (the next largest was 175.26cm, 5ft9) so it may be argued that their genes and statures were isolated from, and not representative of, the population. Therefore their heights shall be ignored and the original averages of the Herculaneum populace

However, Ciarallo and Carolis (1999), while reporting the average heights of a Roman from Pompeii have failed to disclose how they determined these heights or to fully reference the data. They state that an average male was 1.66m, with a range between 1.63-1.69m. The average female was 1.53m, with a range between 1.511.55m. The Pompeii population’s heights were ‘similar to that of other ancient inhabitants of Italy’. Ciarallo and Carolis (1999), also give the weights of the population ‘Using formulas based on average body size’. Again the formulas are not given, although there is the statement that the sizes were similar to other ancient Italians. The average male was 65kg, with a range between 61-69kg. The average female weighed 49kg, with a range between 47-51kg. It should be noted that at each point ‘range’ is the range for the average due to differing methods of calculation. It is not the complete range for Italian weights in the ancient world.

At the time of writing, stating that a legionary couldn’t marry was accepted shorthand that he didn’t have a local family, despite the existence of vicus near forts. This is an example of the social customs an author writes within being transposed on to and influencing understanding of the subject they write about. Now as society evolves, this assumption cannot be made. For the purpose of this study, the pertinent point is that no other adults in Pompeii have been documented as having been his height, so if he had adult children that have been found, they were not as tall as him.

1 

101

Investigations into the Dyeing Industry in Pompeii It is interesting that when a modern Body Mass Index (BMI) calculation is applied to these skeletons that the average person in Pompeii was healthy and of the desired build for their health. For example a person of 166cm (5ft4.5) and of 65kg would have a BMI of about 23. 20-25 is seen as healthy. But it should be noted that these average weights for well-being would probably have provided the data necessary for an estimation in build of the skeletons, so to use these calculations would result in a circular argument.

some before what is widely regarded as the modern era of excavation, Pompeii is regarded as a ‘time-capsule’, a site that shows a city at a certain point in time, allowing the study of a society in its entirety. In Pompeii the entire populace died within a few hours and so there is the argument that an entirely representative sample of the population has remained, (Henneberg and Henneburg, 1999). Admittedly there is also the accusation that those parts of the population that were able to flee would not have been found among the dead within Pompeii and Herculaneum, but for the purposes of studying a crosssection the information is reasonably complete. It was noted that young males appeared to be under-represented in the sample and over 60s appear to be over-represented, but this could be due the young men leaving the city when the volcano first erupted, and the elderly were unable to leave, (Henneberg and Henneburg, 1999). Also Pompeii was a retirement resort for legionaries – it is very likely that there would have been a greater number of over 60s to start with. This may not be the case though as the gender ratio was 1:1 (Henneberg and Henneburg, 1999), unless women were living as long as the legionaries and men not in the army were dying earlier.

Ciarallo and Carolis, (1999) list the bibliography at the end of each section with no indication of which authors have been cited in each section or the source of specific data. While it may be technically possible to determine the source of their skeleton data, this approach leaves the original resources open to interpretation as they may contradict. Capasso (2001) published a 1043 page book detailing findings from the skeletons of Herculaneum. In terms of stature he was primarily interested in the heights, age and sex of the skeletons, from this working out the demography of the population with the fertility and death rates. Capasso used various bones from each individual, sometimes varying the bones used between individuals, to determine the final height, but other than referencing the equation used fails to specify the data used in each equation. This means that it is not possible to check the calculations (either individually or through the application of another equation) or to use the skeletal data for any further studies. The skeletons have each been examined for pathological remains, but measurements and photographs of the individual parts have not been published except where of interest pathologically.

5.9 Understanding modern ergonomics The majority of explanatory ergonomic texts are aimed to aid the engineer in designing devices (‘error output would be of most use around the prototype stage’), or to allow the evaluation of apparatus already in place when used by a skilled worker. They are not aimed at, but do allow, the evaluation of structures already in place (Stanton and Young, 1999). It may be demonstrated that only quantitative data may be gleaned from studying the vats in situ and the replica, as although a person matching the height, build and reach of an average Roman and a vat can be provided for the study, the behavioural practice is still unknown and so there is not firm qualitative data. It is demonstrable that a vat can be used or not, and how easily it may be used, but as it is not possible to watch a skilled Roman dyer working on a complete vat it is not possible to know their method or rate of working, and what the error or loss rate was. However there is data on whether the vats are usable: each dyeing apparatus was ergonomically tested by the author during the survey. It may be assumed that the vats were designed for use by the persons who were employed or owned to use them. This does not means that they were custom designed for each person, but that each vat should have been usable, (an idea expanded from Crumlin-Pedersen, 1999, who stated that a ship is designed to sail). Therefore each vat must be usable by the persons present. According to modern ergonomic ideas each item should be designed to take the variety of a population into account, so that 95% of the population should be able to use it. The distances involved should be small enough for the smallest person to reach, (Dul and Weerdmeester, 2001).

To work backwards from the final height using the equation to work out the measurements of each bone is not possible. Only where one bone has been used to find the height may this be done with any accuracy. However, the final heights have been given as a single figure. The calculation used should result in a range for the final height, the range decreasing as more bones are included in the equation. It is not possible to tell whether Capasso has taken the average or a part of the range to give the final number unless the individual bone lengths had been published. To try to use the final figure to determine the individual bone lengths would result in a circular argument as the individual bone lengths are what is used to determine the accuracy of the final figure for stature. 5.8 Studying Pompeii and Herculaneum Pompeii and Herculaneum are regarded as perfect sites for studying a complete population. Although it has been demonstrated that there was some disturbance in antiquity following the eruption (Allison 1992) and 102

Application of Ergonomics to Apparatus and Skeletal data

The question arose during the study of how different people react in a manual environment. It is true that two young adult British females who were unused to heavy lifting in a Southern Italian summer would react differently to two young adult males acclimatised to the area. A further study was made on how each part of the environment would affect the dyeing procedure and accuracy of results. Environmental temperature, humidity, airflow and the amount to be lifted all effect the ability, speed and accuracy of workers. The health and age of the workers, their acclimatisation to the environment and how practised they are at the work involved also affect the outcome of the work. There is a ‘comfort zone’ in each environmental factor and studies have shown that the further the worker is from the ‘comfort zone’ the greater the number of accidents and errors. Furthermore the age and acclimatisation of the worker affects where their comfort zone is, so that an older or less fit worker may not be able to tolerate a greater heat than a younger or fitter worker, (Nat Instit 1981). Dye works are highlighted as environments where the heat is so much higher than other environments that workers are at far greater risk from working there, (Murrell, 1957). While heat stress is tolerated in different ways by different people young adult males who are acclimatised can tolerate the most, but they have to be acclimatised. Even then only very few can. So while it is arguable that the persons working in this study are young and adult, they are both unlikely to have been able to operate in the Roman conditions. But where they were operating the replica these conditions were not replicated in full so this was not an issue. However it was possible to carry out dyeing in a 30°C and above heat wave. So based on all this information it is more than possible that healthy young adults could have worked in ventilated dye works with no problems. But in enclosed dye works with an increased heat and loss of ventilation other dye workers would have certainly encountered problems, although some could have become acclimatised.

for a length of time afterwards. This would have made emptying and cleaning the vat difficult and timeconsuming. Furthermore there was the danger of lead contamination, which although the Romans recognised as possible, would not have prevented its handling over a long-term period (Vitruvius). The temperature of the dye works would have been extremely high. If temperatures of 30oC and above were not unusual in Pompeii, to have 10 dye vats heating at the same time would have increased the temperature and the staleness of the surrounding air. This would in turn have increased the likelihood of heatstroke and other heat-related problems, especially in old or young persons or new (and as yet unacclimatised) workers. A person working in a dye works would have required between 10m3 per hr and 18m3 per hr of clean air for the environment to have been bearable, (Dul and Weerdmeister, 2001). While it is possible that in some of the dye works this would have been difficult, the exhaust gases released from the braziers as the fuel burned would have made this increasingly unlikely. This is before the extra oxygen that would have been needed to sustain each fire was taken into account. If there was insufficient oxygen to the fires there would have been incomplete combustion of the fuel leading to the production of carbon monoxide. Even complete combustion would have led to the production of water and carbon dioxide. Steam from the vats would have also added to the water vapour in the air. The evidence so far suggests that while working in the dye works could have been unpleasant, working there long-term as an unskilled worker (and therefore doing tasks that involved contact) could have been incredibly dangerous. 5.10 Lifting Maximum lifting limits are now law or recommended in many countries (including Britain, America and Canada) with a further guideline from the World Health Organisation. There are also guidelines about lifting differences between the sexes and through age and build. In the US (Nat Instit 1981) it is noted that although there is a limit, 25% of men and 1% of women have the ability to work above this. This begins to show the discrepancy between the sexes in terms of lifting ability and what can reasonably be expected. Murrell (1957) goes into further detail stating that men may be expected to lift 120 lb (50.8kg) occasionally and 100 lb (45.4kg) constantly, while women may only lift 60 lb (25.4kg) occasionally and 45 lb (20.43kg) constantly (although only two thirds of this weight should be expected for each realistically). This means that lifting a wet fleece is well within the realms of possible (as has been demonstrated repeatedly during this study) as is lifting the empty dye vat (the stainless steel dye vat) as this only weighs 13kg. The lead vat presents problems however as it is a much greater weight, 35kg. However many of the vats were in fact plumbed in and so lifting was not possible anyway, and buckets needed to empty

The apparatus to be handled and used also needed to be considered. The modern guidelines are that metal should not be above 50oC if it is to be touched for under a minute, (Dul and Weerdmeester, 2001), otherwise burns could result. Humidity should be between 30%-70%. It should be noted that although it was possible that any slave of any size or fitness could have been useful in a dye works, studying the above guidelines and limits the slaves used are beginning to look increasingly ill due to the high temperatures and the enclosed, unventilated surroundings. The vats would have regularly been above 50oC, although they would not have been touched whilst heating was ongoing. However they would have had to be cooled and it was found during the experiment that even when the fleece and water are removed the vat remains extremely hot (enough to cause burns) 103

Investigations into the Dyeing Industry in Pompeii or clean other vats would not have been of a prohibitive weight.

breaks is the most productive, (Fraser, 1980). This is the Roman working day (8am – 12, possible break for lunch, 12-4pm, according to Grose-Hodge, 1944). A break does not necessarily have to be a cessation of all activity – a change to a less strenuous activity is also restful. However it is not known whether this working day was allowed for slaves.

Continuous exertion of muscles should be avoided in working conditions as this causes pain and discomfort. Whilst this may not worry the master of a dye works, the prevention of further work would have caused concern. However, other than holding a siphon still to empty a vat or draining a fleece, there should not be anything in the dyeing process that would cause a constant strain. A repetitive strain could be caused in grinding dyes, but this is different as there are breaks in this process.

5.12 Nutrition Dye workers could only have worked to the maximum that they were capable of working through their calorific intake. A worker who receives less calories than they need for both their own well being and a specific task will be unable to perform the task to the fullness of their ability and become unwell. To perform ‘light sedentary work’ a male needs 2700kcal a day, while a female needs 2250kcal a day, (Fraser, 1980). For working in a dye works a male would need 3900kcal a day and a female 3250kcal a day, as this is medium-heavy work, (Fraser, 1980). It should be noted that although a male figure is given for heavy work such as mining (4200kcal a day) no figure is given for women, possibly because it is so strenuous. Both sexes are capable of carrying out dye working, so long as the environmental conditions allow it, but it would have been cheaper to feed female workers.

Fraser states that on average, female upper body strength is two thirds that of males, (Fraser, 1980). Also that if an apparatus is designed for the median/mean of the population then it will be useless as one half of the population will be unable to use it. It is better to design for the standard distribution and 95% of the population, making allowance for the rest. It was possible to have practically anyone working in a dye works. Lifting could have been carried out by anyone present who was nearing or at their adult size (so long as they were fit enough to bend and lift), as they would have been capable of reaching the bottom of the vat, tending the fires and fleeces and lifting anything that needed lifting. Persons that were not adult size would have still been capable of tending fires and fleeces, and if they were children would be acclimatised by the time they reached adulthood, with many years of knowledge and experience.

Nutrition could have also affected the height and build of the dye workers. If they had been raised as slaves it is possible that they would not have received sufficient nutrients in their diets to allow for them to express their full growth potential or for full muscular development. Both of these could have lessened their effectiveness when working in a dye works as lifting and carrying would have been required. This may also mean that the assumption that Roman dyers would have been fitter and stronger than the modern replica operatives is incorrect.

It has been noted that modern ergonomic data may be unreliable. Although figures may be given for the average male and female heights and build in an American population in the 1960s (Dreyfuss, 1966), data such as this tends to be incredibly specific thereby limiting its usefulness in design. As Fraser (1980) states, most of the data available at the time of his writing was gained from the United States military, which by its very nature means that it will only include healthy young adults, usually male and Anglo-Saxon in origin. There will be no further distinctions made within the data set between urban and rural populations or geographical variation, so any trends within the data will be averaged out. Methods such as this has led to problems such as the inability of female military engineers to use the tools that they are supplied with as these tools have been designed for a reach and grip much larger than their own.

It is not possible to determine the diet or calorific intake of the dyers. It is possible to determine if growth of a skeleton was interrupted through dietary deficiency, through examination of Harris Lines (White, 2000), but it is not possible to state which of these skeletons belonged to dyers. It is also not possible to determine whether the diet received in adulthood differed from that in childhood for the dyers. All that may be determined is that there was a specific calorific value necessary within the diet to allow dyeing to take place and that if this was not met the dyeing would have been substandard. It is also possible to state that healthy female dyers would have been cheaper to feed as they would have required less food than larger male counterparts, while still being just as capable of dyeing.

5.11 Working day It is possible that the Romans had an awareness of good working practice, certainly for the free. According to the World Health Organisation the eight hour day with

Note: The guidelines stated by the World Health Organisation and others would be within the physical 104

Application of Ergonomics to Apparatus and Skeletal data

Figure 5.2. Schematic showing the ability to reach the base of a dye vat without a step if the vat is small.

Figure 5.3. Schematic showing the need for a step when reaching the base of a taller vat. Note: the additional height of the step offsets the additional height in the vat reducing the reaching distance.

105

Investigations into the Dyeing Industry in Pompeii limitations to ensure that workers would not become ill. The slaves in Pompeii were not subject to these guidelines and so could be worked much closer to their physical limits. However, each guideline was usually accompanied with the absolute limits of endurance and size, which were of use in this study.

beneath the apparatus. There were also steps that were not wide enough for a person to stand reaching into the vat.

5.13 Necessary steps

The height of the brazier means that an apparatus may be one of two states:

It may be determined that there are four basic states that a vat may be:

For a step to be judged practical and necessary there must be a need for a step to be used when reaching the bottom of the dyeing apparatus. Without a step it is possible to reach the base of a dyeing apparatus. This is shown schematically in Figure 4.2.

• A step is necessary • A step is not necessary However, this sub-divides: • • • •

However, some vats are so tall that it is not possible to reach the bottom of the vat without a step. Therefore a step must be used. This is shown in Figure 4.3. For the step to be practical it must be of a height and width that a person may use it to reach the bottom of the vat. It is therefore possible to construct a step that is so tall that it is unsafe for the person to use as they would overbalance. It is also possible to construct a step that is too short as even with the step the person is unable to reach the bottom of the vat. In Pompeii it was discovered that there were steps constructed that may only have been used when there was no fire

A step may be necessary and is present A step may be necessary but is not present A step may not be necessary, and is not present A step may not be necessary but is present

A step being both necessary and present suggests that the apparatus is well preserved. A step being necessary but not present suggests either that the apparatus has not been well preserved or that the apparatus has been altered. This warrants further investigation. A step being unnecessary and not present suggests good preservation. A step being unnecessary but having been constructed anyway also warrants investigation.

Figure 5.5. Schematic demonstrating that it is not possible to reach the bottom of a kettle in a brazier if there is still over 1.01m of height after the step. The human body does not flex allowing such a reach.

Figure 5.4. Schematic showing that a step can be too tall to reach the base of a vat. To use this step the person on the left would overbalance. This step is unnecessary.

106

Application of Ergonomics to Apparatus and Skeletal data

Below is an examination of each vat in each of the properties. The presence of a step was determined at each one and whether a step would have been required. If an unnecessary step is present or a necessary step is missing this shall be investigated. Following ergonomic assessment it may be assumed that a step is required if the brazier is at least 1.01m tall with a vat base that is below 0.2m tall. An additional test is whether a person of Roman height, proportions and build is able to reach the base of the kettle while standing on the floor of the dye works, or where a step is present, on the step. Note: where the height of the base of the vat is unknown, the height of the firebox was used instead. Property I viii 19

0.34

Table 5.2. The presence of steps on apparatus in Property Vi4

1.14

0.36 0.21

0.4

In Property 1 viii 19 the three smaller vats (Vats 1, 2, 3) did not require steps and did not have steps. Vat 5 did require a step and had two, one placed in the middle of the back of the vat and one placed at the front of the vat. The step was constructed from rubble and was topped with marble. What is of interest about this step is that to use it with the fire in progress would mean burning the leg, see Figure 4.6. It may only be used by the left foot, requiring the right foot to be dangled over the firebox. It may be supposed that the step was only used when the apparatus was cold, or that it was used with great care or with practice.

Too tall or not required?

0.99

Yes

Necessary

No

Yes

With step

No

5

Vat

3

Vat base height (m)

0.38

0.84

Step height (m)

0.86

No

Brazier height (m)

Vat base height (m)

No

No

Step height (m)

No

2

Brazier height (m)

Necessary

1

Too tall or not required?

With step

Figure 5.6. Photograph of dyeing apparatus five in Property I viii 19. Step is necessary but if used when the fire is in place would result in burning as the flue from the fire hits the leg. Vertical measure: 1m. Horizontal measure: 0.2m.

Vat

Table 5.1. The presence of steps on apparatus in Property 1 viii 19

4

No

Possibly

0.73

0.42

5

No

No

0.610.66

0.20

6

Yes

Yes

1.05

0.15+

0.30

7

Yes

Yes

1.08

0.25

0.31

8

Yes

Yes

1.20

0.25

0.43

pristine, linked and of uniform height. If they were linked in antiquity they are likely to have been of comparable height. As restoration was so complete it is now not possible to know. This must be considered before they are examined.

It appears that each of the dyeing apparatus have not been adjusted. The heights and the presence of a step appear to be genuine and original. Volumes calculated from these apparatus would be accurate.

The dyeing apparatus had been restored. Restoration in this context means that following the original discovery of each apparatus, concrete and cement were applied to preserve the fabric and structural integrity of the remains. This has been a mixed blessing: the apparatus

Property Vi4 These vats had been the subject of obvious restoration dating from their original discovery. They are now 107

Investigations into the Dyeing Industry in Pompeii have survived into the modern era, but it is not possible to discern the accuracy of the restoration or the absence of detail. The attempt was made to restore features that were not understood, and while a faithful attempt at restoration of each feature was made, it is not possible to tell how accurate each attempt was. This is a problem when studying the apparatus, especially when making a functional study, as features that may have altered the function, such as steps, height, or the addition of flues, may have been amended, added or obliterated.

It was noted when examining this property that it was missing all content except for apparatus that had been physically affixed in some way to the walls or floor. It is possible that the steps could have been removed during the removal of any excess rubble, as the original excavators would have been unaware that steps were required when using the apparatus. To reach the back of each of the vats steps would have been required and not all of the vats that required steps have them. It may be supposed that they had been removed, that they were portable or that the structures had been incorrectly restored. However, it is probable that a further implement would have been required to actually reach the base of each vat due to their size.

The steps present in this property are all necessary and functional. They appear to have been heavily restored. As they all facilitate use of the apparatus it may be argued that they have not been amended beyond restoration and may be used to confirm the original heights of the apparatus and thereby volumes of the vats. Vat 4 does raise questions as it would be easier to use this apparatus if a step were in place, but there is nowhere physically to place a step that would not cause an obstruction. It is possible that a portable wooden step was used instead.

The requirement for steps suggests that the vats have not been shortened in height since their original use. However, their restoration, the absence of the steps and the ‘swept out’ feel to the property suggests that the vats may have been made taller during restoration. This is a feature of the vats in property IX iii 2 (see below). However as the heights appear to have been unaltered, volume calculations may be deemed accurate.

The trail of mortar leading up the wall from Vat 4 suggests that originally this vat was flued or rendered. It may be argued that these remains are of a flue as Vat 8 has the remains of a flue resembling these. For flues to have been restored the remains of each must have been obvious, as those who undertook the restoration would have been unlikely to have been able to discern whether a flue had been necessary, being unfamiliar with the newly discovered Roman remains. Vat 8 is so close to the door that an external chimney is unnecessary, but a flue to enable better combustion would have been desirable. The presence of relatively large vats in a relatively small property raises practical questions regarding storage and ventilation. The trail of mortar and presence of flues suggests that these vats have been restored to their approximate original height.

Property VII ii 11

Vat base height (m)

Step height (m)

Brazier height (m)

Too tall or not required?

Necessary

With step

Vat

Table 5.4. The presence of steps on apparatus in Property VII ii 11

0.39

?

0.39

?

3

No

No

Property Vi5

4

Yes

No

0.26

5

Yes

Table 5.3. The presence of steps on apparatus in Property Vi5

6

Yes

7

Yes

8 9

Vat

With step

Necessary

Vat base height (m)

1.48 1.14

Step height (m)

Yes Yes

Brazier height (m)

Yes Yes

Too tall or not required?

1 2

1

No

Yes

1.11

0.51

2

No

No

0.99

0.47

3

Yes

Yes

4

Yes

No

Too tall

0.99

0.60

0.31

1.02

0.60

0.38

0.78

0.50

Too tall

0.82

0.37?

No

Too tall

0.80

0.37?

?

No

Unneeded

0.81

0.16

0.27

No

Unneeded

0.77

0.16

0.21

Yes

No

Unneeded

0.69

0.27

0.35

Yes

No

Unneeded

0.73

0.27

0.20

The steps present in this property are problematic. They appear to be unnecessary, a fact borne out by the reconstruction and ergonomic evidence. They are a hindrance. However, the steps were constructed as permanent structures and so were seen as necessary. The narrow gap between Vats 3, 4 and 5, in which the steps have been constructed are too narrow for an adult to stand facing each vat from the side. The steps are

108

Application of Ergonomics to Apparatus and Skeletal data

altered, becoming shorter since antiquity, as the top of each appears to be missing. However, the final heights of both vats 8 and 9 are unknown. Property VII xiv 17

Necessary

Too tall or not required?

Brazier height (m)

Yes

No

Unneeded

0.95

3

Yes

No

Unneeded

1.22

2

4

Yes

5

Yes?

7

?

6

Figure 5.7. Step between vats five (on the right) and four in property VII ii 11. The step between the two is unnecessary as each brazier is only 0.8m tall. It is not wide enough for an adult to stand on facing either vat. The vats have not shrunk and the steps are not additional. Vertical measure: 1m. Horizontal measure: 0.2m

Yes

No

No No

Yes

Yes

8

No

No

10

Yes

9

No

?

No

Unneeded

Unneeded Unneeded

1.11

?

1.14

?

0.86 0.88 0.51

1.00 0.82

1.06

?

? ? ?

0.50

Vat base height (m)

With step

1

Step height (m)

Vat

Table 5.5. The presence of steps on apparatus in Property VII xiv 17

0.38

0.40

0.42

0.43 0.43 ?

0.43 0.39 0.43

0.50

Although this property has been extensively restored, it appears that the vats were of approximately the same height. Where they are not of identical height they were constructed to complement matching apparatus. Some of the vats had steps, while some retained adjoining rubble that had probably been steps originally. The vats were not so tall that steps were necessary, but the presence of steps would have been convenient. It must be noted that although the vats were tested for ease of use by a person of average Roman height, half of the dyers would have been below this height, including any children, and so may have benefited from the presence of steps.

unusable due to both this and their height. It would not be possible to use the steps without overbalancing. It is possible that the steps were constructed to assist a child in the use of the dyeing apparatus. It is only necessary to reach the base of the dyeing apparatus to empty or clean it. Both of these tasks are unskilled and time consuming and may be performed by children. If a child worked in a dyeing workshop and continued into adulthood the owner would have gained an adult worker with a sound practical knowledge of the dyeing process and who was themselves a skilled dyer. Over the generations the child dyers would replace the older skilled dyers, learning the trade as they grew. The steps on the apparatus in this property are wide enough for a child to stand or kneel on while reaching into the vat. As the steps became too small (as the child grew) the child would be able to reach the base of the vat by standing at the front. The progression of trade and property from slave or freeborn adults to children is discussed in greater detail by Mouritsen (2001).

Property IX iii 2 The base of these kettles would have been 0.22m lower than it is possible to reach when standing at the side or front and using the steps. These vats have been heavily restored and it is possible that this may have affected the remaining structure and design. On closer examination it is possible to discern a line in the external mortar above which a different mortar has been used in the restoration. This line occurs 0.22m from the top of the braziers. This is illustrated in the photographs in Figure 5.8. It may be argued that during restoration the heights of the braziers were altered. These vats are of a

The heights of the vats are undisputed as there is no evidence in the adjacent wall that they were originally taller. The heights of Vats 8 and 9 may have been 109

Investigations into the Dyeing Industry in Pompeii

With step

Necessary

Brazier height (m)

Step height (m)

Vat base height (m)

1

Yes

Yes

1.33

?

0.33

2

Yes

Yes

1.30

?

0.57

3

Yes

Yes

1.35

?

0.48

Too tall or not required?

Vat

Table 5.6. The presence of steps on apparatus in Property IX iii 2

issues of size the larger the person the easier it would have been). However, the application of ergonomic principles has led to doubts over the authenticity of some of the heights of the original apparatus, although it is possible to correct or allow for these heights in calculations. The presence of unnecessary steps is of concern and suggests endemic entities to the dyeing process or the use of children when dyeing. As one way of increasing slave ownership and training dyers was to keep and train the children of slave dyers (after Mouritsen, 2001) it is probable that some adjustment had to be made to the structures to allow their use by all available dyers. 5.15 Further work

differing design to each of the other vats that have been discovered and following the application of ergonomic principles it is possible to argue that they have been altered.

It was concluded that having reviewed the apparatus in context and determined the differences in design there is a need to understand the influence that these differences have on the dyeing process. Further experimentation was needed to examine the effects of differing design on the ventilation, fuel requirement and operational parameters of the apparatus. This work is presented in Chapter Six.

5.14 Summary By visiting the vats in Pompeii it has been determined that each of the vats would have been usable by the average male and female Roman (although due to

a

b

Figure 5.8. Photographs demonstrating amendment to vat one in Property IX iii 2. Photo (a) indicates the dimensions of the vat, Vertical measure: 1m, Horizontal measure 0.2m. Photo (b) indicates the line of the new mortar.

110

Chapter Six

Flued Experimental replica 6.1 Introduction

6.3 Vat size discrepancies in the replicas

Once a modern equivalent to a Roman dyer had been established it was possible to further explore the apparatus, in particular the operating parameters and differences in design. Further experiments using the replica were possible leading to the challenging of former assumptions. The apparatus could be amended to allow the affect that a change in design would have on the ventilation, fuel use and operational parameters to be determined. It was also possible to examine the vats more fully and to determine the stress that each kettle had been subject to during operation.

The project had been grounded in data from the original remains in situ gained through survey. This meant that any findings were based on the archaeological remains and so have an accurate foundation. However, this led to a problem in the amendment of the reconstruction. For an experiment in experimental archaeology to be valid only one change may be made at any point, (Mathieu, 1999). This allows the effects of that change to be determined and understood. The original replica apparatus was constructed based on an actual apparatus from Pompeii. The original replica apparatus was a copy of the design of Vat 5 in Property VII ii 11. On completion of the survey it was discovered that the vat was in fact a copy of vat 6 in this property. This vat therefore is accurate, even if it was discovered that it was accurate for a vat that it was never intended to be a copy of.

6.2 Flued vats It was discovered during the survey that there were two basic designs of dyeing apparatus: the flued and unflued vat. The flued vats were of the same design as the unflued, the only difference being the presence of a flue. The flues had been incorporated as part of the apparatus in their original construction, so it may be concluded that the vats with flues were designed purposefully to have flues and that those without were intended to be unflued. This raised the question of why the flues had been constructed and what difference the flues could have made to the dyeing process. The apparatus in property VII xiv 17 and in property VII ii 11 were of virtually the same size and design, the only difference being that those in VII xiv 17 had flues. It was noted during the survey that the flued vats existed in properties that were sheltered or enclosed. It has been supposed from this that as the flue would have required extra labour, materials, time and technique to build that it was necessary. It has been deduced that there was a need for a flue for each of these vats as the ventilation was different on each of them. It was therefore deemed necessary to understand the difference that a flue makes to a vat, an answer that may only be understood through further experimentation. The decision was made to add a flue to the previously constructed replica to allow a comparison of the two combustion processes. To add this flue will cause the replica vat to cease to be a replica of vat 5 in property VII ii 11 and instead become a replica of vat 4 in Property VII xiv 17. The replica will still be reasonably accurate as the height, width, size and volume of the two braziers are the same, as is the volume of the two kettles. However, there is a discrepancy between the two braziers which has caused a problem with the reconstruction.

The next stage of the experiment was to add a flue to the replica apparatus to gauge the difference that this would make to the operation of the vat. The flued vat is a copy of vat 4 from Property VII xiv 17. However, there is a problem. The flued vat is of a differing design to the vat that has already been replicated. The base of the flue must be level with the base of the metal kettle for it to be accurate but the original vat has a smaller fuel box than the flued vat. The replica apparatus has a fire box (and therefore kettle base) that is 200mm tall. The flued vat to be copied has a fire box that is 420mm tall. To add a flue at the correct height for the second replica would mean that the base would be against the side of the kettle and so this would not aid ventilation. To add a flue at the base of the kettle would mean that only one factor in an experiment would be changed at any one point, but no apparatus has been discovered that matches these dimensions and design. To simply add the flue here at the base of the kettle, while accurate in design, would diverge away from archaeology as there is no archaeological data to support this design. It would take the replica into the realms of ‘experience’ archaeology and not ‘experiment’ (as defined by Reynolds, 1999) as each experiment must be based on remains that have been discovered for it to be archaeologically valid. An alternative to allow the archaeological integrity of the experiment would be to raise the height of the firebox to 420mm and then construct the base of the flue at the correct height. However, this would double the height

111

Investigations into the Dyeing Industry in Pompeii of the fire box and so allow for a greater fire and would not allow for the effects of each individual change to be gauged.

was not possible in these workshops. The replica reconstructed for the experiment was unflued. The difference a flue made to the operation could be gauged by the addition of a flue.

It has been concluded that ‘the ends justify the means’. The aim of the project was to discover through replication of the apparatus the amount of textiles that may be produced given the equipment. The aim was not to build replicas for the sake of the replicas themselves, but to allow an understanding of influences of the design on the process of operating the apparatus. This means that an apparent disregard of archaeological information is justifiable in the short term when this intermediate stage is seen in this context. This would cause the experiment to loose its integrity in archaeological science as more than one factor would have changed. It was decided that the experiment was to replicate the concept using the example of the design. To meet the aim of the experiment it was necessary to continue use of the same apparatus to allow for comparative results. To amend the apparatus is to apply different parameters to the same geometry and underlying design of apparatus. While this may be argued to be diverging from archaeology it is still a valid scientific approach.

6.4.1 Hypothesis The addition of a flue would alter the airflow of the apparatus. The addition of a flue would decrease the amount of fuel required. The addition of a flue would lead to a more controllable flame. 6.4.2 Diagram

6.4 Experiment Three The first experiment successfully established how a dyeing apparatus operated and allowed an understanding of each part of the design. This allowed an analysis of each vat identified by Moeller and the discounting of one apparatus and the identification of alteration in a further three (for details see Chapter Four, Section 4.3.2). However, it was noted during the fieldwork that there were two basic designs of dyeing apparatus, the flued and the unflued. The flued vats tended to be in enclosed spaces, leading to the theory that without the flue complete combustion

Figure 6.1 Front view of apparatus. Apparatus was as described in Figure 3.3. Source: Author

Figure 6.2 Aerial view of apparatus. Apparatus as described in Figure 3.4. Source: Author

112

Flued Experimental replica

6.4.3 Apparatus Replica kettle (stainless steel) Replica brazier 9kg pine off-cuts Lid, 2cm thick wood (x2) Thermometer Stopwatch Sheep fleece, 2kg 90 litres of water 6.4.4 Method A replica Roman brazier was constructed. The brazier was 0.8m tall and 0.65m diameter. The kettle was 0.55m tall with a diameter of 0.55cm. The firebox was 0.2m tall with an opening 0.2m2. The brazier was constructed from modern brick and mortar and Figure 6.3. Photograph of flued dyeing apparatus rendered with mortar. The 6.4.5 Results kettle was constructed from stainless steel. This differed from the original Roman construction methods (lime Table 6.1. Fuel used in the dye run one of the flued mortar and rubble were used for the brazier and a lead apparatus. Ambient temperature 30oC kettle was used instead of a stainless steel one) but the conductive properties of the materials were the same Time Temp (°C) Fuel (kg) (Watling, 2004) and they were of the same size so the 0.00 22 1 +125g +530g +475g design remained the same. The replica was therefore a 0.15 25 550g +450g true replica. 0.25

A hole was placed level with the base of the metal kettle and a flue constructed. The flue was a double flue constructed according to the design of apparatus in property VII xiv 17. The kettle was lifted into place. Ninety litres of water was placed into the kettle. A small fire was lit underneath using pine off-cuts and fed with off-cuts throughout the experiment. The lids were placed directly on to the kettle top. The water was brought to the boil and simmered for 20 minutes. The fleece was then added and simmered for one hour. Following this the fire was extinguished and the whole apparatus allowed to cool naturally. The fleece was allowed to cool in the dye liquor then removed.

38 poss 28

425g

0.35 0.40

400g

0.45

49

500g

0.55

65

150g +160g

1.05

75

225g +800g +250g

1.15

88

475g

1.25

98

1.30 1.45

97

2.00

99

2.15

97

2 hrs and 15 min

113

75g +225g Put out Fuel 6.815kg Ash 175g

Investigations into the Dyeing Industry in Pompeii to heat the apparatus. The flame was easier to control and there was less expulsion through the front of the firebox. The flued vat was safer and easier to use than the unflued vat. It was noted that flued vats occurred in enclosed workshops, and it has been theorised that while flues may aid ventilation in such an environment, it may not cause an obvious difference in the environment in which the replica was constructed as this was already well-ventilated.

Table 6.2. Fuel used in the dye run one of the flued apparatus. Ambient temperature 30oC Time

Temp

Fuel

0.00

23

930g + 1.13kg + 750g

0.15

25

360g + 390g

0.30

38

600g + 180g

0.40

52

580g +505g

0.55

72

430g + 345g

1.00

80

200g

1.10

90

255g

1.15

94.5

340g

1.30

96

6.5 Stress on the vats It was noted during the survey that each of the vats were different sizes unless they had been constructed as a matching set. This raised the question of the effect that each change in design would make to the dyeing process through the structural differences between the apparatus. As the kettles were constructed from lead a significant change in the ratio between the volume and the diameter of the vat would result in a significant change in pressure on the base of the kettle. This would result in a difference between each apparatus in how long it would take for the kettle base to fail. It was therefore necessary to determine the significance to the pressure on each kettle base caused by the design of the dyeing apparatus.

Table 6.3. Fuel used in the dye run one of the flued apparatus. Ambient temperature 30oC Time

Water Temp Fuel (kg) (°C)

0.00

21

1.105 + 500g

0.15

25

1.080 + 540g + 1.255

0.30

42.5

380g + 160g

0.40

55

480g

0.50

66

650g

1.00

77

1.5

1.20

99

1.35

99

In Mechanics (Gere and Timoshenko, 1997), assuming that the loads (or forces) are uniformly distributed, the stress on a body (s) is defined as the force (F) divided by the cross sectional area (A), i.e.

1.50 2.05

98

2.20

98

2 hr and 20 min

Put out 7.645kg used 230g ash left

s=

6.4.6 Discussion

F A

The fundamental relationship between weight (W) , mass (m) and acceleration due to gravity(g) is given by Newton’s second law, F = ma, where a = acceleration. Weight is,

The apparatus operated in the same parameters as in the first experiment. The same quantity of fuel was used as the first experiment. The time taken to heat the vat and for the vat to cool was similar to the first experiment.

W = mg Where m is in kg and g = 9.81 m/s2. The unit of weight is the Newton (N). The weight of a body having a mass of 1 kg is

The flame was more controlled. It remained within the brazier, whereas the flame had often spilled outside the brazier during use of the unflued apparatus.

1kg × 9.81m/s 2 = 9.81N ≈ 10N

A greater amount of exhaust gases travelled through the flue. There was still expulsion around the vat and a limited amount from the front of the firebox, but the flue was effective in evacuating exhaust gases.

The units of stress is the Pascal (Pa), where 1Pa = 1N/m2. For a dye vat the uniform stress acting on the base of the vessel given by

6.4.7 Conclusion to Experiment Three The flue made little difference to the ventilation of the vat, the quantity of flue required and the time taken

s=

114

F = A

mw g (π D 2 4 )

Flued Experimental replica

Table 6.4. Stress on dyeing kettles

Property

Vat

Diameter (m)

Area (m2)

Height (m)

Volume (Litres)

Force (N)

Stress (Pa)

I viii 19 I viii 19 I viii 19 I viii 19

1 2 3 5

0.98 0.84 0.74 0.97

0.75 0.55 0.43 0.74

0.51 0.54 0.63 0.75

384.62 299.20 270.90 554.13

3846.19 2991.99 2709.02 5541.31

5100 5400 6300 7500

VI4 VI4 VI4 VI4 VI4

4 5 6 7 8

0.97 0.61 1.2 0.74 1.2

0.74 0.29 1.13 0.43 1.13

0.8 0.52 0.8 0.78 0.81

591.07 151.94 904.61 335.40 915.92

5910.73 1519.40 9046.08 3354.02 9159.16

8000 5200 8000 7800 8100

VI5 VI5 VI5 VI5

1 2 3 4

0.95 0.5 0.99 0.85

0.71 0.20 0.77 0.57

0.6 0.91 0.68 0.69

425.21 178.64 523.34 391.47

4252.13 1786.44 5233.44 3914.67

6000 9100 6800 6900

VII ii 11 VII ii 11 VII ii 11 VII ii 11 VII ii 11 VII ii 11 VII ii 11 VII ii 11 VII ii 11

1 2 3 4 5 6 7 8 9

1.06 1.06 0.96 0.76 0.5 0.86 0.55 0.99 0.94

0.88 0.88 0.72 0.45 0.20 0.58 0.24 0.77 0.70

0.69 0.74 0.6 0.56 0.43 0.54 0.55 0.38 0.55

608.80 652.91 434.21 253.99 84.41 313.62 130.65 292.46 381.62

6087.92 6529.07 4342.12 2539.94 844.14 3136.16 1306.46 2924.57 3816.16

6900 7400 6000 5600 4300 5400 5500 3800 5500

VII xiv 17 VII xiv 17 VII xiv 17 VII xiv 17 VII xiv 17 VII xiv 17 VII xiv 17 VII xiv 17 VII xiv 17 VII xiv 17

1 2 3 4 5 6 7 8 9 10

0.68 1.17 1 0.99 0.6 1.09 0.68 0.95 0.51 0.94

0.36 1.07 0.79 0.77 0.28 0.93 0.36 0.71 0.20 0.70

0.5 0.65 0.76 0.68 0.43 0.48 0.51 0.63 0.43 0.6

181.55 698.70 596.79 523.34 121.56 447.82 185.18 446.47 87.8247 416.31

1815.50 6987.04 5967.90 5233.44 1215.57 4478.19 1851.81 4464.74 878.25 4163.08

5000 6500 7600 6800 4300 4800 5100 6300 4300 6000

0.73 0.83 0.9

0.42 0.54 0.64

1.33 1.3 1.35

556.55 703.25 858.67

5565.51 7032.46 8586.71

13300 13000 13500

IX iii 2 IX iii 2 IX iii 2

Yellow = approximate values Green = minimum values Pink = maximum values

Note: Volume of Vat II, Property I viii 19 was derived from circumference of the kettle (2.64m), not the height and diameter, due to extensive damage. Circ/π = D 2.64/3.141 = 0.84 115

Investigations into the Dyeing Industry in Pompeii

Ε=ΔL LO

Where D = diameter of vat and mw = mass of water. The mass of water can be obtained from a knowledge of the vat’s volume (V) and density of water (rw) since

Where Lo is the original length and Δ l is the change in length.

mw = V r w

Strain is a measure of the amount of deformation (shape change) that a body undergoes when subjected to applied loads. Strains may be elastic at stresses up to the yield stresses or plastic at higher stresses. At room temperature for most metallic materials strain is time independent, i.e. a given stress will generate a particular value of strain which is unchanged for however long that the stress is maintained. Creep is time-dependent strain. In other words, for a given applied stress the strain increases with time. Creep is a thermally activated process which takes place when at temperatures above 0.4 – 0.5 Tm, where Tm is the melting point expressed in Kelvin. The melting point for lead is 300oC (or 573 K), so for lead at room temperature (20oC). The value of T/Tm = 293/573 = 0.51. Thus lead will creep at room temperature and at increasing rates at higher temperatures. In fact at room temperature lead creeps slowly under self-weight, (Ashby and Jones, 2005). Any evaluation of the operation of the original dye vats needs to take the creep phenomenon into account.

and rw can be taken as 1g/cm3, i.e. a litre of water occupies volume of 10 cm3 and exerts a load of 10 N. Using this information and the measurements recorded during the survey it is possible to determine the stress pressure on the base of a dye vat. The data is given Table 5.1. It may be noted that the pressures exerted on the base of the kettles were between 3800Pa and 9100Pa. There are three kettles that had stresses significantly greater than this (13,000 – 13,500 Pa). However, it has already been demonstrated through the application of ergonomics principles that these apparatus had been significantly altered and so stresses determined from the current dimensions would be inaccurate. The greater stress of these vats may be taken as further evidence of their alteration. It may be seen in Table 5.1 that despite the metal kettles being of differing dimensions and proportions, the pressure exerted on each was of the same order of magnitude. This suggests that, whether the manufacturers were aware of the effects of stress or not, they had sufficient understanding to allow the construction of similar apparatus. It is possible that the dyeing apparatus were constructed to an empirically determined design that was altered in ratio during construction according to the needs of the dyers.

6.6 Summary The addition of the flue to the replica dyeing apparatus made little difference to the operation of the apparatus, although the flame was easier to control. It was postulated that the flue had been constructed to aid ventilation where an apparatus was in an enclosed location. The kettles of each dyeing apparatus were exposed to stress of the same order of magnitude to each other, regardless of actual shape or size. This stress would have been great enough to cause strain and creep. When examining the dyeing apparatus in Pompeii it was noted that the majority of the intact lead kettles exhibited signs of bowing, (see Chapter Four, Section 4.4). Each kettle that bowed did so uniquely. It was postulated that this shape was caused by a damage accumulation process caused by the repeated heating and cooling that is part of the dyeing process. It may be supposed that the changes observed in the apparatus were due to creep. The process of creep and the affect that this would have on the dyeing apparatus is considered in more detail in Chapter Seven.

The short term room temperature yield stress for lead is in the range of 20 – 40 x 106 Pa, (or 20 – 40 MPa). At applied stresses below the yield stress lead behaves as an elastic Hookean solid. At stresses above the yield stress lead deforms plastically and as stress increases would ultimately fail by ductile rupture. The short term yield stress decreases with increasing temperature. At 100oC the yield stress for lead is ~ 12 MPa, (Ashby and Frost, 1982). The stresses in the dye vat are well below these values, being in the range 0.51 – 1.35 MPa, so the vats are able to support the mass of water in the short term. However there is a complication which is the phenomenon known as creep. Subjecting a body to a load not only generates stresses, it also creates strains. In simple terms strain is defined as

116

Chapter Seven

The Finite Element model use all forty apparatus. If this is combined with the differences in the influences of the surroundings on each apparatus during dyeing it may be argued that each of the forty must be constructed in surroundings that replicate their surroundings in Pompeii. If it may be argued that to allow a true comparison each of the dyeing apparatus should be constructed in a similar, controllable environment this would require such an environment to be found where each can be used in conjunction with only those that would have originally influenced it.

7.1 Introduction During the survey it was noted that the majority of the intact lead kettles displayed signs of bowing, although each had bowed uniquely. It was postulated that this was the affect of a damage accumulation process caused by repeated heating. It was deemed necessary to investigate this process as any damage to the kettle would have shortened its lifespan. However, the replica kettle was constructed from stainless steel, not lead, which although it shared many physical properties did not react to stress in the same way and so could not be used to investigate the strain of the dyeing kettles. Furthermore, as each kettle had bowed uniquely, the properties and behaviour of one could not be transposed to another. It was therefore decided that a virtual replica should be constructed to allow the construction of a replica with entirely authentic materials to allow a fuller understanding of the creep phenomenon.

It has been determined that the only way in which it would be possible to resolve each of these issues would be to construct the dye vats in an artificial environment that would allow the variables to be controlled and would eventually allow for the construction of all forty vats. Such an environment does not exist in the real world and would mean construction of the apparatus inside a computer. This has additional advantages. Once a single vat has been constructed it may be amended to allow an examination of the other 39 vats. The surrounding of the vats may be controlled and each subject to the same influences. The program used to construct the apparatus was designed to allow the breakdown and understanding of the physical principles within apparatus, specifically creep within metals and the transfer of heat. This would allow a greater understanding of the apparatus than would have originally been possible through the actual reconstruction of a physical replica. It should be noted that although a virtual vat may seem extreme or unrepresentative of a physical apparatus, the design and materials of the physical apparatus are the basis for the reconstruction. The temperatures and times that shall be used in the virtual replica are those taken from the physical replica and so are accurate and empirically gained.

7.2 Replicating the material behaviour of the apparatus When Reynolds stated that to construct a fully accurate replica it was necessary to use entirely accurate materials his statement was significant on many differing levels. During the reconstruction of this apparatus the materials were chosen to be of similar size and to have similar thermal properties to the original Roman materials. This was because using the original Roman materials was deemed too hazardous. As each of the properties that were to be studied were matched by the replica materials it was decided that Reynold’s statement was fulfilled in the replica as the materials were all representative of the properties that were being tested and the design itself was a replica. However, it may be argued that in fact Reynolds was correct as at a greater level there is no way in which to replicate an artefact without using the original materials. This may be demonstrated in the replica as there is no way in which to replicate the creep experienced by the lead without using the lead itself. In the replica the kettle was constructed from stainless steel, a material that will not creep in a similar way to lead.

7.3 Creep Creep is time dependent strain on a body subject to constant load or stress. Creep is a thermally activated process: the higher the temperature the faster the body will creep. Creep is thought to be of importance only at high temperatures, (Greenfield, 1972). However what constitutes a ‘high’ temperature is relative, (Pomeroy, 1978). Metals creep at temperatures above 0.3Tm, where Tm is the melting temperature in Kelvin, (Greenfield, 1972; Pomeroy, 1978). Lead has a melting temperature of 327oC, or 600K. Due to this, lead will creep at room

However, there are further difficulties in using a full scale replica. While each dyeing apparatus was of a similar design, each was of a unique construction, as only when apparatus was constructed as a matching set were they the same. To fully understand each apparatus it would be necessary to construct and 117

Investigations into the Dyeing Industry in Pompeii The micro-mechanisms of creep are complex and involve void formations and coalescence arising from grain boundary sliding, diffusion induced grain shape changes (ie elongation in the stressing direction) and softening arising from thermally activated dislocation climb. Relevant detail of these mechanisms, which often occur in combination, have been described by Ashby and Jones, (2005). 7.4 Changes over time

Figure 7.1. Schematic representation of a high temperature creep curve, (after Evans 1993). eo = Initial Elastic Strain (Ashby and Jones, 2005)

When a dye-run is in progress the apparatus is subject to changes in temperature. At the beginning of the dye run the kettle is filled with the water necessary to become the dye liquor. If madder is to be used, the dyestuff is added at this point. This is then heated with a fire underneath the kettle. When the water reaches 95oC it is held to temperature and simmered for one hour. Following this the fire is extinguished and allowed to cool. Figure 7.2 is a schematic of the temperature/time profile.

During a dye-run a dyeing apparatus is subject to heating under load, holding the temperature under load, cooling under load and then unloading once cool. This will place a strain on the metal kettle. As this cycle Figure 7.2. Graph showing a standardised temperature profile of the kettle during a is repeated the affect of strain on thermal cycle. Source: Author the dyeing kettle will accumulate and will ultimately result in failure. However, the rate at which creep occurs and temperature. Lead pipes sag without support and lead the number of cycles that it would have taken before on roofs crack during their life through exposure to the kettle fails are still unknown. It would be possible to heat, (Greenfield, 1972). determine how many cycles it would have taken for the kettle to fail by constructing a model of the apparatus Creep deformation occurs in three phases, the primary and simulating the cycles until the kettle fails. It should or transient phase, the secondary or steady-state be noted that after each thermal cycle the kettle would phase and finally the tertiary phase, which results remain in its new shape until heated again. This means in the material’s failure, Figure 7.1, while the first that each subsequent heating would effect a kettle that and final stages may take place relatively quickly, the had already been permanently affected, something that second stage, depending on values of applied load and must be allowed for in a full determination of the full temperature, may last for many years. (Greenfield, effect of creep. 1972).

118

The Finite Element model

1: The kettle is loaded with water, then the fire is lit 2: The load is constant, the temperature rises 3: The temperature is held as the dye liquor simmers, the load is constant 4: The fire is extinguished and the vat left to cool. As it approaches ambient temperature the fleece may be removed and the kettle emptied. Source: Author Figure 7.3. Schematic showing the relation of temperature of the kettle and the load the kettle is subject to during one dye run. The load and temperature are mutually exclusive but change over time as the dye run progresses. The combination of these two factors increases the rate at which the kettle to fails.

The dyeing apparatus is exposed to two differing forms of stress simultaneously during a dyeing cycle (see Figure 7.3). The water, dye and fleece are placed in the apparatus at the beginning of the cycle, thus providing the load stress by the weight of the consumables to be used in dyeing. Once the consumables have been added, the fire is lit below the apparatus and the apparatus is heated. The thermal loading increases as the heat the apparatus is subjected to increases. The apparatus is then held at a steady temperature and cooled with the consumables in place. Once the fire is extinguished the thermal stress begins to subside as the thermal loading is reduced as the kettle is allowed to cool. The load stress remains constant from the beginning of the cycle until the thermal cycle is complete and the consumables are removed.

ft = f (σ1 T)

Time to failure is a function of strain and temperature.

Figure 7.4 shows the combined load / thermal cycle produced by one dye run. Successive imposition of the cycle on the dye vat will result in creep and ultimately, as a result of damage accumulation, will result in failure of the vat. The question remains as to how long this would take both in time and in the number of cycles. This is expressed schematically in Figure 7.5. Grainsize must also be considered as the lead kettles had an increased resistance to creep due to their production method. The kettles were constructed from cast lead, not sheet (or ‘rolled’) lead. Cast lead has a coarser grain structure than sheet lead with a larger boundary area between the grains. This will have a beneficial effect since coarse grained materials creep at lower rates than fine-grained structures, (Ashby and Jones, 2005).

When the vat is subject to load and the resultant stress (caused by the self-weight of the structure of the vat and the load) in conjunction with the thermal cycle it will eventually fail. Time-to-failure (ft) is a function of load, strain and temperature. For a fixed load an increased temperature will lead to a decreased time to failure. For a fixed temperature an increased load will lead to a decrease in time to failure. This is illustrated schematically in Figure 7.4.

7.5 A virtual replica When determining how long it would have taken for a dyeing kettle to fail there are a number of problems in approaching the question. Construction and use of an 119

Investigations into the Dyeing Industry in Pompeii

0: The instantaneous elastic strain caused by loading the kettle 1: The strain caused by heating the water in the kettle, and thereby the kettle 2: Steady strain caused by holding the water, and thereby the kettle to temperature. This stage is assumed to be secondary creep due to the nature of the lead 3: Strain increases but the temperature effect reduces during cooling as the kettle is still under stress 4: Instantaneous reduction of strain as the water is removed from the kettle. However, as the lead kettle is inelastic the effect of strain has not reduced. Source: Author Figure 7.4. Schematic of strain on lead kettle during one thermal cycle

experimental replica has proven useful in determining the operating parameters of the apparatus, but unfortunately can not be used to determine the failure of the materials used in the original. This is because while stainless steel shares attributes of lead, it does not have as low a melting point and so does not creep in the same way. Furthermore, if a replica kettle was to be constructed of lead and the tests rerun, how long failure would take is still an unknown and it is impractical to wait the years that may be necessary. There is the additional problem that once the failure method of the model were determined it would be necessary to construct and test a further 39 models as each dyeing apparatus in Pompeii was unique.

the construction of the virtual replica already exists and is used extensively in engineering design as a means of validating and optimising designs before they are manufactured, (Fagan, 1992). 7.5.1 Modelling the dye vat The model is not of the material used in construction but of the design of the dyeing apparatus. It is possible to change the design without changing the material and cause the properties of an item to change drastically. The load will remain constant, the material may be the same, but if the design geometry is altered, for example by the wall of the vat becoming thinner, the apparatus would respond in a different way. The material has known properties, but the design and construction of the apparatus are what ultimately decides the behaviour of the apparatus. It is possible to use software to investigate behaviour of an article when influenced by external conditions to validate a design

It was therefore determined that a virtual dyeing apparatus should be constructed. This approach is not within the realms of experimental archaeology, itself a relatively new branch of science, but falls instead within the remit of engineering. Software that allows 120

The Finite Element model

Figure 7.5. Schematic showing the strain at different times during a dyeing cycle (heating an cooling once). See Figure 7.4 for explanation of a single thermal cycle. As lead is inelastic, the effect of strain continues after stress is removed resulting in an accumulation of strain and eventual failure. As time progresses the base of the kettle bows and fails. As the cycles accumulate the base fails. Source: Author

before manufacture. The apparatus shall be modelled using Finite Element Analysis software to determine the changes that occur. Once the virtual model of the design of the apparatus has been constructed it will be used to investigate the combined effects of load and temperature on the apparatus. The modelled apparatus shall be used to investigate the influence of thermal stress as the vat already under load stress is subjected to heating while under load.

finite elements, each part represented by an equation. Each material has a known set of physical properties and this too is included in the equations. As a force or temperature is applied to one surface or point of the material, its effect continues through the structure. Its influence reduces as it continues, so the effect on the opposite surface is much reduced yet unknown. The equations at each point allow the changes to the material caused by the external influence under the effect of the material properties to be gauged. The points at which specific changes occur may be located and studied. These calculations are used when exploring the effects of the application of temperature or stress to a material and are used when determining fatigue.

There are three different types of loads influencing the vat during its use: the stresses due to self-supported weight, the unloaded vat; the load caused by the water, the hydrostatic pressure; and the thermal stress caused by heating the apparatus. A model can replicate boundary conditions that a dyeing apparatus would have been exposed to, such as mechanical and thermal load and the contact between the kettle, supports and the floor.

The principles of finite element analysis are straightforward, yet the application is complicated. Before the advent of computer software that was able to solve the equations, each part had to be determined by hand. This was laborious, time consuming and could produce inaccuracies as the methods involved extrapolation from existing data to predict outcomes twenty years hence and there was difficulty in predicting the behaviour of complex structures. The application of computers has made the calculations

7.5.2 Finite element analysis Finite element analysis is used to determine changes within a structure when an external influence is applied. The structure is divided into discreet parts, 121

Investigations into the Dyeing Industry in Pompeii quicker to undertake, however has brought with it a new set of requirements. To use a specific FE program it is necessary to have an understanding of how the program operates and also how to design and input the code for the problem correctly so that the program can understand it. To write the code for archaeological artefacts is difficult as they are not uniform, they are incomplete and they have never been deconstructed in this way or coded before. What is worse is that as computer software is used so extensively these days there is little information available about how to approach the problem from an unfamiliar background. 7.5.3 Modelling creep A material’s response to an applied external load can be either linear (i.e. elastic or Hookean) or non-linear. Creep is a form of non-linear behaviour. For elastic behaviour the state of stress at a point is defined as shown in Figure 7.6.

σ = normal stresses and t = shear stresses Figure 7.6. The component of stress at a point (Moaveni, 1999)

The simplest way is through use of a ‘time hardening’ power law model (ABAQUS User’s Manual, 2002)

Because of equilibrium requirements only six independent stress components are needed to characterise a state of stress at a point, (Moaveni, 1999). Thus the state of stress at a point is defined as,

[s ]

T

ε cr = Aq nt m

(7.1.)

Where

= s xx s yy s zz t xy t yz t xz 

For an isotropic elastic material, stress and strain are modelled using the generalised form of Hooke’s law (ABAQUS User’s Manual, 2002).



ε cr



q = the uniaxial equivalent deviatoric stress



= the uniaxial equivalent creep strain rate

t = the total time

A, n and m are constants which are a function of temperature. For physically reasonable behaviour A and n must have positive values and m be in the range -1 to 0. Equation 7.1 plots in terms of creep strain rate and creep strain versus time as shown in Figure 7.7 (a) and (b), respectively. The coefficients to equations such as equation (7.1) are usually obtained from empirically derived data. Although a number of investigations of creep in lead have been published (Michalopoulos and Brotzen 1968, Hofman 1970, Mohamed et. al. 1973) the results presented can not be used to generate a creep model either because they are presented in terms of creep strain rates for different test temperature and stress levels or the test were conducted at too high a temperature. Fitting coefficients to Equation 7.1 required the original creep strain-time data. Fortunately, data in this form was available from the study of the creep of lead published by Sahota and Riddington (2000). An Excel spreadsheet was created to curve fit the empirical data of Sahota and Riddington

In FE packages such as ABAQUS isotropic elastic behaviour is modelled by inputting appropriate values for the material properties, Young’s modulus (E), Shear modulus (G) and Poisson’s Ratio (n). Modelling nonlinear behaviour is much more complex and involves the use of constitutive equations. ABAQUS offers a number of different ways of modelling creep behaviour.

122

The Finite Element model

Table 7.1. Values of A, n and m used in the Time Hardening Creep Model implemented within the ABAQUS simulations. Temperature (C) 20 40 80 100 120

A

n

m

1.1 x 10-4 2.8 x 10-4 5.0 x 10-4 8.0 x 10-4 1.5 x 10-3

2.0 2.0 2.1 2.1 2.1

-0.70 -0.75 -0.78 -0.80 -0.85

(2000). Estimate values of A, n and m obtained through this approach were validated via an ABAQUS simulation of the experimental creep tests of Sahota and Riddington (2000). This demonstrated an acceptable degree of correspondence between the FE simulation and the experimental data (see Appendix 4 for details). The derived values of A, n, M used to model creep behaviour of lead in the ABAQUS simulations are presented in Table 7.1.

 Figure 7.8 shows creep curves for lead at q = 5 MPa modelled using the data presented in Table 7.1

Figure 7.7. Schematic representation of Equation 7.1.

Figure 7.8

123

Using computer simulation in Finite Element Analysis modelling has replaced the earlier method of extrapolation from the known material properties. Computer simulation is

Investigations into the Dyeing Industry in Pompeii a more sophisticated method allowing a more accurate representation of reality. The cross-over between these two methods was in the early 1980s when the computer software that allowed this modelling was developed, (Boyle and Spence, 1983).

the apparatus lost heat, it was not possible to determine exactly where this loss took place. Also, it was not possible to calculate the movement of heat through the apparatus. It was therefore decided that the experimental dye-runs should be re-run with temperature sensors in place to record the temperature during heating and cooling at strategic points around the vat.

7.6 Requirements for the model A computer simulation of the dyeing apparatus consists of two parts in combination:

Experiment Four was a re-run of experiment Three. The only difference was that this time it involved the use of temperature sensors to record the temperature at different points of the apparatus throughout the duration of the experiment.

• The geometric model • The thermal load To construct the physical model it is necessary to input:

7.6.1.1 Hypothesis

• The measurements of the apparatus, thereby constructing the design • The properties of the materials used in the construction: The elastic properties, the plastic behaviour and the creep behaviour. These are inputted as functions of temperature to allow the measurement of the material response to see if it’s plastic or elastic under 1) self-weight, 2) selfweight with the addition of water and fleece. • The presence of 90 litres of water. This will allow the model to respond to the load. The density, thermal conductivity, specific heat capacity and the presence of the water as a liquid should also be inputted. • Time. This allows a determination of the physical effects of the gravity and additions of load, and a fuller understanding of how the effects occur and progress.

The use of temperature sensors will allow the recording of temperature at different points within the apparatus during the experiment. The apparatus shall operate as it did in experiment three. 7.6.1.2 Apparatus Replica kettle (stainless steel) Replica brazier 9kg pine off-cuts Lid, 2cm thick wood (x2) Thermometer Stopwatch Sheep fleece, 2kg Shetland 90 litres of water TC-08 datalogger with Type K Thermocouples (x6) Laptop with power supply 7.6.1.3 Diagram

The dyeing apparatus was heated over time and allowed to cool. To model the physical effects of heating it was necessary to input:

Below are diagrams of the apparatus at the start of the dye run. Sensor five contained a piece of wire that wrapped under the rim of the kettle, indicated by the red line. Sensor one was directly below sensor four.

• The physical model and ensure that it was stable, reliable and would withstand alteration • The presence of 90 litres of water • The temperature profiles of the experimental apparatus showing the rate and temperature of heating, holding at temperature and cooling.

7.6.1.4 Method The replica brazier used in experiment three was reused for this experiment.

However, to construct a model of heating based on the actual temperatures of the dyeing apparatus during operation, there is a need to know the actual temperatures at various positions around the dye vat. This required additional experiments with the replica.

The sensors were placed around the apparatus (see diagram). Channels: 1 = base of flue (only used once, it melted) 2 = Handheld probe. Held above fire, but not in contact with vessel. 3 = Air at back 4 = Top of flue 5 = Round top of vat 6 = Outer base of flue

7.6.1 Experiment Four During the original archaeological work only the temperature of the water had been monitored. There had been no temperature recordings from the rest of the apparatus itself. While it was possible to calculate that 124

The Finite Element model

Figure 7.9 Front view of apparatus with sensors in place. Source: Author

Figure 7.10 Aerial view of apparatus with sensors in place. Source: Author

The kettle was filled with 90 litres of water. Recording of the temperatures was started as the fire was lit in the firebox. The water was heated to 95oC and the fleece was added. The water was then simmered for an hour. Following this the fire was then extinguished and the fleece was allowed to cool naturally.

7.6.1.5 Results Two types of results were obtained. 1. The fuel and time required to heat the dyeing apparatus 2. The temperature of the apparatus during heating

The procedure was repeated three times to allow a range of results to aid understanding and comparison. 125

Investigations into the Dyeing Industry in Pompeii 1) The fuel and time required Table 7.4 The fuel used and time taken to heat the water to 95oC

The amount of fuel used to heat the vat, the times at which this fuel was added and the time taken to heat the vat were recorded.

Third Experimental Re-run Time (Min) Fuel(g) Water temp OC 0 725 20 15 1725 16 30 2675 26 45 3250 43 55 4450 52 63 5100 65 66 70 5800 80 6450 82 85 7450 90 90 100 97 110 100

The first run was to test the apparatus. The temperatures recorded through the sensors are shown in Figures 7.11, 7.12, 7.13 and 7.14. As the temperature was recorded one hundred times every minute the overall picture of the changes of temperature may be appreciated best in graph form. The actual readings may be found online at https://doi. org/10.32028/9781789697421-online. Table 7.2 The fuel used and time taken to heat the water to 95oC Time (min) 0 5 15 40 50 65 75

First Experimental Re-run Fuel (g) Water Temp oC 775 2575 3775 4275 4800 5200 6150

Table 7.5 The fuel used and time taken to heat the water to 95oC Fourth Experimental re-run Temperatures oC Time (min) Fuel Water Ambient 0

10

Table 7.3 The fuel used and time taken to heat the water to 95oC

Time (min) 0 5 10 15 25 30 35 45 55 65 70 75 80 85 100 110 120

15 20 25 30 32

Second Experimental re-run Temperatures oC Fuel (g) Water Ambient 1000 18 20 925 750 24 22 30 21 500 775 36 40 225 48 24 1200 58 350 70 775 22 80 98 100 100 17

40 45 50 55 60 65 70 75 80 85 90

100 110

126

250g 1050g 50g 450g 775g 700g 700g 875g 450g 900g

20

24

21

23

26

27

38

31

48

30

750g 100g 500g

57

25

66

27

800g 250g 325g

76

23

87

26

95

24

Simmer 20 min. 100

The Finite Element model

2) The temperature of the apparatus during dyeing

Figure 7.11 Graph indicating the temperature at each point during the first dye run

Figure 7.12 Graph indicating the temperature at each point during the second dye run

127

Investigations into the Dyeing Industry in Pompeii

Figure 7.13 Graph indicating the temperature at each point during the third dye run

Figure 7.14 Graph indicating the temperature at each point during the fourth dye run

128

The Finite Element model

7.6.1.6 Discussion of Experiment Four

allows the movement and transfer of heat, the loss of heat, and the starting and finishing temperature at points on the apparatus to be determined.

Following the dye runs it was important to measure the temperatures at each part of the dyeing apparatus during its use to gauge the changes in temperature during the heating, simmering and cooling. This will allow an understanding of movement and loss of heat through the apparatus. The recorded temperatures may be used to reconstruct how the apparatus reacts and behaves to changes in heat and these figures may further be used to understand other physical changes within it during its use.

7.6.2 Average temperature profile To use the recorded temperatures in further work it was necessary to formulate an ‘average’ dye run that represented the temperatures that an apparatus would be exposed to during a dye run. This could then be used in the model to represent the temperature changes that occurred within the apparatus. However, on examination of the temperatures recorded during operation of the dyeing apparatus it may be noted that there is a great discrepancy between the temperatures reached during each dye run. This is caused by the use of an actual fire to heat the apparatus, which by its nature is unpredictable.

The replica brazier used in experiment three was reused for this experiment to maintain consistency between the experiments and allow comparison of results. The method of the dye run was undertaken as a copy of the first sets of experiments so that the figures gained would be representative of a dye run and comparable to previous results. Furthermore the recording of weight of fuel used ensures that comparisons between the experiments may be made as the dye run may be seen to be representative of the previous dye runs.

During a dye run it is necessary for the water to reach the activation temperatures required for the dyeing process to work. It is also necessary that each of the temperatures when reached should be held for the time necessary for the chemical process to take place. For example, if an insufficient temperature is reached the dye may not be released from the dyestuff, or if heated for insufficient time it may not bind to fabric or mordant. How long the apparatus must be heated and how long the temperature is held for after the time has been reached is unimportant, (unless the dye may be ‘stewed’ and process fail). Therefore it is possible to determine from the graphs the activation temperatures, the points at which these are reached, the time the temperatures must be held for and a representation of temperatures reached during the dyeing process. It is important to gauge a representative graph, not an average graph as the average graph would not be representative. However, it must be noted that the representative graph will look nothing like any of the original graphs from which it is constructed.

Comparison of the graphs of dye runs in 2002 and 2003 with the dye run of 2005 demonstrates that there is no significant difference between any of these replica experiments. Therefore the temperatures recorded during 2005 may be used to represent the temperatures that the apparatus would have reached in 2002 and 2003. This may also be used to represent the original Roman apparatus as the replica was constructed to have the same physical properties and design as the original. It may be discerned that there are small differences within the recorded temperatures and the quantity of fuel used each day. This is due to natural variation. Wind speed and the ambient temperature differs from day to day, as does the make-up of the fuel used on the real fire. However, the graphs indicate that the difference between each set of data is so small that each may be taken to represent the whole. It should be noted that the original Roman dyers also worked with natural variation and no two dye runs would have been the same.

Below are the temperature measurements taken as representative from the recorded figures. The beginning temperature of the water was 20oC. This would rise through heating to 95oC. The rise from 20oC to 110oC at channel two represented the lighting of the fire and original heating of the apparatus. As more fuel was added the temperature would drop, represented in channel two as a drop to 35oC, as the energy required to pyrolyse the lignin and other products within the wood detracted from the heat that the fire could produce. At this point the fire had not yet caused the apparatus to heat fully, represented by the continuation of channels four and five at 20oC. As the fuel broke down and the energy was utilised and realised as heat the temperature rose at channel two. This was reflected by a rise at channels four and five as the heat transferred through the apparatus.

7.6.1.7 Conclusion to Experiment Four These recordings allow a solid foundation in any future work involving the understanding of the process of heat transfer within the dyeing apparatus. Prior to this work no such recordings have taken place and so any calculations regarding heat transfer were estimated based on the ambient and water temperatures. These recordings have placed the archaeological experiment of how to physically operate the apparatus into the engineering context by providing the data that 129

Investigations into the Dyeing Industry in Pompeii The consequence of heating one channel may be seen in the rise in heat at another channel, but there is a thermal time lag between these changes. The rise in heat at the second channel will be delayed as the heat has to transfer through the system. The addition of fuel to the fire temporarily lowers the temperature as the energy lost as heat is instead used to pyrolse the lignin and other products within the wood. Once the lignin has been broken down the energy within these products is converted to heat and the temperature of the fire rises.

Table 7.6: The activation temperatures at the necessary times at each point of the vat during heating Average Amplitude curve for four days of temperatures of the flued vat Time Channel 2 Channel 4 Channel 5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120

20 110 35

140

20

20

20 200

See Figure 7.15 for a graphic representation of how these channels relate.

130

7.7 Constructing the finite element model 140 60 140

75 145

The model was constructed in the order described above.

75 150

140

150

145

110

110

110

7.7.1 Modelling the apparatus in the computer To model the dyeing apparatus the design must be simplified to allow its conversion to code and its input into the computer program. The computer sees a functioning apparatus as part of a system, not as a solid object operating alone. It is therefore necessary to construct the apparatus in a way that the computer can ‘see’ it, then construct the system in which the apparatus operates in the computer containing all of the inputs, outputs and losses within the system. As constructing a dyeing apparatus in this way had not

Figure 7.15 Graph showing average amplitude, gained through merging each dye run

130

The Finite Element model

Figure 7.16 shows the plane of the cross section taken.

yet been attempted, it was necessary to design all of the code for the apparatus from first principles. This required breaking the apparatus down into the simplest form through which the system operated and then building up back into a three dimensional system.

Following the construction of the geometric model the material properties of the constituent parts of the design could be included. This included the elastic and plastic behaviour of the materials during heating. This would allow a replication of the behaviour of the design and apparatus to the application of differing temperatures from 20oC to 120oC. The loading of the dyeing apparatus through the application of mechanical strain through the addition of weight and the application of thermal stress through heating could then be observed. A thermal cycle could then be undertaken and repeated to allow the discernment of the effects of elastic and plastic strain on the apparatus and the eventual effect of creep on the kettle.

First it was necessary to simplify the apparatus. It is not necessary to replicate the rough surfaces and joins in brickwork, but instead to define a uniform surface that the computer can understand the relevant values of and which it can see in relation to each other surface and within the system as a whole. It was therefore necessary to take the design of dye vat and turn it first into a 3D design, a geometric model. It was decided that the apparatus should be replicated as a 2-dimensional object and then expanded to become a 3-dimensional object once it was understood how the 2-dimensional design operated, so the geometric model should be constructed in two dimensions.

The dyeing apparatus was reduced as shown in Figure 7.16. Once the final cross section has been determined it must be broken down into elements.

It was realised that as the object was symmetrical and subject to the same external influences throughout that only half of the object required replication at this stage. Therefore the model should be axi-symmetrical, the model having circular symmetry around the vertical axis through the centre of the apparatus. This allowed a good approximation of the apparatus, although it did assume continuous support for the metal kettle instead of the pillars exhibited in some of the apparatus. It was decided that when the object was eventually replicated using three dimensions it would be necessary to replicate all of it.

Figure 7.17. Labelling of elements and nodes

131

Investigations into the Dyeing Industry in Pompeii enough elements to allow a change across a surface to be fully represented. If the item is to be three dimensions then it must be broken up in all three dimensions. These elements should each be labelled in such a way that the computer can recognise the location of each one and each may be plotted without overlap.

7.7.2 Geometric model When using finite element analysis to model an item, it is first necessary to break the item into theoretical pieces called elements. Each point where an element meets another element (a node) is described by an equation, so it is important to ensure that there are

Figure 7.18. Diagram showing the division of the base of the dyeing kettle into elements, 5 x 275

Figure 7.19. The apparatus simplified and broken into elements

132

The Finite Element model

Each separate part of the item must be broken up using a separate set of elements and nodes. This means that the computer can be told that each part is separate (such as the water being separate from the metal kettle), allowing each part to behave in a different way. However, in this example the elements and nodes had to be labelled in such a way that there was both no repetition and that the elements stayed under 10,000 in number (due to the limits of the Abaqus educational licence). This proved problematic, but was eventually solved. To code the apparatus into the computer, the apparatus must be further broken up. The kettle must be split along the join in the metal to allow the base and the side to be coded separately, otherwise the computer will not see their physical relationship correctly. The base was modelled using 1x1 mm2 elements. However, the mesh size was varied in the side wall (using the BIAS command in ABAQUS) with 1x1 mm2 elements at the bottom, increasing in size to 1 x 10 mm2 at the top. A high mesh density was used at the base and bottom of the side wall since this is the critical region and would increase the accuracy of the solution, (Fagan, 1992). 7.7.3 Material properties Following this the physical properties of each of the materials used within the construction were entered into the program. These included the material properties such as density and elasticity of the lead and the elastic/plastic behaviour and creep over a temperature range. This also included the thermal properties of the lead such as the thermal conductivity and expansion. The values for elasticity, plasticity, creep, coefficient of thermal expansion, specific heat capacity and conductivity were each entered as a function of temperature.

Figure 7.20. Diagram of elements used in model

• Hydrostatic loading: the dyeing apparatus had to support the weight of 90 litres of water. • Thermal loading: The dyeing apparatus was subject to repeated heating and cooled. 7.7.4.1 Self-weight The dyeing apparatus originally had to support their own weight before any additional loading (with water, fleece and then through heating) took place. It was therefore important to construct an accurate design to establish how the materials and design of the apparatus would have affected how the apparatus supported its own weight.

7.7.4 Loading There are three ways in which the dyeing apparatus was subject to loading:

Figure 7.21 demonstrates that while the lead kettle wall is supported by the brick support, the base of the kettle

• Self-weight: the dyeing apparatus had to support the weight of each part of it 133

Investigations into the Dyeing Industry in Pompeii

Figure 7.21. Schematic show self-weight. The apparatus is supported only by the brick support. The majority of the lead base is self-supporting.

is mostly unsupported. It is important to establish the effect that the lack of support would have had on the lead before the addition of water. It is possible that the weight of the lead alone could have been detrimental to the apparatus.

Figure 7.22. Schematic demonstrating hydrostatic pressure

pressure of the water that is important to include in this simulation, it is possible to include the hydrostatic pressure that the apparatus was subject too rather than the actual water itself.

7.7.4.2 Hydrostatic loading During a dye run, the apparatus was loaded with 90 litres of water prior to heating. It is necessary to simulate the presence of this water to allow an examination of the effect of this weight, and the effect of the thermal loading on the apparatus subsequent to the imposition of the boundary conditions. To actually use water broken down into elements and nodes is impractical and unnecessary in this simulation. Instead, it is possible to simulate the presence of water through the inclusion of hydrostatic pressure as a boundary condition.

The material properties defined for water were each entered as a function of temperature. The material properties for water are consistent below boiling point and it was possible to add the surface co-efficient between the water and the air above it. This was entered as ‘5E-6’ as the co-efficient is 5 x 10 –6 . (Pers. Comm. Robinson, 2007). The water was tied to the metal vat and allowed to flow as a liquid, but a co-efficient was added to ensure that the water did not heat to the same temperature as the metal. Nodes were identified along the centre line to allow the heating of the water to be observed.

Hydrostatic pressure is the effect by which the pressure that water is subjected to at the bottom of a container is greater than the pressure it is subjected to at the top of the container. Water within a container is subject to the weight of itself and the weight of water above it. The water further down a container will have a greater amount of water above it than the water nearer the top of the container. This means that the water further down the container will be under a greater weight and therefore pressure than the water above it. As the rate of this change is known, and as it is the weight and

7.7.4.3 Static load Once the design and material property had been programmed, the water added and the effect of gravity placed on the dyeing apparatus, this was declared a finished physical static model. At this point the model was a representation of a loaded dyeing apparatus at 20oC. Once the model has been completed it is possible 134

The Finite Element model

to apply time to determine how the apparatus will react at 20oC and create a mechanical model of failure. The times taken during the dyeing process were inputted and the effect of the loading and unloading of the apparatus was observed to determine that the model was a functioning simulation. This was done to validate the integrity of the model before temperature was applied.

The temperatures programmed into the model were taken from channels 2 and 5 on the dye run, (see Figures 7.9 and 7.10). Channel 2 was taken as the source temperature. Channel 5 was taken as the sink temperature. Heat lost between these two has been lost to heating the water and the external elements (the losses in the system). However, a problem was encountered as there were times during the rerun when the temperature at channel 5 was greater than the temperature at channel 2. This was due to the heatsource being a real fire. This means that it is not uniform – it is possible that it can tip or be ventilated un-uniformly, causing it to burnt greater under channel 5 (the air for which will conserve heat as it is rewarmed as it rises next to the vat) than it did under channel 2 (taken to be the temperature of the fire but which was exposed to real changes in wind and air temperature). However, this has been allowed for and the data inputted into the model was adjusted accordingly.

7.7.5 Introduction of temperature Following the construction of a physical model to allow the observation of the effect of loading it was possible to combine the design model with the thermal history of the dyeing apparatus to allow the determination of how the dyeing apparatus would have changed over time under the combined action of stress and thermal loading experienced during a dyeing cycle. After the temperature history has been outputted into a file, it may be read back with the mechanical loading, superimposed on the thermal history. This was done as a coupled-thermal displacement analysis, including the creep model for lead. Following the rerun of the full-scale replica dyeing apparatus with the inclusion of thermal-couples it was possible to input a real temperature set. Use of these temperatures will result in the vat being exposed to the correct temperatures and times and, when combined with the already existing mechanical model of failure, it will be possible to simulate the behaviour of the dye kettle during a dye cycle or series of dyeing cycles.

The Abaqus input decks used to construct this model may be viewed in Appendix Five. 7.8 Results Figure 7.23 shows the constructed dyeing apparatus at 20oC, unloaded. At this point the only load that the kettle is subject to is its self-weight. It is supported by the brick (bottom right). There is no discernible effect on the kettle caused by this weight.

Figure 7.23. The strain to the dye kettle when under self-weight

135

Investigations into the Dyeing Industry in Pompeii Figure 7.24 shows the base of the kettle in greater detail.

Figure 7.24. The strain to the dye kettle when under self-weight

The kettle was then subjected to the weight of ninety litres of water. At this point the ambient temperature and the temperature of the water were both 20oC.

Figure 7.25. The strain to the dye kettle when under self-weight and loaded with ninety litres of water

136

The Finite Element model

As may be seen in Figure 7.25 the stress caused by the addition of ninety litres of water causes the kettle base to become strained. Following the addition of the water, the apparatus was heated from 20oC to 120oC in ten minutes. The temperature was held for one hour. The combination of thermal stress and the load stress caused by the ninety litres of water caused the kettle to deform.

Figure 7.26. Loaded kettle following heating to 120 oC over ten minutes during first thermal cycle

Figure 7.27. Loaded kettle following heating at 120oC for one hour during first thermal cycle

137

Investigations into the Dyeing Industry in Pompeii

Figure 7.28. Loaded kettle following cooling during first thermal cycle.

Figure 7.29. Loaded kettle following heating to 120oC during second thermal cycle

138

The Finite Element model

Figure 7.30. Loaded kettle following heating at 120oC for one hour during second thermal cycle

Figure 7.31. Loaded kettle following cooling during second thermal cycle.

139

Investigations into the Dyeing Industry in Pompeii It may be noted from examination of Figures 7.24 to 7.31 that the dyeing apparatus begins to show signs of strain from the time at which the kettle is filled with water. The greatest points of strain are at the brickkettle interface and as the base of the kettle alters as it reaches the floor. The kettle stretched at the points of greatest strain. The strain was not above 3% in the simulation, a relatively low strain, that suggests that the kettle was resilient.

a better understanding of the progression of the dyeing apparatus. 7.9 The constructed dyeing apparatus Following the application of time, load and temperature to the model it was discovered that the dye kettle is intensely subject to creep. The base of the vat is predicted to come into contact with the floor within the first few thermal cycles. Following this the base flattens at each dye run, but did not break during the simulation of 160 cycles. This is illustrated in Figures 7.23-7.35 and in the avi file of an animation of the dye vat dropping avaliable at https://doi.org/10.32028/9781789697421online. It was determined that if the dye vats did this in Pompeii they would eventually fail at the brick-lead interface, the point of contact between the lead kettle and the brick support in the apparatus. It may also be suggested that the lead itself would not break at the base of the kettle, despite the kettle resting on the base of the firebox. It may be further suggested that it is because of the additional support provided by the floor that the kettle did not break.

Following the simulation of two thermal cycles it was decided to simulate 160 thermal cycles. If the dyeing apparatus was heated and cooled once a day, 160 thermal cycles would be equivalent to six months use of the apparatus. If the apparatus was heated and cooled three times a day, this would be the equivalent to two months use of the apparatus. The simulation suggested that the kettle was under stress but that it would not fail during the first few thermal cycles. The simulation of 160 thermal cycles would allow a fuller examination of the nature of the strain and the possible method of failure. It is presumed that the Roman month was alternately 30 or 31 days in length and that the dyeing apparatus was not used for four days of every months due to holidays and feast days. While this figure may be amended, the simulation of 160 cycles would still allow

It should be noted that while the simulation provided evidence of strain, the strain did not exceed 4%. This meant that although the stress was great enough to

Figure 7.32. The dyeing kettle after 10 thermal cycles

140

The Finite Element model

Figure 7.33. The dyeing kettle after 50 thermal cycles

Figure 7.34 The dyeing kettle after 100 thermal cycles

141

Investigations into the Dyeing Industry in Pompeii

Figure 7.35. The dyeing kettle after 160 thermal cycles

produce strain that caused the vessel to deform it was not great enough to cause it to fail. The greatest strain was at the brick-kettle interface and at the most central part of the kettle base that was not supported by the floor. The central part of the kettle did not appear to be under a strain greater than 2%. This is probably because it was supported by the floor.

the results of the simulation as the elimination of the sharp edge at which the kettle failed may prolong its working life and the absence of such an edge on the remaining dyeing apparatus suggests that they had been amended to delay the failure of the kettle. This is an area requiring further investigation. 7.10 Summary

The photographs and description of each apparatus were reviewed. It was discovered that Vat VI in property VII xiv 17 exhibited behaviour that matched the simulation. This suggested that the simulation was of a level of accuracy that allowed a representation of what may have occurred during the working life of the kettle. It was also discovered that the supports for the kettles had the shape edge removed, regardless of the level of preservation and restoration. This supported

The results of the dyeing simulation indicate that the kettle did not fail in the manner expected: the lead was not as fragile as originally thought and so did not immediately fail, yet was so malleable that the kettle shape changed drastically during the first few thermal cycles. This raises the question of why the dyeing apparatus was constructed from lead. This shall be discussed in Chapter Eight.

142

Chapter Eight

Discussion (Coles, 1973:14), scientific experiments in archaeology have only really been undertaken since the 1960s. The experimental archaeology that has been carried out has been sporadic with a definite, although narrow, goal in mind. This was beneficial to each project and it fulfils the ‘scientific’ aspect – there was a definite aim and the aim was fulfilled. However, while each project may be beneficial to those involved with the project, it can shed little light on similar projects.

8.1 Introduction The understanding of the dyeing industry of Pompeii before this work had been purely archaeologically derived, mainly theoretical and with limited practical work to form a fuller understanding of the apparatus and their place in the industry. This work involved further investigation through the construction, use and amendment of a replica, a new and fully comprehensive survey of the remains in situ, an ergonomic assessment of the apparatus, and the exploration of the affect of the materials used through finite element analysis of a constructed virtual replica. This has allowed archaeological presumptions to be identified and challenged and unsolvable endemic entities to be defined. The findings of this study and its sequential progression through differing disciplines, the setting of this study and the parts within it in context and recommendations for further work are discussed below.

The closest experiment to the dye vat was not an experiment, but a piece of entertainment for the public, constructed and run by the Mary Rose Trust. This did not fulfil any scientific criteria, other than to see whether such a construction was possible, but it did make the archaeology involved accessible to the public, which was part of its aim. Therefore, while this experience was of little direct benefit to dye vats in Pompeii, the project fulfilled its role absolutely. It indirectly benefited the project, as it was possible to see how an unflued dye vat of that size would have looked and how quickly the fire led to the cracking of the wall as the ventilation had been insufficient. The closest project to the Pompeii dye project is currently being undertaken by John Edmonds of the Chiltern Open Air Museum. He is experimenting with dyes (mainly woad) and mordants to understand the chemistry and physical principles behind the process. The project is recognised for its scientific endeavours (Edmonds, 2003), but also benefits the public by making the knowledge of the dyeing industry and the chemical processes accessible.

8.2 This work in context It is necessary to place this study within its wider context to understand the significance of the findings. The nature of this study and the novelty of its approach shall be examined. The findings of this study shall be placed within the wider context of the discipline of experimental archaeology. The significance of the survey of the remains in situ shall be examined. The application of modern ergonomic theory to a Roman population and the use of this to establish a modern equivalent and to discern the usability of the dyeing apparatus shall be examined. The findings of each part of the previous work of this study shall be summarised and placed together to form a foundation for the later findings. The archaeological assumptions held at the beginning of this study and how each was challenged shall be explored. An examination shall be undertaken of the significance of the use of lead in the construction of the kettle and a comparison shall be made between Roman manufacturing and modern manufacturing systems theory. The findings of each of these shall be concluded and followed with a recommendation for further work.

The origin of this work fits with the original experimental nature of experimental archaeology (as typified by Heyerdahl, 1971), with only the later work being of a more predictable nature, (as typified by The Matthew Trust, 1997). The analogy of shipbuilding put forward by Crumlin-Pedersen (1999) describes the original construction: it may be assumed that each vat found functioned as a dyeing apparatus otherwise it would not have been found. Any vats that did not operate as dyeing apparatus would have been amended over time to a design that worked. Therefore the reconstruction of an apparatus, while experimental, was of a design that was believed to work prior to its construction. However, there were unexpected findings from the original reconstruction. It was expected that wood would not work as a fuel in the dyeing apparatus as the airflow would be insufficient. Wood burnt fiercely. It was then expected that charcoal would burn as fiercely. The charcoal did not work at all. The explanation for

8.2.1. Experimental archaeology Experimental archaeology is a relatively new branch of archaeology when viewed as a discipline in its own right. While investigations were undertaken to determine how artefacts were manufactured as early as 1860, 143

Investigations into the Dyeing Industry in Pompeii this was in the chemical composition of both fuels and the oxygen needed to ignite and sustain combustion. Bell stated that ‘It is difficult to take account of the role of chance and chaos in interpreting the results of such an experiment’ (Bell et al, 1996:243) a point that is valid for both the Overton Earthworks (Bell et al, 1996) and the dye vat. It depended on how hot the day was, or which way the wind was blowing, or how the wood was stacked as to how quickly the vat heated. As time progresses and these factors come under greater control this should be less of a problem, but at present they must be noted and allowed for in calculations.

The question arose during the study of the dyeing apparatus as the amendment of the apparatus to add a flue would have rendered it without archaeological precedent. However, to add the flue at the correct height for previously discovered artefacts would have required an amendment to the vat that would have caused it to cease to be an experiment (the results would not have been comparable). It was decided to continue the apparatus as an experiment. Each time the dye vat was altered to affected the thermodynamics and heat transfer with in it. This was expected and allowed for within the calculations. To add a flue to the side of the vat involved creating a hole in the vat, re-rendering the sides, and creating a layer of insulation layer around the hole where the flue was attached. A flue was not added until it was absolutely necessary to further the research.

In basic terms the experimental aspect of this investigation was to determine the output of the vat, through using a replica of the vat to create an output. As the method of using the vat and the physical principles behind its use was what was to be determined the design of the vat was the element important to the experiment. This meant (Mathieu, 2001) that only the design had to be replicated in its entirety. Although it was desirable to replicate the method for reconstruction and the materials used, it wasn’t absolutely essential so long as both the design was accurate and an allowance for alternative materials had been made in the physical calculations. Each material used had to respond thermally in the same way to the original materials used in the Roman construction as well as having the same physical properties. This substitution was acceptable for the aim and method of the experiment and so withheld Reynolds criticism that to not use the original materials degraded an experiment to ‘experience’ rather than experiment and so made it invalid, (Reynolds, 1999).

Each stage of the vat experiment has been carried out separately, reviewed separately, and the mathematics have been determined separately. This has been to avoid circular arguments or the application of a theory that fits neatly with one stage of the experiment to another stage with the expectation that it will also fit. Each part of the experiment has been determined and confirmed through experimental, literary and theoretical engineering means before moving on to the next stage. The experimental elements have been carried out before the engineering and some literary work, but this resulted in an experiment free from preconception – there wasn’t the subconscious urge to ‘make’ the experiment work or fail according to the results from the other strands of research, as can be seen by the use of charcoal in the brazier.

The dye vat project shares many points with the Overton work. The dye vat had clear objectives at each stage, but unlike Overton the aims of the dye vat project overall changed as the project grew. One of the main strengths of the dye vat project is its multi-disciplinary, and therefore triangulating viewpoint. This has the disadvantage that as each discipline sheds new light on the methodology of the vat, the aims of the experiment change to take in the new problems, ideas and enquiries. However, this has the advantage that at the inclusion of each new idea the findings may become both more accurate and precise.

The problem of non-repeatable experiments has been overcome in the study of the dye vat by constructing the vat in such a way that it should stand for years and as many experiments can be run on it as necessary for a full understanding of the procedure. Long-term projects can lead to unexpected findings as new tests and advances are made. The dye vat project started as a project to determine the methodology behind a dye vat and has evolved into a study of not just the times and consumables involved but also the physics and thermo-fluids principles behind how a vat operates and the placing of the apparatus within the wider context of the manufacturing processes of Pompeii.

Changes to the apparatus can not be made following the completion of its construction, unless they have been allowed for. This led to a conflict between ‘accuracy’ and ‘investigation’ – should the artefact be usable or faithful. This is alluded to by Dixon who states that although the building of the Telleborg reconstruction (a replica Iron Age meeting hall) was considered successful ‘it is, after all, still standing’ it was discovered afterwards that it wasn’t accurate, (Dixon, 1976: 63).

8.2.2 Definition of experiment In archaeology, an experiment may have differing definitions. An experiment may be successful in its physical outcome, which calls for the apparatus to stand for many years, such as the Overton Earthworks. Blockley argues that a successful experiment is one that tests ideas and generates ‘powerful memories and 144

Discussion

infectious enthusiasm that persists for life’ causing the undertaking of further work to explore the ideas raised, even if the experimental apparatus itself may have lasted less than a year, (Blockley, cited in Stone and Planel, 1999:32). In this context Blockley’s statement is suggestive, as while Blockley also states that the effect of exposure to experimental archaeology at a young age is unknown, the existence of this study of dye vats shows that it does indeed have some (admittedly unquantifiable) effect.

reviewed all the standing remains associated with each dyeing workshop and placed each in the wider context of supply routes, access, ventilation and water supply. Undertaking this survey meant that each theory formed about a dyeing apparatus was grounded in solid archaeological evidence, each replica or amendment to a replica was accurate and the simulation was based on a dyeing apparatus that had actually been discovered. Each theory stated about the supply of consumables or the disposal of waste was made with an understanding of the context in which the dyeing workshops operated. Undertaking the survey after the construction and use of the first replica (itself based on unpublished data from a review undertaken by Robinson and Janaway, 1994) meant that each dyeing apparatus could be reviewed and assessed with a practical knowledge of the dyeing process. This allowed the exclusion of ‘dyeing apparatus’ that had been identified in Moeller’s original survey but which could not have operated as dyeing apparatus.

Very few archaeologists have attempted to explain why something happens in a reconstruction. Speth just reported the physics behind the creation of stone tools and did not attempt to explain the terms to the archaeology public, (Speth, 1977). The majority of reconstructive archaeologists have reported the results of the experiments and then not explained the science behind what happened, (for example Coles, 1973). Cotterell and Kamminga attempted to explore and explain the physics behind a selection of archaeological artefacts, but chose such a sporadic collection that it was not possible to learn and understand the principles to apply them to other artefacts and findings, (Cotterell and Kamminga, 1990).

Through this survey it was also discovered that the digital plan constructed in 2004 from the aerial photograph taken in 2004 is unreliable. This may have implications of further work that relies on the use of this plan.

This work differs from previous projects not only in its inclusion of principles from outside archaeology but in its sequential nature. It is possible to follow the understanding of the artefacts and the examination of their context from the original classical archaeological through the different principles applied, the thermodynamics, experimental replica and through to Finite Element Analysis of a virtual replica. This results in an understanding of both the artefacts and the principles applied. This means that what would otherwise be unknown principles to archaeologists are presented in a clear way with physical demonstration. It is then possible to take these principles and apply them to other artefacts. The ability to take the findings from the project and use them in multiple disciplines to further understand each one is a new development. The provision of a working example from an archaeological background through the application of a number of techniques more usually associated with engineering in a way that makes each accessible to differing disciplines is new. The provision of explanation and example that allows the transfer of techniques to other archaeological problems is new. This is the first work of its kind and supersedes all previous work, which is fragmentary in its approach and nature.

8.4 Ergonomics This study was the first to review the Roman skeletal population, apply the data to a living dataset and modern ergonomic understanding and apply these findings to the Roman industrial equipment to assess its usability. Each step of this process, its use in this context and within the aim of this project was new. Combining these processes to produce a final outcome and the final outcome produced were both new. It has been discovered that the Roman population of Pompeii and Herculaneum matched the population of the United States between 1900 and 1960. It has been discovered that each of the dyeing apparatus were usable except for those that had been amended to change height (those in property IX iii 2), the unusability of which may be taken as further evidence of their inaccurate restoration. It was discovered that the replica dyeing apparatus could also have been constructed and used by the Roman population, which added to the accuracy of its use in this investigation. It may therefore be assumed that any conclusions made regarding the physical use of the apparatus withhold scrutiny as the apparatus would have been operable by members of the Roman population.

8.3 Review of standing remains

8.5 Summary of each section of work

The survey undertaken in 2002 was the most comprehensive survey undertaken since Moeller’s original survey in 1976, (Moeller, 1976). In some ways it was more detailed than Moeller’s as it specifically

Each section allows a better understanding of the influences and processes within the system and within the context of a living social, economic and 145

Investigations into the Dyeing Industry in Pompeii political system of a city. Dyeing was labour extensive, not intensive, although other parts of the textile manufacturing industry may have been, including possibly transport and supplies. Each section builds upon the last so each must be summarised here to give foundation.

• The addition of a flue does not alter the amount of fuel required. • The change to ventilation could not be perceived in either the fuel used or the time taken for the apparatus to heat. However, the flame was easier to control. • In addition, the idea originally put forward in 2002 that dyeing apparatus could be divided by size into vats that could take 1, 5, or 10 fleeces is inaccurate. The original idea arose due to inaccurate recording of the apparatus due to inaccessibility. • The amount of fleece dyed per person annually was: Minimum: 15 + 28 + 13 + 40 + 39 + 19 = 154 fleeces per day Maximum: 17 + 32 + 16 + 41 + 42 + 19 = 167 fleeces per day

8.5.1 Original work, the foundation of study The original study was published in 2002, (Hopkins, 2002). Following the use of an experimental replica of a dyeing apparatus it was concluded that: • Each dyeing apparatus could only have been used once a day. • The vats were discovered to fit into categories of 1,5,10 fleeces according to the size of vat required to contain a fleece and 90 litres of dye liquor. • Fuel used: 7.5kg of pine per dye run.

It would have been possible to dye between 154 and 167 fleeces a day in Pompeii.

7.5kg x 18650 kJ/kg = 139875kJ each dye run

There were 318 working days in Pompeii.

• The dyeing apparatus was only 20% efficient • Used a lid. • Charcoal could not have been used to fuel the apparatus. Either the fuel was wood or had a similar calorific weight to wood.



Minimum: Maximum:

154 x 318 = 48, 972 167 x 318 = 53, 106

Between 48, 972 and 53,106 fleeces could be dyed in Pompeii per year.

It concluded:

48,972 =

32,436 fleeces manufactured annually = 2.703 fleeces per person minimum. 12,000 people

12,000

53,106 = 12,000

42,612 fleeces manufactured annually = 3.551 fleeces per person maximum. 12,000 people

4.081

This may be rounded down to 4 fleeces per person.

4.4255

This may be rounded down to 4.4 fleeces per person.

Ryder states that the closest fleece to that of a Roman fleece (in terms of size, weight and make-up) is that of a Shetland sheep, (Ryder 1990). Therefore a fleece was taken as weighing approximately 2kg.

3.551 fleeces per person maximum.

2.703fleeces X 2 = 5.406 kg 3.551fleeces X 2 = 7.102kg

4.081 x 2kg = 8.162 kg

There were between 5.4kg and 7.1kg of fleece manufactured annually in Pompeii per person.

4.42 x 2kg = 8.84 kg

Therefore between 8.162 kg and 8.84 kg dyed fleece was produced annually per person in Pompeii. This may sound like a lot. However, in modern terms this is a ‘washing machine load’ (albeit a large one). It could therefore be reasoned that the dyeing industry of Pompeii was extremely small, possibly specialist, and not of a scale large enough to export, (as Moeller, 1976, had suggested).

8.5.2 Preliminary work The work completed in 2003 has already been published in conjunction with Watling (Watling, 2004) in 2005 (Hopkins et al, 2005). This work was undertaken following the reappraisal of the standing remains and detailed the use of the flued apparatus. It concluded:

Suppositions based on the conclusions so far include:

• Each of the apparatus could have been used by an average sized Roman.

• The availability of fuel could have been the limiting factor. 146

Discussion

• There was insufficient airflow through the apparatus to allow the complete combustion of charcoal. It may be presumed that fuels more calorie rich per gram weight than the wood used may have not burned completely. • The alternative design of flue was due to a difference in ventilation between the properties. Those properties with less through-flow of air contained vats with flues. • The dyeing process itself was not labour intensive. A workforce of between one and three people were required during the dyeing process, the most labour intensive parts being the set-up and dismantling of the apparatus. • The quantity of water used in the apparatus had the greatest effect on the amount of fuel required and time taken to heat the apparatus, (Watling, 2004).

was left undyed. It may be presumed that textile was redyed as this was cheaper than manufacture. It may be noted that some textile was left undyed. However, to define these quantities relies on the understanding of intangible endemic concepts such as fashion, cost and the social status of dyeing and re-dyeing and the rate at which re-dyeing would have occurred. It has been noted that while an estimate may be made of these quantities, a final determination may not be. This problem has been examined in Chapter Two. It is possible to argue that as a third of Pompeii remains unexcavated, conclusions drawn from current evidence may yet change as new evidence is uncovered and that intangible entities may be further explored in the future. However, further exploration is unlikely as the remaining third is buried in solid rock and has the modern town of Pompei built upon it. Therefore conclusions, while possibly inaccurate due to incomplete data, may remain unchallenged for a great period of time. It is therefore necessary to make the most accurate conclusions that the evidence allows, but be prepared to amend when new data come to light.

Following each of these previous studies it was concluded that the dyeing industry was not large enough to export. The use of dyed material as a decoration and not as the main body of some garments may have led to an underestimation of the figure produced, but the need to scour the textile and the use of scouring plants leads to the conclusion that this is not a vast underestimate. Moeller’s findings (Moeller, 1976) are still an overestimation.

8.7 The new findings from this study It may be noted that the introduction of a differing approach to the original 2002 work has allowed an examination of factors that were previously unappreciated.

8.6 Assumptions applied to the industry

• It was demonstrated by Watling (2004) that the material used in the construction of the surround had the greatest influence over the quantity of fuel required to heat the dyeing apparatus. This study was the first to examine fully the significance of the use of lead. It had been concluded that lead was an intriguing choice from which to construct a vessel that was to be exposed to heat, but its significance had not been fully explored. The influence of each material used in the construction of the apparatus has now been examined. • While the presumption was made that a dyeing cycle took 24 hours and that the industry worked at maximum output, the apparatus, workshop and dyeing industry as a whole had not been examined in terms of a system. There had also been no comparison between the Roman and modern systems of manufacture.

Following the preliminary findings of 2002, a set of assumptions were defined that were applied to an understanding of the industry. It is important that these are stated as they form part of the context in which the findings from this study are placed. 1. It was presumed that the dyeing industry was working at its maximum capacity. It may be argued that the minimum output of the industry could be zero at any point. Therefore determining a minimum would prove problematic. However, determining a possible maximum would allow a greater understanding of the industry and its significance and influence within the economy of Pompeii and Pompeii’s place within the Roman world. It is possible to attempt to determine a maximum maximum and a minimum maximum however, that is, to determine a maximum output within a range. 2. There are a number of unknowns that shall remain unknown until further evidence is discovered. It is possible that further evidence may not be discovered.

8.7.1 Significance of lead Experiment Four re-ran the dye runs of Experimental Three but this time with greater precision and with the inclusion of accurate recording of temperature and time. The same recipe and time was used as for Experiment Two and Experiment Three. The temperatures then

Such unknowns include the rate at which textile was re-dyed instead of being replaced and how much textile 147

Investigations into the Dyeing Industry in Pompeii allowed a theoretical reconstruction of the apparatus. This was the first time in which a dyeing apparatus had been reconstructed and the temperatures at each stage of use had been recorded.

although the FE model offers a method of failure as only one kettle appeared to be affected it is possible that other kettles may have been affected by a differing method of failure. However, it should be noted that there were other kettles where the base had malformed into concentric rings, suggesting creep as the failure method. It may be suggested that this effect was caused by exposure to repeated thermal loading followed by adjustment of the shape of the base (to return it to an approximation of its original shape) and that the only reason for the shape of kettle VI to differ was that its own shape had not been amended following heating. The effect of amendment to the base following heating is an area that requires further investigation. As there are so few intact kettles present in situ it is not possible to discern how each of them failed, so it is still possible that the method of failure of the majority of them was that indicated by the FE model.

The apparatus was deconstructed and replicated as part of a model generated using the ABAQUS Finite Element program. The system took the physical construct of the apparatus and the properties of the materials that had been used in the construction and allowed the changes to occur to each material under the influence of different temperature and stresses. The stresses were obtained from the replica apparatus and the remains in situ. The temperatures were taken from the physical replica. This meant that the virtual replica was grounded in the physical replica, itself grounded in the original remains. This was the first time that the actual temperatures of a dye run and the flow of heat through the system had been recorded. It also allowed the first replication of the dyeing process within the theoretical model.

The question remains that if lead was so readily deformable during use, why it was used. Lead has a melting point of 327oC, significantly below other metals. Other metals are stronger and are less subject to creep. However, there are advantages to lead. Lead was cheap and readily available as it was a waste product from silver processing, (Hodge, 1992). Its malleability meant that it could be transported in sections and made up on site, (Hodge, 1992). It was a mordant for dyeing which gave a brighter, stronger colour, unlike iron or copper which caused a duller colour in the finished cloth (see Chapter Two). This would mean that if the lead dissolved during dyeing there was a reduced need for additional mordant, thereby reducing cost and storage need, and there would have been no need to counter the effect on the material caused by the vessel used for dyeing. If the chemical advantages of lead were combined with the cheapness and availability it may be seen that there is a persuasive argument for its use.

Following the use of the Finite Element Analysis and modelling of the dyeing apparatus it was demonstrated that the dyeing apparatus was probably subject to intense creep, eventually likely to fail at the brick support. This raised a number of points. The photographs and records taken during the survey of the dyeing apparatus in situ in 2002 were re-examined. It was discovered that dye vat VI in property VII xiv 17 matched the FE model. It had been presumed during the survey that the shape of this kettle had been an anomaly and so had been ignored during previous surveys. The evidence from the FE model suggests that in fact this kettle exhibited the most likely occurrence and that in fact it was now the other kettles that presented a mystery. It was also noted that the kettle supports present in the majority of vats had rounded edges, whether or not they had been restored. It was further noted that only the supports that had been restored with a topping brick or tile had a sharp edge. This supported the evidence that the kettle failed at the brick support – it is possible that the edge had been smoothed to prolong its working life by reducing the stress concentration affect of a sharp edged brick support. The presence of smoothed edges in both the restored and unrestored apparatus also demonstrated that the restoration had been more faithful than originally thought. Prior to gaining a working knowledge of the dyeing apparatus it was not possible to discern the accuracy of restoration of the apparatus. Following the original reconstruction in 2002 it was possible to discern the attempts to preserve the fire boxes, kettle supports and flues in the original apparatus.

The malleability of lead was not so great a problem as it first appears. The archaeological assumption that the kettles were lead was only confirmed in 2005, (Monteix and Pernot, 2005). Therefore the question of its creep and malleability has not been investigated in any previous study. The main archaeological assumptions were either that the property of lead was unimportant, (after all, if it did not work the kettles would not be in place), or that the failure of the kettle would have been the limiting factor on production in the industry. The FE results suggest that the lead maintained a greater integrity than originally expected and that its malleability instead of presenting a problem would allow its continuation as it allowed amendment to the kettle shape that would prolong its working life. The evidence from the FE analysis is that the kettle would have been subject to creep during the first thermal cycle, but that there would be enough of a gap between the floor and the kettle to amend its shape. If it had

It may be noted that only a single kettle, VI in property VII xiv 17, showed signs of the method of failure indicated by the FE model. This may suggest that 148

Discussion

been left unamended the base of the kettle would have reached the floor during the first few thermal cycles. If the kettle had been allowed to rest on the floor this would have prolonged its working life as the floor would have supported its weight but the space within the firebox would have been greatly reduced. Given the Romans’ pragmatic nature, the malleability of lead and the relative cost for repair (in both downtime to the dyeing process and repairing the damaged apparatus), it is possible to suppose that they adjusted the shape of the base of the dyeing vessel between dye runs if it had been observed to warp. As the vessel would have been too fragile to repair once it had begun to rest on the floor of the fire box it may be supposed that either a support was added beneath the kettle during heating or that repair was undertaken rapidly once the requirement was recognised. It may be noted that the dyeing apparatus modelled in the FE analysis has a gap of 175mm between the base of the kettle and the floor. The majority of the dyeing apparatus had a greater gap, which would have allowed a greater room for leverage in repair.

number of people would have made the process faster as more apparatus could be prepared at the same time, and an additional dyer could assist with lifting, but once the dyeing has started only one person is required to ensure that the fire does not extinguish during the dyeing. For the single dyer (or untrained slave or child) to be able to monitor each of the apparatus they must be aligned in such a way as to allow monitoring to take place simultaneously. This alignment must also allow sufficient ventilation to each apparatus. It has been noted that each of the vats is arranged in each dye works in a U-shape or line. This would allow a single dyer to view each vat simultaneously and monitor and correct during the process. This positioning of manufacturing equipment and the arrangement of the work shops is used by modern manufacturing companies today, such as Toyota, (Monden, 1993). It should be noted that the Toyota production system is recognised as the state-of-theart production system which other manufacturers are attempting to emulate. This lay-out is used to allow a factory foreman to view each of the processes simultaneously as they occur in order. Toyota go further and link each of the U-shaped sections within the factory. This causes the production process to flow and eliminates isolation, (Japan Management Association, 1985; Monden, 1993). It may be argued that the Romans did this, but across a city not a factory. This could have been due to space constraints – they were fitting into a city recently destroyed through earthquake – or through intangible reasons such as economics or transport or tangible reasons such as water supply or storage. This is difficult to determine as it is still not possible to determine the ownership of the workshops or of the dyed textile at each stage of production. It should also be noted that, especially following the earthquake, the Roman dyers would have had to fit the location and lay-out of their dye works within the preexisting physical and socio-economic fabric of the city, (Hoffman, 1967:185).

It is unfortunate that while the method and place of failure may be determined, the timespan to failure may not. The FE analysis was run for the equivalent of six working months and yet the kettle did not fail in the simulation. Further work may allow the lifespan of the kettle to be established, particularly if the modelling is combined with empirically derived creep data for cast Roman lead. 8.7.2 Comparison of modern manufacturing systems to Roman dyeing It has already been noted by Edmundson (1989) that although the Romans did not possess the terminology of modern economic theory, they already had an empirical understanding of economic concepts and their application. It may be argued that there was also an empirical understanding of systems theory in manufacturing. While Roman manufacturing was preindustrial and so on a relatively small scale, it may be determined through examination of the dye works that there was a similar empirical understanding of manufacturing systems theory.

Groover, (1987), defines a manufacturing company as one that manufactures discrete parts, that is, single items with a clearly defined beginning and end to the process. Using this definition it may be seen that Roman industry consisted of manufacturing companies. The Romans did not possess the equipment or technology to form processing companies (the other form of manufacture, according to Groover, 1987), a company that processes liquid such as food or petroleum from original resources to finished product. The closest that Roman industry could have come to this was the movement of water through the aqueducts from source through cleaning to the eventual disposition on the streets of Pompeii or the production of chemicals that would be used in other manufacturing industries.

It was determined during Experiment Two and Experiment Three that although the size of the dyeing workshop could change, the number of dyers that would be required would remain at two or three or one with occasional assistance from a second. This was because the dyeing itself was not a labour intensive process once the apparatus had been set up. The dyeing apparatus would have to each be cleaned, filled with water and the fire lit, but there is only so many people that can fit around a dyeing apparatus. An increased 149

Investigations into the Dyeing Industry in Pompeii The term ‘company’ may be an inappropriate term regarding Roman industry as there is ongoing debate as to how the industry was organised. In this context the use of the word ‘company’ is purely as a name used for the organisational system as a whole within which manufacture of a particular commodity took place. While the term ‘workshop’ may be seen as inappropriate because it suggests production on a small scale, ‘workshop’ shall be used instead of ‘factory’. It is important to consider that although it appears through modern eyes that the Romans produced items on a small scale and in a pre-industrial sense, they did industrially produce items and were possibly the most efficient manufacturers that the World had seen to date. It shall however be argued that the term ‘factory’ is inappropriate, as although they industrially produced items, there was no mechanisation as would be understood in the modern sense. It shall be argued that any reference to the term ‘factory’ shall be in referring to any mechanised or organisational process devised subsequent to Arkwright’s Cromford Factory of 1771.

line within a workshop and the system and production line of the manufacture as a whole. The process of textile manufacture has already been discussed in Chapter Two. It may be seen that the process was linear, each part requiring the completion of the previous step. However, as each step was a discrete part it was possible to complete any number of products at each stage and store them until the subsequent part of the process could be undertaken. This would mean that the only limiting factors would be the number of previously completed items and a delay on the provision of consumables. However, a delay at one step would delay all subsequent processing. Within the dye workshop itself, processing was also linear. Each fleece had to undergo a number of steps toward dyeing otherwise each subsequent step could not have been undertaken. For example, a fleece had to be pre-mordanted before dyeing otherwise the dye would not bind to the fleece and the colour would not be fast, (Grierson, 1986; Grieve, 1992). However, these steps were closely related – the pre-mordanted fleece had to still be damp with mordant when it was placed within the dyeing apparatus and each step of the dyeing recipe had to be carefully adhered to. It was therefore necessary to plan the processing of each fleece in advance to ensure that the entire procedure could be undertaken without interruption in one step. At present it is not possible to determine whether scouring (cleaning the fleece) and pre-mordanting took place within the dye works, such information is an intangible endemic entity, but it may be suggested that it did.

The manufacturing system used by the Roman dyers would have been determined through empirical methods, (Hoffman, 1967:185). It is therefore unlikely that it would exactly match a modern system as it predates any standard or standardised theories on the ideal method of manufacture or the geographical or systematic lay-out of a workshop or procedure. However, it is possible to examine modern systems and apply the theory to the Roman system to determine how the manufacturing process was arranged and why. Roman manufacturing was industrialised but on a small scale (by modern definitions). The historic term is ‘cottage industry’ as the scale of manufacture meant that it could be undertaken in rooms in private houses as well as in specific workshops. It is important to realise that although through modern eyes all preindustrial revolution manufacturing was on a small scale, there was in fact great diversity within that scale of manufacturing. It is also important to remember that modern eyes will judge with hindsight, whether there is acknowledgement of this or not, and that it should not be forgotten that prior to the industrial revolution the larger scale workshops of the cottage industry was both the greatest amount of manufacturing seen up until that point and was also the greatest amount that was believed to be possible. The Romans were effective, efficient engineers, (Hodge, 1992): if a dyer believed that they could have cost-effectively increased scale of manufacture to out compete his rivals, he would have done so.

The textile industry in Pompeii sourced and processed raw materials into dyed textile and finished items. It is therefore possible to argue that the industry was a processor (processing the raw materials), a converter (converting the raw materials into the textile) and a fabricator (turning the textile into finished garments or items. It may be argued that the industry as a whole had vertical integration as all of the processes involved in the conversion of raw materials into finished products were present in the city of Pompeii. However, it is not possible to break down the industry as a whole into the horizontal division of processing or to state the economic or social relationships between each part of the industry as a whole. The ownership at each stage of production is at present an intangible endemic entity: only those present within the society at the time would know and those studying the society do not know as they are not a part of it, (Clarke, 1978). Therefore offering a horizontal division of the processes within the dyeing and textile industry as a whole is inappropriate. However, it is possible to define the physical places, the workshops, in which the processes took place and so divide the industry

It is important to distinguish that within Roman manufacturing there was the system and production 150

Discussion

within these tasks. Dyeing, for example, is a discrete part of the textile manufacturing industry with prerequisite processes that must be undertaken before the textile may be processed. A dyeing workshop is also recognisable in the archaeological remains. This is why dyeing has been chosen for further examination to allow an understanding of the role of the textile industry as a whole.

will be available at once and that any shortage of fleece will remain as a delay until the next season. It is possible that the fleece was stored and dyed throughout the year and that there was a continuous supply, starting at the beginning of the year with the need to store a year’s worth and then diminishing storage requirements as the year progressed. But it is also possible that the fleece was owned and stored elsewhere and brought to Pompeii only when in demand. This again is an intangible supply based on economics, ownership, transport costs and the physical storage of the items, all of which shall remain unknown.

The processing within a company may be defined by the types of production that the company uses to manufacture, (Groover, 1987). In Job Shop production the products are manufactured as a ‘one-off ’ or in low volume, and each is built to specific customer requirements. This requires that the production equipment is flexible and that the workers are highly skilled to perform different tasks, (Groover, 1987; Muhlemann et. al., 1992). In Batch production the company manufactures a medium size load of identical products, but the production rate exceeds demand. The purpose is to satisfy customer demand for a specific product. An inventory is kept and when the stock of that particular item begins to be depleted another batch is produced to those specifications. In this way there will always be a stock of whatever the customer specifically requires but the items themselves can be produced in an efficient cost-effective way (in a larger batch). The equipment used for manufacture is general purpose and designed for high rates of production, (Groover, 1987; Muhlemann et. al., 1992). The third means of production is Mass production, in which there is a constant high rate of production of identical products. There is high investment in equipment as the skills have been transferred from the worker to the apparatus. The complete workshop is dedicated to the production of this item, (Groover, 1987; Muhlemann et. al, 1992).

The method that the workshops employ is an (automated) flow-line (Groover, 1987). In this context it does not mean that automation of production was involved, instead it means that processing takes place at a series of workstations with a device (in this case a slave or freedman) that moves the item between the workstations. The unusual part to the linear nature of this manufacturing system is that it was possible to process products concurrently. While each step must be completed for a single item before it may be processed further, it is possible to process several items during one stage simultaneously (Muhlemann et. al. 1992; Hopp and Spearman, 2001). It is also possible to stagger the processing of the items, so that while one set are reaching the end of dyeing in one apparatus another set may be just beginning to heat in another dyeing apparatus. This meant that theoretically it would have been possible to keep a set of dyeing apparatus occupied while only needing a single person to attend them: If ten apparatus each took ten minutes to start and finish, they could be kept in cycle so that there was never a break in use.

It is difficult to define the mode of production used in the Roman world due to the intangible nature of ownership and the economics involved. At present the demand for textiles is unknown, so even if it were possible to determine how much could be produced the consumption rate would still be indeterminable. It is probable that the process of manufacture was determined through either a new, specific customer order, thereby making the process a job shop manufacture, or by a customer requirement from a batch that had already been processed. Which of these was actually the case would have been determined by the rate of demand and by the rate at which a customer could place a specific order that the dyeing workshop had not previously manufactured. Unless records are subsequently unearthed, these are intangible demands.

There are advantages to using an automated flow line. Within a workshop single process distances and time between stages may be minimised as may be the number of staff required. Within the system as a whole it is possible to put a specialism in at any stage in production and to integrate production as a whole. It is also possible to run a process concurrently if there would otherwise be a surplus of an earlier stage, (Groover, 1987; Muhlemann et. al., 1992; Hopp and Spearman, 2001 ). It may be observed that the design of workshops allowed this through the division of the process according to tasks. The movement between workstations within a process or between processes within the system of Roman textile manufacture may be described as ‘asynchronous transfer’, (Groover, 1987). This method of transfer has several advantages to the dyeing system. In this system the items are each moved to the workstation for the next

A further consideration is the seasonal nature of textiles in the pre-industrial world. All sheep will need shearing during the same season. This will mean that all fleece 151

Investigations into the Dyeing Industry in Pompeii process when they are completed. Each workstation works independently of other parts and only depends on the previous completion of preceding items within the system. This method can be used to compensate for line balance problems: If a preceding item may be completed within half the time of a subsequent process it is possible to complete twice as many and then split them between two parallel subsequent workstations. This method of transfer also allows the storage of surplus parts to be processed. In this way breakages or shortages of consumables in the system would not delay any previous processes, instead it would also allow for a build-up of part to be processed.

By considering the system of manufacturing and the processes within it it is possible to determine the downtime at each stage and within the system over all. This could have been influenced by a lack of consumables or the failure of an apparatus. Depending on the ownership of dyed materials downtime may also include time taken to re-dye anything that was unsuccessful. The analysis of flow line performance in a modern factory may not be relevant to understanding the process of production in the Roman world. To determine the average production rate it is assumed that there is synchronous transfer between workstations and that time spent at each workstation is the same. In the unmechanised dyeing this is not the case. Analysis through considering the amount of time that the production line is operating (as opposed to downtime) is also irrelevant as each dyeing apparatus within a workshop and each workshop operated physically independently to each other. There may have been dependence through economic considerations, prerequisite processes and the considerations of supply and demand, but these are intangible. However, for a maximum output to be calculable it shall be assumed that they were not limiting factors. It is still possible to determine the amount of average time that an item should take to process from start to finish, as it is possible to determine the length of time that each process should take. Due to breakdown in supplies and equipment the process will on average take longer than the determined time. Therefore it is possible to determine a maximum outcome but it is unlikely that this was ever realised.

Cycle rates are usually slower in systems that use asynchronous transfer than in other systems, (Groover, 1987; Hopp and Spearman, 2001). This is not important in the context of the dyeing industry in the preindustrial world. The other methods (continuous transfer and intermittent transfer) are impractical in the context of Roman dyeing. Also, the method used was the most efficient method available and with the allowance of time for the completion of the process. The application of modern transfer mechanisms (methods to actually move items between workstations, not the actual time in the system for the move of completed items) is inappropriate to the Roman industry. The industry did not have automation in the modern understanding of the word, but instead relied on the muscle power of slaves. Any items were therefore carried between workstations or transported between workshops or storage areas through either lifting or removal with a cart.

The Romans had an empirical understanding of processing the fleeces in a manual assembly line using multiple workstations. However, they took it one stage further by separating each stage of production horizontally and so grouping all similar tasks together. It may be argued that as Roman manufacturing was pre-industrialised in fact they had done the reverse and hadn’t yet grouped the tasks apart. However, as the original industries would have operated on an even smaller scale it may be argued that they were brought together to pool resources for each stage of production: the first part of industrialisation in an industry.

It is possible that the dyers possessed an inspection stage in manufacture with an associated feedback loop. This would have taken place at the end of each stage of manufacture to allow an assessment of whether the item was ready to proceed to the next step of processing. The final inspection would have been before either sale or return to the client. However, the action following inspection may be arguable as it would depend on ownership of the item at each stage of process. If a client owned an item that needed to be redyed and the dyeing didn’t work, it may be possible that the dyers had another attempt. However, if the item was due for sale, it may be possible that a missmatched colour was deemed to be natural variation and that the item was put for sale anyway. A failure in a preceding stage may have been removed from the system, amended or allowed to continue, again with the argument of natural variation. Ownership at each stage of dyeing is currently unknown, so although it is possible to speculate as to the action taken if an item failed inspection during manufacture, it is not possible to say definitely what would have occurred.

The sequence of production of an item is limited by ‘precedence constraints’, that is the order in which tasks in production must be completed. Each part of the process of manufacture will take a differing amount of time. The workstations are divided into sections according to the time taken to complete each series of tasks. Line balancing is the arrangement of workstations into sections so that each section takes the same amount of time to complete a series of tasks. If a section isn’t balanced then there will be a delay in the 152

Discussion

system as the slowest part of production dictates the production rate overall, (Groover, 1987; Muhlemann, et. al., 1992; Hopp and Spearman, 2001) In the Roman method of production this was overcome with the use of storage buffers and parallel processing. While it is possible that ten dye vats in a dye works may have been used simultaneously, the source of fleece was seasonal, so a steady supply of fleece to a dye works and to each subsequent production process depended on sufficient storage and transport before dyeing and between each stage. The modern term of line balancing would not have been used, but the design would have been developed through theories derived from an empirically derived understanding.

horizontal division of the full processing of the fleece. Each task is then divided into work elements. This further division allows specialisation at each point. ‘By giving each worker a limited set of tasks to do repeatedly, the worker becomes a specialist in those tasks and is able to perform them more quickly and more consistently’, (Groover, 1987). In the textile manufacturing industry this would mean that dyers became extremely specialised. It may be argued that this was necessary as dyeing is precise chemistry, and these dyers had to dye without the aid of thermometers or pH meters. This assertion is supported through contemporary ethnographic studies of pre-industrial potters, who manufacture without the use of modern sensing equipment. The potters are skilled in the choice of materials and the manufacture of the product, (Vitelli, 1984).

The order in which a process must be undertaken is not the only constraint on the order of manufacture or the location or sequence of workstations. Zoning constraints may exist, either positively or negatively. A positive zoning constraint would be the construction and use of all dyeing apparatus together as they are specialist equipment and use in one location means that dye spillage will not effect any other part of the process. A negative zoning constraint could be the isolation of dyeing so as to not cause ruination to any finished fabrics. There should not have been any position constraints caused by the size of the dyed material, however, the size and needs of apparatus could have caused constraints in designing the lay-out of the workshops and the addition or amendment of the apparatus.

In modern manufacturing there are two ways of transferring products between workstations. In mechanical lines the products are moved along with a conveyor belt. This would not have happened in the Roman dyeing industry as the technology did not exist to do this and the distances between workshops were too great. In non-mechanical lines the items are passed along by hand. This is the method that would have been used by the Romans. However, this can result in starving or blocking of stations as each relies on the completion of the previous task. Balance delay is the loss of efficiency caused by one section of the line waiting for the other section of the line if the line isn’t balanced, (Groover, 1987; Muhlemann, et. al., 1992; Hopp and Spearman, 2001). The Romans would have been concerned over line balance as they would have wanted the most efficient use of materials and labour. However, they would have accepted that the process had to happen in a specific order and that it would take as long as it took. The chemical reactions involved could not be speeded up. This would have meant that each process would take as long as necessary and the person manually handing the item on to the next person after processing. This is the method that the Romans would have used as you can not hurry a chemical reaction or the laws of physics and there was no other method to transfer an item other than manually.

The Dyeing workshops were set out in a process layout, (as defined by Groover, 1987; Muhlemann, et. al., 1992). This means that the production line was separated according to process. It may be argued that the Romans did not possess the technology to undertake a modern product-flow layout, where the item is processed in one continuous flow. However, dyeing does not lend itself to this method of manufacture. The textile manufacturing process was already divided horizontally according to process with zoning defined by the equipment required. The tasks within the production process were separated into workshops through the city. This breakdown may have been through lack of space and economic considerations such as rent and consumables supply. Furthermore, within each workshop the apparatus were independent, allowing the use of differing apparatus for different procedures simultaneously. The lay-out of the apparatus is of interest. As has already been discussed in this section, the Romans empirically derived the workshop arrangement used today by Toyota and other companies to allow the simultaneous monitoring and amendment of differing stages of the process, and to minimise the number of dyers required, (Japan Management Association, 1985; Monden, 1993).

Automated control is one way in which to improve line balance, but this was a method unavailable to the Romans (as it relies on electricity). Dividing work into elements, inventory buffers between stations, and parallel stations are all methods that can be used, (Groover, 1987; Muhlemann, et. al., 1992; Hopp and Spearman, 2001). The Romans used each of these methods in manufacture, combining them with flexible manual assembly based on the chemical reactions and the horizontal division of tasks located through zoning of equipment.

In mass production work is divided into small tasks. In the textile manufacturing process that would be the 153

Investigations into the Dyeing Industry in Pompeii If a line of production moves ‘too fast’, the workers have difficulty in processing the items, they become stressed, complain and may even sabotage the line to slow it. If they are allowed to control the rate then it improves. In the Roman version the slaves would get no say officially. However, if the slaves were seen as sufficiently important to the process or had technical knowledge that the master required it is possible that they were accorded respect and that their views on the speed of the process were listened to. At the very extreme, if they were replaced it would be expensive and if the process continued not to work the master would eventually realise that there was a problem. The time this would take and the waste in resources that this would take would depend on how reasonable the master was or how familiar he was with the process. This is another intangible concept.

1. Raw materials 2. Equipment 3. Tooling and Fixtures 4. Energy (electrical energy) 5. Labour There are two outpus: 1. 2.

The completed workpiece Scrap and waste

The production of scrap and waste is determined only after the final process as it is possible that the item could be ruined during the final process. In the Roman textile, or more specifically the dyeing, industry, the inputs vary from the modern list. 1. Raw Materials: sheep fleece, mordant, dyestuff 2. Equipment: Dyeing apparatus, kettle, 3. Tools: buckets, poles (stirring) and a means to empty everything 4. Energy: Slaves, fuel 5. Labour: Slaves, paying freedmen. Whether freedmen were paid is debatable, (Mouritsen, 2001).

Assembly techniques evolved when humans were the only assembly machine available. But people are expensive to run. While the Romans would not have been concerned with paying a wage, there was still the running costs of food, clothing, shelter and downtime through sickness. ‘Human beings are the most dextrous and intelligent machines’. If a human were to be replaced the machine would have to have been quicker, more accurate, more flexible and have the ability to problem solve. The Romans never possessed this technology in the dyeing industry.

There is also a variety with outputs: 1. 2.

8.7.2.1 Inputs and outputs

The completed workpiece Scrap and waste

The completed workpiece may be the textile at any stage of manufacture. The scrap and waste could be reduced by redyeing or increased if the dyeing did not work exactly as required. The frequency of this occurring is unknown.

It is possible to argue that the Roman textile industry contained within it all stages within manufacture as defined by Groover, (1987). Processing is the collection of raw materials and rendering them into a usable state in manufacture. In the textile industry this would be the collecting and drying of dyestuffs, the collection of mordant and the collection of sheep fleece. Assembly would have been the use of these materials in the manufacture of the textile. It is fairly straightforward to understand the role of these two steps in the fleece processing, however the role of handling and storage causes problems. The storage used in the dyeing industry could have been within the workshops themselves, or somewhere within the city, or somewhere outside the city. As the only requirement for storage would be space with a roof in which items could be stored away from the elements it would be difficult to determine where items were stored and how much storage (in terms of physical size and location and the timespan involved) would be required. Inspection and testing and control would all have been one step and any faulty items would have been either discarded or an attempt made to amend them.

The lack of electricity may not have had a detrimental effect on the dyeing process. It should be remembered that the process evolved without electricity. It may therefore be argued that it will already be as efficient as possible, using natural light and positioning of the working day to increase this efficiency. A lot of modern ideas can not be applied to the Roman world without consideration of context, for example transport by mechanical methods that involve electricity will be inappropriate. 8.7.2.2 Buffers The Roman manufacturing system was determined empirically and was the most efficient that they could make it. Line stoppages were overcome with the inclusion (whether deliberate or not) of storage buffers. A buffer is used to break up a production line and a storage area may be placed within it. Each part of the line processes items from one storage area and places

According to Groover, (1987), most manufacturing processes involve five inputs: 154

Discussion

them in storage in the next area. In a modern factory, unless these buffers are in place, if one part of the production line breaks down, the line grinds to a halt as the workstations after the breakage are starved. If the buffers are in place then only the section of the line that contains the breakage stops manufacture; every other section continues to work as those below the breakage feed into the store and those above use older parts from the store to continue, (Groover, 1987; Muhlemann, et. al., 1992; Hopp and Spearman, 2001). In Roman dyeing each process was split into workshops. Therefore a stoppage in one workshop would not impact on any other workshop unless there was insufficient storage to allow the preceding workshop to continue. This meant that the production process was as efficient as it could have been – production at every other level of the process would continue. It may have been possible that the lead dye vat fractured during dyeing. This could result in the temporary loss of apparatus, a loss of the textile and possibly a temporary loss of a slave to scalded feet. Furthermore, each apparatus operated independently within each workshop. If one apparatus failed, the others would continue to operate.

would be a need to store each batch until either sale or requirement (depending on ownership of the raw materials). If the job-shop method were employed, there would not be a need for storage of finished products except until collection by the customer. At present the ownership of the fleece at each stage of manufacturing is unknown. It is possible that the dyers owned the materials or that the dyers were subcontracted to by the owner of the materials. While it is advantageous to store materials and completed products for as little time as possible due to space and money constraints (King-Scott, 1971), at present the storage requirements and location in Pompeii must remain unknown. This is not to detract from the other needs for storage of the production industry. There is a need for storage of all raw materials. There is also further consideration that the dyes should not be allowed to get damp or to rot and that the fleeces, once scoured, should not be allowed to become dirty. This will affect where storage can take place. If job production is required it may be that there is minimal storage as each component could be ordered in as needed. For example, there is no point storing a batch-load of woad if the requirement is for madder. Both dye stuffs degrade and there would be the risk of loosing either. In batch production a new batch could be made depending on what was selling best. This again would require storage to hold just below the amount required of each item for a batch. The difficulties in storage would be in placing degradable items, in keeping scoured items clean and in storing the dyed textile prior to finishing or sale. If the materials were owned by a customer prior to processing then the finished items could be delivered alleviating storage problems. If they were not or the items were to complete the manufacture process, then it may be possible to take them straight to the fullers, again alleviating the need for storage.

It should be considered however that as each workshop was dedicated to a single operation, an entire level of a system could have ground to a halt for want of one consumable. While slaves did not have to be paid and, depending on contract, freedmen didn’t require payment either, rent and other outgoings would still have been and there would be demand for items further along the process. However, having said this it may still be argued that the horizontal division of the process into workshops resulted in the most efficient way of processing. According to Groover (1987), ‘there are other reasons why line stops occur which are not directly related to workstations… These other factors have to be taken into consideration’. These may be planned for, such as a need to change tools or replace a worn part of a machine. However they may be unplanned, such as a jam in a machine. In Roman dyeing these may also be planned or unplanned. It is possible to plan that there will be time taken at the end of each dye run to clean all of the equipment. It is even possible to plan for occasional occurrences, such as repair of the lead kettle. However, dropping the fleece or having the transport cart stolen are events that may not be planned for but which would also have to be taken into consideration.

8.7.2.4 Information Process Cycle The Information Process cycle (Figure 8.1) is also an intangible part of the industry. The Roman Empire was the most bureaucratic seen before the modern day, and there are records showing the development of a complex economic and legal system, (Temin, 2001). However, trying to understand the system of this economy has been likened to trying to understand the system of processing of the London Stock Exchange by only reading the Financial Times: It will be possible to discern day-to-day occurrences, but not to develop an understanding of the framework in which it occurred, (Temin, 2001). Therefore, while it may be argued that the Roman dyers did receive orders and produce textile for customers, the framework of planning and information transfer in which this occurred currently

8.7.2.3 Storage However the order for manufacture happened would determine the need for storage for the industry. If the batch method of manufacture were employed, there 155

Investigations into the Dyeing Industry in Pompeii horizontally so each workshop was associated with one process. The U-shaped layout in the workshops was just a geographical distinction to allow the overseer to monitor all apparatus at once and to reach each with the minimum time and effort.

Figure 8.1. The Information Process Cycle, from Groover, 1987.

Toyota specialise each task and change tasks to alleviate boredom. It was not necessary to do this in Pompeii as the majority of workers were slaves. However, there was specialism within the textile manufacturing industry as each process was divided and required the necessary skills. It may be supposed that there must have been some form of variation otherwise the records of lead poisoning (written by Vitruvius) would have mentioned the dyers as being at high risk.

8.8 Conclusions

remains unknown. Our only understanding of the process may be discerned through the physical remains, the occasional record of a transaction or legal edict and the application of logical deduction based on the knowledge that the Romans were efficient and aware of the importance of legal redress.

This thesis has eleven main conclusions: 1. The dyeing industry was smaller than Moeller originally anticipated. A maximum of 8.84 kg of fleece was dyed per year, per person. While a slave may not require a substantial amount of clothing yearly, this figure also includes soft furnishing for public buildings, so it may still be concluded that the industry was smaller than Moeller originally thought. 2. The water used in the dyeing process was hard water taken from the public fountains and disposed of in the street. 3. The wool was dyed ‘in the fleece’. The most common dyestuff used was madder, the most common mordant was alum. 4. The dyeing apparatus were each operable by an ‘average’ Roman, except for the three that had been altered (vats I-III, property IX iii 2). This may have been taken as further evidence of their alteration. Dyeing apparatus had been amended with steps to allow operation by the ‘average’ Roman or by children. Such steps were found on vats III, IV and V in property VII ii 11. 5. Dyeing may have been an unpleasant occupation, but the dyers were ‘better off ’ than originally supposed. They were economically valuable

If it was supposed that there was a limitless supply of consumables in the stages preceding and proceeding the dyeing stage of manufacture it is possible to determine the cycle time of dyeing and the influences of each factor within dyeing. This may then be placed back into the context of the manufacturing system as a whole and the system further examined. When creating the manufacturing area as a U-shaped it is necessary to be wary of isolation. The Romans avoided this by breaking the processes down horizontally. Each stage was buffered as it operated separately so each stage could build a sufficient store for processing. The isolation was not geographic so there was no breakdown in communication. Buffers between levels of production enabled each other stage to continue if one halted. It should be noted that Toyota run a factory floor where the flowline is along a U-shape so the overseer can view all processes at once. This is not how the Romans operated. Their processing was divided 156

Discussion

6. 7.

8.

9.

10.

11.

and so care was taken of the trained dyers. Dye works could operate with only two-three people and so was not labour intensive. Unfortunately the value of some dyers and any preferential treatment may have been detrimental to the social conditions of others, although some were able to gain freedom and a raised social status. The dyers did not succumb to lead poisoning and there was no lead in the drinking water or water used for dyeing. The survey undertaken as part of this study is the most comprehensive survey to date of the dyeing equipment in Pompeii. It supersedes Moeller’s original survey and even corrected misidentifications. This allowed a more accurate understanding of the remains and thereby the industry. The lead kettles did not fail in the way predicted through archaeology. In fact, the lead was advantageous and the material of choice. Whether the kettle was amended to retain its shape, supported during heating or allowed to rest on the floor of the firebox is unknown, but the working life of the kettle was greater than previously expected. The simulation has demonstrated that it is possible that the kettles did not fail. It is possible that following repeated heating the base rested on the floor, which allowed sufficient support to prolong their working life. It is also possible that the base was amended following heating to retain the original shape. It may be argued that only lead allows this amendment and this flow of the material without breaking. This work is sequential allowing the evolution of the project to be followed. This means that it provides a working example not just of the background to an archaeological problem, but the application of a number of techniques more usually associated with engineering in a way that makes each accessible and transferable to differing disciplines. This is the first work of its kind and supersedes all previous work, which is fragmentary in its approach and nature. The use of lead, the use of a method now used by Toyota, the understanding of economic theory, the understanding of chemistry and the physical properties and needs of the dyeing process, all indicate that the Roman civilisation was even more advanced than originally thought.

a more reliable model. At present the data for creep of lead above 40oC was derived through extrapolation and so its reliability should be tested. 2. Following the construction of the simulation it will be possible to expand it from 2-dimensions to 3-dimensions to determine the affect of stress throughout the apparatus. It may also be possible to determine the time it would take and the number of thermal cycles for the kettle to fail and the method by which it failed. It may be supposed that it failed at the lead-brick interface, but further investigation is required to confirm this. 3. To test the theoretical simulation it is possible to construct a physical model from lead to determine the accuracy of the simulation. This may be used to amend the simulation. 4. It would be possible to test the effect of amending the base following the first thermal cycle. If the base was amended to retain the original shape of the kettle before heating this may prolong its working life. 5. The findings of the physical reconstruction, the lead model and the simulation could be applied to the other dyeing apparatus of Pompeii. Once the simulation has been demonstrated to work, the geometry used in the simulation can be amended to simulate the other 39 dyeing apparatus in Pompeii. 6. As only a single dyeing kettle was discovered with a base that matched the simulation, there should be an investigation into the bases of each of the other kettles, as it is possible that they were affected differently. It is possible that the unaffected dye kettles were used for cold dyes and so were unheated, but this requires further investigation. 7. The rate at which the lead would have dissolved from the kettle into the dye liquor is unknown. It would have depended on the content and pH of the dye liquor. It may be supposed that the lead present would have had some effect on the dyed fleece, but at present that effect has not been investigated as it is beyond the scope of this study. The effect may be determined through further work in the understanding of the effect of the dye liquor on the lead kettle. 8. It should now be possible to determine the size of storage required in the city for the consumables and the finished fleeces. Unless the dye workshops of Vi4 and Vi5 possessed additional storage to that on-site, the storage may indicate a possible output total, or even limitation, of the dyeing workshop. 9. It may be possible to determine the amounts of consumables, such as water and fuel, used in the dyeing process.

8.9 Further work There are ten areas recommended for further investigation. 1.

Experiments regarding the creep of lead should be undertaken to gain the data necessary for 157

Investigations into the Dyeing Industry in Pompeii 10. The scale of manufacture and the possible output of the industry have been determined. However, the quantity of textiles required by each inhabitant is still unknown as it relies on intangible factors

such as economics, fashion and personal taste. To apply the quantity of textiles manufactured to the quantity of textiles required would require further work in discerning these intangible factors.

158

Glossary The terms below are grouped to allow comparison between each. mortar is of a right angle at 45o up from each brick edge, hence a point. The process ensures the wall remains weather-proof and prevents further damage.

Definitions: Scale of manufacture: the amount of items that may be produced. A tangible, objective figure that may only change after future excavation or revision.

Render: The layer of mortar placed over the external surface of the dyeing brazier. This served a dual purpose: it hid that the brazier was constructed from rubble and assisted in maintaining structural integrity. Rendering the vats also assisted in the dyeing process: the smooth surfaces did not catch the fleece as it was manoeuvred and allowed the vat to be rinsed down after use.

Economic significance: the importance of something, such as the scale of manufacture, within the economy. An intangible, subjective figure that may change through interpretation. Definitions – Engineering:

Restored: Some dyeing apparatus had been restored since their original excavation. In this context this means that they had been re-rendered, re-pointed and in some instances reconstructed. When originally surveyed the extent and reliability of this restoration was unknown. However, it was noted that an attempt had been made to faithfully reproduce any feature, even though it was possible that the purpose of these features were unknown at the time of restoration. When exploring how the vats had operated, unrestored vats and features were used.

Hydrostatic pressure: The increase in pressure that a liquid is subject to the deeper within a container that it is due to the additional weight of the liquid above it. Consumables: Items used during a manufacturing process that are necessary for the manufacture, by either allowing the process to take place (such as fuel) or becoming part of the finished product (such as dye). Definitions – Archaeological: Amphorae: Terracotta container used to transport and / or store perishable goods.

Taphonomy: The natural break down of artefacts. In buried archaeological artefacts this is due to microorganisms in the soil in which they are buried. In Pompeii this particular mechanism of break down was not present, so artefacts broke down differently, broke down following discovery, or remained preserved.

Endemic entity or concept: An idea that may only be known within the society that holds the idea due to its intangible nature. Such ideas should be allowed for when interpreting finds or the social constraints of a society.

Computers

Harris Lines: Lines of increased bone density that represent the position of the growth plate within the long bones at the time of arrest in development, due to insufficient nutrition or another factor preventing growth.

FE: Finite Element Analysis. The process by which an article is broken down into smaller parts so that changes that take place within it may be examined. Each part is an element and as each part is discrete these are finite elements. Each finite element is represented as an equation. The analysis is the examination and combination of these equations to understand how they relate and how a change occurs within the material, such as the movement of temperature or force.

Lararium / Lares: Lares were the household gods, worshipped at the household shrine, the Lararium. The shrines are of significance in understanding Pompeii as they were used to make requests to the gods by way of depiction. This means that everyday scene were depicted with clarity and an expansive palette, ranging from Vesuvius before the eruption to everyday dress and activities.

ABAQUS: a computer program designed c1983 to undertake finite element analysis. In some areas this program has been surpassed, but it has all of the different processes necessary in this analysis. Prior to the development of this program it was necessary to perform calculations for each stage of analysis, including analysis of the effect of temperature and changes to the structure of the material due to creep.

Pointing (subsequently ‘to point’, or ‘re-point’): To renew the mortar between bricks or stones within a wall through cleaning or removal of damaged parts and re-application of mortar. The shape of the renewed 159

Investigations into the Dyeing Industry in Pompeii Dyes:

an individual’s weight warrants further investigation. However, it is also used to gauge the possible build of an excavated individual based on their height and signs of activity prior to death.

Free movement: When material is dyed there must be enough water in the receptacle to allow it to move freely, otherwise it would dye unevenly. However, there must not be too much water, otherwise the dye would be too dilute. The amount of water required for one 2kg Shetland fleece is 90 litres.

Weight (kg) = BMI Below 20 Height (m) 2

Indigo: A dyestuff used to create a blue dye. Originally from India.

20-25 25-30

30+

= Underweight =Desirable weight

= Overweight = Obese

It has been noted that this is an unreliable measure if a person is particularly active or has an ancestry that results in a larger build.

Indigotin: The chemical name of the blue dye found in woad and indigo. Chemically this is the same dye, so when indigotin is found it is not possible to discern whether the dyestuff was woad or indigo.

Standard Distribution: An article designed for use by the general public is designed for the requirements of the mid 95% on a bell curve.

Madder: A dyestuff used to create a red dye. Contains purpurin and alizarin, a red and a purple chemical dye. Differing quantities of these give the overall dye its red or purple shade. Age, storage and pH of the dyestuff when heated cause differing amounts of alizarin and purpurin to be present in the overall dye. The dyestuff is the dried and ground root of the plant. This may be stored, but over time the relative amounts of alizarin to purpurin contained within the stored root change.

Genetic: Genotypic – relating to Genotype: The genes that a person carries that go towards the qualities of that person, such as height. A person can carry the genes but not express them due influences in their environment. For example, a person may have the genotype to be tall, but not have received a diet that allowed growth resulting in a phenotype of being short.

Mordant: a chemical, usually a metal salt, included in the dyeing process that causes the dye to bind to the material that is being dyed. Usually the material must be ‘pre-mordanted’ prior to dyeing, treated with the mordant to allow the later uptake of dye.

Phenotype: The expressed genes of a genotype. For example, a person may be short because their genes are for a short person. Or they may be short because while their genes are for a tall person they have not received the nutrition to allow their growth. Phenotype is how a person looks.

Murex / Imperial Purple: Dyestuff derived from the murex mollusc. Extremely rare and expensive it may take 12,000 whelks to produce 8g of dye. Chemical name: Di-bromo-indigotin. The only chemical difference is the additional two bromine atoms on an indigotin molecule. Di-bromo-indigotin is unstable and oxidises to indigotin as it loses the di-bromine.

Genetic Drift: The process over time of the natural change within a population due to differences of genetic make-up between individuals and the inheritance of their resultant offspring.

Scouring: Cleaning a fleece to remove the lanolin and other dirt prior to mordanting or dyeing. Without scouring the mordant or dye would be unable to ‘stick’. Scouring took place using urine as a detergent during the Roman era. In this study commercially available detergent was used.

Textiles: Dyer: A person that dyes either fleece, yarn or fabric, prior to further processing. Fuller: A person that ‘finishes’ fabric. Fulling of wool causes the fibres to mat, thickening the wool fabric, causing the fabric to become wind and water proof.

Vat dye: a dye that does not require a mordant and so may be used directly to dye.

Gamma patterns: Named after the Greek letter that they resemble, these were patterns made in fabric by weaving coloured thread into a plain, undyed background.

Woad: A dyestuff used to create a blue dye. Originally from northern Europe. Ergonomics:

Clavi: Straight lines woven into a fabric to pattern it, these were usually made by a dark thread being woven into a plain, undyed background.

BMI (Body Mass Index): A measure of relative weight to height, used as a benchmark to determine whether 160

Glossary

Places:

Sheep:

Fayyum: Also spelt Fayum and Fajum. Egyptian archaeological site, famous for the preservation of funereal face masks and textile remains from the Romano-Egyptian period.

Shetland. White-woolled domesticated sheep weighing 25-40kg depending on location. Fleece approximately 2kg. The size and build of the sheep, and the size and type of wool (including hair/wool ratio), mean that this sheep is equivalent to the Roman sheep.

Mary Rose: Flagship of Henry VIII’s fleet, which sank in the Solent on its maiden voyage in 1545. It was raised from the seabed in 1982 and is currently being preserved by the Mary Rose Trust in Portsmouth. The preservation of artefacts on board was almost complete due to its burial in silt.

Soay (Ovis aries). Brown woolled primitive breed of sheep, weighing 40-50 lb, 18-23kg. Equivalent to Bronze Age British Sheep. Smaller and hardier than modern domesticated sheep. Measures

Vindolanda: Romano-British fortification on Hadrian’s wall. Famous for the preservation of textile, leather and written remains, including letters posted to soldiers from their mothers. The preservation of these organic remains was caused by their deposition and rapid burial in an anaerobic and waterlogged environment. This has resulted in a high number and variety of remains being preserved.

Within the photographs of the dyeing apparatus varying measures are used. The larger of the two is a 1m length, divided to 0.5m lengths, the smaller is a 0.2m length, divided to 0.1m lengths.

161

References Boyle, J.T. & Spence, J. 1983 Stress Analysis for creep. Butterworths, London. Bray, P. 2004. Choosing and justifying our conclusions: ‘Inference to the Best Explanation’ in the context of the transition to iron based economics in Eurasia. MPhil dissertation, University of Bradford, 2004. Brophy, K. and Cowley, D. 2005. From the Air Understanding Aerial Archaeology. Tempus, Stroud. Brunello, 1973, The Art of Dyeing in the history of Mankind. Vicenza: Neri Pozza Editore. Burnside, C.D. 1979. Mapping from Aerial Photographs. Series: Aspects of Modern Land Surveying. Granada, London. Callebat, L. (eds) 1973. Vitruve VIII Paris (editions Budé). Cited by Hodge, 1992 in: Hodge, A Trevor. 1992. Roman Aqueducts and Water Supply. Duckworth, London. Callister, W.D. 2000 Materials Science and Engineering: An Introduction, John Wiley & Sons, Inc. New York. Capasso, L. 2001. Palaeobiologia delle vittime dell’eruzione vesuviana del 79 d.c. I Fuggiaschi di Ercolano. di Bretschneider. Cardon, D. 2001 ‘On the road to Berenike: a piece of tunic in damask weave from Didymoi’ The Roman Textile Industry and its Influence A Birthday Tribute to John Peter Wild. Ed: Walton Rogers, P. Bender Jorgensen, L. Rast-Eicher, A. Oxbow books, Oxford. pp. 12-20). Carolis, E. 2001. Gods and Heroes in Pompeii. di Bretschneider. Cary, M. & Scullard, H. H. 1975. A history of Rome down to the reign of Constantine. New York: St Martin’’. 3rd Ed. Çengel, Y.A., & Boles, M.A. 1998 Thermodynamics: An Engineering Approach, McGraw Hill Ciarallo, A & De Carolis, E. 1999. Pompeii – Life in a Roman Town. Electra, Milan. Clarke, D.L.1973, Archaeology: the loss of innocence. Antiquity. 47:6-18. Cited by Bray in Bray, P. 2004. Choosing and justifying our conclusions: ‘Inference to the Best Explanation’ in the context of the transition to iron based economics in Eurasia. MPhil dissertation, University of Bradford, 2004. Clarke, D.L. 1978 Analytical Archaeology. 2nd Ed. Methuen & Co Ltd, London. Coghlan, H. 1977. Notes on Prehistoric and Early Iron in the Old World. Pitt Rivers Museum, University of Oxford. Coles, J. 1973. Archaeology by Experiment. Hutchinson University Library, London. Cotterell, B. and Kamminga, J. 1990. Mechanics of PreIndustrial Technology. Cambridge University Press. Crabtree, D.E. 1977. The Obtuse Angle as a Functional Edge. Chapt 2 in Ingersoll, D. Yellen, J.E. Macdonald, W. 1977. New York. Experimental Archeology Pub: Columbia, New York

ABAQUS 6.3 /Standard User’s Manual, Vol II, (2002) Hibbitt, Karlsson and Sorensen Inc. 2002 Allison, P. 1992. The Distribution of Pompeiian House Contents and its significance. Volumes I and II. University of Sydney, Australia. Almond, J.K. 1975: 11, Technical Aspects of Alum Making, Cleveland Industrial Archaeology Society, (2), 11-20. cited in Miller, I. 2002. Steeped in History The Alum Industry of North-East Yorkshire. Pub: North York Moors National Park Authority. Arthur, P. 1986. Problems of the Urbanization of Pompeii: Excavations 1980-1981. The Antiquities Journal. Oxford University Press. 66:29-44. Ashby, M F. & Jones, D. R. H. 2005. Engineering Materials 1 An Introduction to Properties, Applications and Design. Elsevier Butterworth Heinemann, Oxford. 3rd Ed. Baginski, A. 2001. Later Islamic and Medieval Textiles from Excavations of the Israel Antiquities Authority. Textile History, 23 (1) pp81-92. Eds Chapman, S & Levey, S. Pasold Research Fund. Balfour-Paul, J. 1999. “Indigo in the South and SouthEast Asia”. Textile History 30 (1), pp98-112. Eds Chapman, S & Levey, S. Pasold Research Fund. Bass, W. M. 1995. Human Osteology: a laboratory and field manual. Columbia: Missouri Archaeological Society. Ed: Bell, M; Fowler, P.J; Hillson, S.W. 1996. The Experimental Earthwork Project. 1960-1992. CBA Report 100. Bender Jorgenson, L. and Mannering, U. 2001. Mons Claudianus: Investigating Roman textiles in the desert. The Roman Textile Industry and its Influence A Birthday Tribute to John Peter Wild. Ed: Walton Rogers, P. Bender Jorgensen, L. Rast-Eicher, A. Oxbow books, Oxford. Pp1-11. Blockley, M. 1999. Archaeological Reconstructions and the Community in the UK. Chapter One, pp15-34. In: The constructed Past Experimental Archaeology, Education and the Public. Eds. Stone, P.G. and Planel, P.G. 1999. R36. One World Archaeology Series. Routledge. Published in Association with English Heritage. Blockley, M. 1999. Archaeological Reconstructions and the Community in the UK. Chapter One, pp15-34. In: The constructed Past Experimental Archaeology, Education and the Public. Eds. Stone, P.G. and Planel, P.G. 1999. R36. One World Archaeology Series. Routledge. Published in Association with English Heritage. Cited by Stone, P.G. and Planel, P.G. R36. One World Archaeology Series. Routledge. Published in Association with English Heritage. Chapt 1:32 Borgard, P.H., Brun, J-P., Leguilloux, M. TufferauLibre, M. 2003. Le Produzioni artigianali a Pompei Richerche condotte dal Centre Jean Bérard. Rivista Di Studi Pompeiani, XIV 2003. 162

References

Croom, A.T. 2000 (2002) Roman Clothing and Fashion Tempus, Stroud. Crowfoot, E. 1989. “A Romano-British Dress of the First Century B.C.?” Textile History, 20 (2), 123-28, 1989. Crumlin-Pedersen, O. 1999. Experimental Ship Archaeology in Denmark. In Experiment and Design Archaeological Studies in Honour of John Coles. Ed: Harding, A.F. Oxbow books, Oxford. Dalby, G. 1985. Natural Dyes, Fast or Fugitive Ashill Publications, Minehead, Britain. De Franciscis, A. 1978. The buried Cities. Orbis Publishing, London. DeHaan, N. 1997. Water Use in Private Baths at Pompeii. The 98th Annual Meeting of the Archaeological Institute of America. American Journal of Archaeology. 101:355-356. Diocletian, 301. Edict on Maximum Prices. In Roman Civilization Sourcebook II: The Empire. Eds. Lewis, N & Reinhold, M. 1966. Harper & Row, New York. Dixon, P. 1976. Barbarian Europe. Elsevier Publishing, Lausanne. P63. Dreyfuss, H. 1967. The measurement of man: human factors in design. New York, Whitney library of design. Dul, J. & Weerdmeester, B. 2001. Ergonomics for beginners A Quick Reference Guide. Taylor and Francis, London. Edmonds, J. Pers. Comms. 1999. Chiltern Open Air Museum. Edmonds, J. 2003. Brewing the Blues, New Scientist, Vol 179, No 2413, 20th Sept 2003, p24. Reed Business Information, London Edmondson, J. C. 1989. “Mining in the later Roman Empire and Beyond: Continuity or Disruption?” Journal of Roman Studies. Pub: The society for the Promotion of Roman Studies. 79, pp84-102 Ellis, M. 2001. Embroideries from Islamic Medieval Egypt in the Newberry Collection, Ashmolean Museum, Oxford. Textile History. 32 (1), pp61-74. Fagan, M.J. 1992. Finite Element Analysis: Theory and Practice. Longman Scientific & Technical: Harlow, England. Finley, M.I. 1973. The Ancient Economy. Chatto and Windus, London. Forbes, R.J. 1955. Studies in Ancient technology. Leiden. Vol 1, 1994. Cited by Hodge in: Hodge, A.T. 1981. Vitruvius, Lead Pipes and Lead Poisoning. American Journal of Archaeology. 85:486-491 Fraser, T.M. 1980. “Ergonomic Principles in the Design of Hand tools”. Occupational Safety and Health Studies. No 44. International Labour Office, Geneva. World Health Organisation. 1980. Frayn, J. 1984. Sheep-Rearing and the Wool Trade. Francis Cairns. Liverpool Frontinus. c35-104AD. De Aquaeductu. Cited by Hodge in: Hodge, A Trevor. 1992. Roman Aqueducts and Water Supply. Duckworth, London. Frost, H. J. & Ashby, M. F. 1982. Deformation-mechanism Maps: The Plasticity and Creep of Metals and Ceramics. Pergamon, Oxford.

Fulford, M & Wallace-Hadrill, A. 1998. Unpeeling Pompeii. Antiquity. Vol 72, No 275, March 1998 pp128-145. Fullick, P. 1994. Physics. Heinemann Education Publishers, Oxford Galer, I. 1987. Applied Ergonomics Handbook. Butterworths, London. 2nd Ed. Gere, J.M. & Timoshenko, S. P. 1997. Mechanics of Materials. PWS Publishing Company, Boston, Massachusetts Goodger, E. 1980. Alternative Fuels. Chemical Energy Resources. Macmillan Press Ltd. Granger-Taylor, H. 1982. “Weaving Clothes to Shape in the Roman World: The tunic and toga of the Arringatore”. Textile History, 13 (1) 3-25, 1982. ed Ponting, K. & Chapman, S. Granger-Taylor, H. 1991 – see Sheffer, A. and Taylor, A. 1991. The textiles from the ‘En-Boqeq excavation in Israel. Translated by H. Granger-Taylor. Textile History. 22 (1) pp3-46. Greenfield, P. 1972. Creep of Metals of high temperatures. Mills and Boon. Grierson, S. 1986. The Colour Cauldron Mill Brooks, Perth, Scotland. Grieve, M 1992. A Modern Herbal. Mayers of Chatham, Kent Grose-Hodge, H. 1944. Roman Panorama. Cambridge University Press. Groover, M.J. 1987. Automation, Production Systems, and Computer integrated manufacturing. Prentice-Hall. Hansen, H.O. 1959. I Built a Stone Age House. Phoenix, London. Cited in: The constructed Past Experimental Archaeology, Education and the Public. Ed. Stone, P.G. and Planel, P.G. R36. One World Archaeology Series. Routledge. Published in Association with English Heritage. Harris, J. 1999. 5000 years of Textiles. British Museum Press. Haynes, A.E. 1975. “Twill Weaving on the Warp-weighted Loom: Some technical considerations” Textile History 6 pp156-164. Pub: Pasold Research Fund. Henneberg, M. & Henneberg, R.J. 1999. Human Skeletal Material from Pompeii. In: Pompeii – Life in a Roman Town. Eds: Ciarallo, A and De Carolis, E. Electra, Milan. pp51-53. Heyerdahl, T. 1971. The Ra Expeditions. Doubleday, New York. Cited in Coles, J. 1973. Archaeology by Experiment. Hutchinson University Library, London. Hodge, A.T. 1981. Vitruvius, Lead Pipes and Lead Poisoning. American Journal of Archaeology. 85:486491 Hodge, A Trevor. 1992. Roman Aqueducts and Water Supply. Duckworth, London. Hoffman, T. R. 1967. Production: Management and Manufacturing Systems. Wadsworth Publishing Company, Belmont, California. Hofmann, W. 1970. Lead and Lead Alloys Properties and Technology. 2nd Ed, Translated to English by Lead Development Association. Springer-Verlag, Berlin. 163

Investigations into the Dyeing Industry in Pompeii Hopkins, H., unpublished 1998-9. Laboratory investigation of the dyeing outcome of madder when used with differing mordants, temperatures and pH. Hopkins, H. 2002. Industrial dyeing processes in Pompeii – Practical research into their methodology and importance. Undergraduate dissertation, Department of Archaeological Sciences, University of Bradford. Hopkins, H., Willimott, L., Janaway, R., Robinson, D., Seale. W. 2005. Understanding the economic influence of the dyeing industry in Pompeii through the application of experimental archaeology and thermodynamics. In Scientific Analysis of Ancient and Historic Textiles, Informing Preservation, Display and interpretation eds R. C. Janaway and P. Wyeth. Archetype Publications, London. Hopkins, G. A. 1992. Pers. Comm. Plumber, worked extensively with lead. Hopkins, R. G. 2002. Pers. Comm. Former Government Analytical Chemist and toxicologist. Hopp, W. J. and Spearman, M. L. 2001. Factory Physics Foundations of Manufacturing Management. Irwin McGraw-Hill, New York. 2nd Ed. Horne, L. 1982. Fuel for the Metalworker. The role of charcoal and charcoal production in ancient metallurgy. Expedition 25(1):6-13 Horne, L. 1982. Fuel for the Metalworker. The role of charcoal and charcoal production in ancient metallurgy. Expedition 25(1):6-13. Cited by Tonks in: Tonks, I. 1998. To determine whether prepared charcoal is anatomically different to burned wood. Undergraduate dissertation, University of Bradford, 1998. Hunter J., Roberts C., Martin A. (Eds.) 1996. Studies in Crime: An Introduction to Forensic Archaeology, London Batsford, 58-85 Ingersoll, D. Yellen, J.E. Macdonald, W. 1977. “Experimental Archeology” Pub: Columbia, New York See - Yellen, J.E. 1977. Introduction to “The Obtuse Angle as a Functional Edge” by Crabtree, D.E. Chapt 2 in Ingersoll, D. Yellen, J.E. Macdonald, W. 1977. New York. “Experimental Archeology” Pub: Columbia, New York James, D. 1990. Bradford. Ryburn Publishing, Halifax Janaway, R.C., 1996. The Decay of Buried Human Remains and their Associated Materials, in Hunter J., Roberts C., Martin A. (Eds.) Studies in Crime: An Introduction to Forensic Archaeology, London Batsford, 58-85 Janaway and Robinson, 1994, unpublished Japan Management Association. (Eds) 1985. Kanban JustIn-Time at Toyota. Management Begins at the Workplace. Translated: Lu, D. J. Jashemski, W. 1979. The Gardens of Pompeii. New Rochelle, New York: Caratzas. Pers. Comm. Jones 2002, Reader in Archaeology, University of Bradford Jongman, W. 1988. The Economy and Society of Pompeii. J.C. Gieben, Amsterdam.

King-Scott, P. 1971. Production Control for Supervisors. Collins, London. Koren, Z. 1994. Analysis of the Masada Textile Dyes. In, A. Sheffer and H. Granger-Taylor. Masada IV The Yigael Yadin Excavations 1963-1965 Final Reports. Israel Exploration Society The Hebrew University of Jerusalem. pp257-264. Lagercrantz, O. 1913. Papyrus Graecus Holmiensis Recepte für Silbur, Steine und Purpur. Laurence, R. 1994. Roman Pompeii Space and Society Routledge, London. Linscheid, P. 2001. Late Antique to Early Islamic Textiles from Egypt. Textile History 32 (1), pp75-80. Lawson, G. 1999. Getting to Grips with Music’s Prehistory: Experimental approaches to function, design and operational wear in excavated musical instruments. In Experiment and Design Archaeological Studies in Honour of John Coles. Ed: Harding, A.F. Oxbow, Oxford. Pp133-138. Maiuri, A. 1942. L’Ultima Fase Edilizia di Pompei. Rome: Istituto di Studi Romani. Maiuri, 1942:138, cited in Allison, P. 1992. The Distribution of Pompeiian House Contents and its significance. Volumes I and II. University of Sydney, Australia. Mann, S. 1994. A Review of the Evidence in Pompeii and Herculaneum on Textile Manufacture and Use. Undergraduate Dissertation, University of Bradford, 2004. Mary Rose Trust. No author named. Consulted 15/11/2001 http://www.maryrose.org/lcity/cook/ galley1.htm Mary Rose Trust. No author named. Consulted 15/11/2001. http://www.maryrose.org/lcity/cook/ galley3.htm Mathieu, J.R. 2002. Experimental Archaeology Replicating past objects, behaviours, and processes. In Experimental Archaeology Replicating past objects, behaviours, and processes. Ed: Mathieu, J.R. BAR International Series 1035, 2002. Archaeopress, Oxford. Michalopoulos, C. D. and Brotzen, F.R. 1968. A Note on the Torsional Creep of Polycrystalline Lead at High Temperatures. The Journal of the Institute of Metals. The Metals and Metallurgy Trust. Ed: J. S. Bristow. Vol 96, 1968. Miller, I. 2002. Steeped in History The Alum Industry of North-East Yorkshire. Pub: North York Moors National Park Authority. Moaveni, S. 1999. Finite Element Analysis – Theory and Application with ANSYS, Prentice – Hall Moeller, W. 1976. The wool trade of ancient Pompeii. E.J.Bril:Leiden. Mohamed, F. A., Murty, K. L., Morris, J. W. 1973. HarperDorn Creep in Al, Pb, and Sn. Mettalurgical Transactions. Pub: The Metallurgical Society of AIME, New York. Vol 4, April 1973. pp935-940 Monaghan, M.D. 2001.Coats of Many Colours: Dyeing and Dyeworks in Classical and Hellenistic Greece. PhD 164

References

Thesis. School of Archaeology and Ancient History, University of Leicester. Monden, Y. 1993. Toyota Production Systems An integrated Approach to Just-in-Time. 2nd Ed. Chapman and Hall, London Monteix, N. and Pernot, M. 2005.Pers. Comm from the forthcoming: Some features of the lead craftsmanship in Pompeii and Herculaneum. Poster presented at conference 2005. Summary of work to date undertaken at Centre Jean Bérard, Naples and Centre Camille Jullian, Aix de Provence. Mouritsen, H. 2001. “Roman Freedmen and the Urban Economy: Pompeii in the First century A.D.” Pompei tra Sorrento e Sarno Bardi Editore Roma Muhlemann, A., Oakland, J, Lockyer, K. 1992. Production and Operations Management. Pitman, London. 6th Ed. Murrel, R.F.H. 1957. “Data on Human Performance for Engineering Designers”. Newgate Press Ltd, London. Reprinted from periodical Engineering. National Institute for Occupational Safety and Health. 1981. Work Practice Guide for Manual Lifting. National Institute for Occupational Safety and Health, Cincinnati, Ohio. Niblett, R. 2000. Roman Verulaneum. St Albans District Council. Peacock, D. 1989. The Mills of Pompeii. Antiquity. 63:205214. Pliny the Elder. Natural History, with an English translation by H. Rackham and others. Thirty seven books in ten volumes. Loeb Classical Library. Heinemann 1958-1962 Pliny the Younger. Fifty letters of Pliny. Selected and edited by A.N. Sherwin-White. 1969. 2nd Ed. Oxford University Press. Translated by Betty Radice. Pomeroy, C.D. 1978. Creep of Engineering Materials. Mechanical Engineering Publications Plommer, H. 1973. Vitruvius and later Roman building manuals. Cambridge Classical Studies. Ponting, K. 1980. A Dictionary of dyes and dyeing. Mills and Boon Ltd, London. Pool, 1997. Prehispanic Kilns at Matacapan, Veracruz, Mexico. In (eds) Rice, P The History and Prehistoy of Ceramic kilns. Ceramics and Civilisation, Vol VII. The Ceramics Society. Pritchard, F. and Verhecken-Lammens, C. 2001. Two wide-sleeved linen tunics from Roman Egypt. The Roman Textile Industry and its Influence A Birthday Tribute to John Peter Wild. Ed: Walton Rogers, P. Bender Jorgensen, L. Rast-Eicher, A. Oxbow books, Oxford. Pp21-29 Raper, R.A. 1977. The analysis of the Urban Structure of Pompeii: a sociological examination of land-use (semi-micro)’ in D.L. Clarke (ed) Spatial Archaeology Academic Press, London. pp189-221 Cited by Robinson, D.J. 1999. in The Shape of Space in Pompeii. Bradford. Reynolds, P. 1999. The nature of experiment in archaeology. In: Experiment and Design Archaeological

Studies in Honour of John Coles. Ed A.F. Harding. Oxbow, Oxford. Pp156-162 Also online: http://www.butser.org.uk/iafexp_hcc.html 22/9/04. Richardson, L. 1988. Pompeii: An architectural History. Johns Hopkins University Press, Baltimore and London. Chapt 5: The Water and Sewage System. pp51-63 Robinson, D. 1999. The Shape of Space in Pompeii. Bradford. Robinson, D. 1999. Robinson, D.J. 2003. Pers. Comm. William Golding Fellow, Brasenose College, Oxford Robinson, D.J. 2005. Pers. Comm. William Golding Fellow, Brasenose College, Oxford Robinson, M. 2007. Pers. Comm. Computer aided engineering consultant, BEC, School of Engineering, Design and Technology, University of Bradford. Robinson, S. 1969. A history of dyed textiles Studio Vista, London. Rogers, P. Walton. 1997. Textile Production at 16-22 Coppergate. British Council for Archaeology. Rosetti, G. 1548. The Plictho of Gioanventura Rosetti: Instructions in the art of the dyers which teaches the dyeing of woollen cloths, linens, cottons, and silk by the great art as well as by the common. MIT Press, London. English Translation 1969. Rossotti, H. 1993. Fire. Oxford University Press. Rostoker, W, and Bronson, B. 1990. Pre-Industrial Iron. Its Technology and Ethnology. Archaeomaterials Monograph 1:61-68. Russell, J.C. 1977. The population and mortality of Pompeii, International Committee on Urgent Anthropological and Ethnological Research Bulletin 19: 107-14 Ryder, M.L. 1987. “Merino History in Old Wool. The use of wool remains in Ancient skin and Cloth to study the origin and history of the fine-woolled sheep that became the Spanish Merino” Textile History, 18 (2) 117-32, 1987 Ryder, M. L. 1990. The Natural Pigmentation of Animal Textile Fibres. Textile History 21 (2), 135-48. Ryder and Stephenson, 1968. Wool growth. Academic Press, London. Ryder, M.L. & Gabra-Sanders, T. 1992. “Textiles from Fast castle, Berwickshire, Scotland”. Textile History 23 (1) pp5-22. Ed: S, Chapman & S, levey. Pub: Pasold research Fund. Seale, W, J. 2002. Pers. Comm, Senior Lecturer in Department of Mechanical Engineering, University of Bradford. Seale, W, J. 2004. Pers. Comm, Senior Lecturer in Department of Mechanical Engineering, University of Bradford. Sheffer, A. and Taylor, A. 1991. The textiles from the ‘En-Boqeq excavation in Israel. Translated by H. Granger-Taylor. Textile History. 22 (1) pp3-46. Sheffer and Granger-Taylor, H. 1994. Masada IV The Yigael Yadin Excavations 1963-1965 Final Reports. Israel Exploration Society The Hebrew University of Jerusalem. 165

Investigations into the Dyeing Industry in Pompeii Singer, C. 1948:17. The earliest Chemical Industry. London. cited in Miller, I. Steeped in History The Alum Industry of North-East Yorkshire. Pub: North York Moors National Park Authority. Sahota M.K. & Riddington J.R. 2000. Compressive Creep Properties of Lead Alloys, Materials & Design, Vol 21, No 3, pp159-167. Speth. J.D. 1977. “Experimental Investigations of HardHammer Percussion flakes.” In Ingersoll, D. Yellen, J.E. Macdonald, W. 1977. Experimental Archeology Pub: Columbia, New York. Stanton, N. & Young, M. 1999. A guide to methodology in Ergonomics Designing for human use. Taylor and Francis, London and New York. Stone, P.G. and Planel, P.G. 1999. Experimental Archaeology, Education and the public. R36 One World Archaeology series. Routledge. Storey, G. 1997. The Population of Ancient Rome. Antiquity. 71:966-978. Storey, J. 1978. Dyes and Fabrics. Thames and Hudson, London. Soprintendenza Archeologia di Pompei. 2004. Digital Map of Pompeii. Unpublished. Taylor, G.W. 1983. Detection and Identification of Dyes on Pre-Hadrianic Textiles from Vindolanda. Textile History, 14 (2), pp115-124. Taylor, T. 2002. Pers. Comm. Reader in Archaeology, University of Bradford. Temin, P. 2001. A Market Economy in the Early Roman Empire. Journal of Roman Studies. Ed: M. Goodman. Pp169-181. 91: 169-81 Tonks, I. 1998. To determine whether prepared charcoal is anatomically different to burned wood. Undergraduate dissertation, University of Bradford, 1998. ToxBase. http://www.spib.axl.co.uk/ Poisons database used by the National Health Service of England, Wales and Scotland. Accessed 30-6-07. Trotter, M & Gleser, G.C. 1958. A pre-evaluation of estimation of stature based on measurements of stature taken during life and of long bones after death. Am. J. Phys. Anthrop, 16 (1): 79-123. Cited in: Capasso, L. 2001. Palaeobiologia delle vittime dell’eruzione vesuviana del 79 d.c. I Fuggiaschi di Ercolano. di Bretschneider. Trotter, M & Gleser, G.C. 1958. A pre-evaluation of estimation of stature based on measurements of stature taken during life and of long bones after death. Am. J. Phys. Anthrop, 16 (1): 79-123. Cited in: Bass, W. M. 1995. Human Osteology: a laboratory and field manual. Columbia: Missouri Archaeological Society. Vitelli, K.D. 1984. Greek Neolithic Pottery by Experiment. In: Pots and Potters Current Approaches in Ceramic Archaeology. Monograph XXIV Institute of Archaeology, University of California, Los Angeles. Ed. Rice, P. pp113-131

Vitruvius. Ten Books on Architecture. Translated by Morris Hicky Morgan. 1960. Dover Publications Ltd, New York. Vogler, H. 1982. “The Craft of Dyeing in Ancient Egypt”. Textile History 13 (2), 159-163, 1982. ed K. Ponting and S. Chapman. Walton- Rogers, P. 1997. Textile Production at 16-22 Coppergate. British Council for Archaeology. Wallace-Hadrill, A. 2006. Pompeii Live! Live interview, Public broadcast, 28-6-2006. Channel Five, United Kingdom. Wallace-Hadrill, A. 1990. The social spread of Roman luxury: sampling Pompeii. Papers of the British School at Rome. 58: 43-97. Cited in Laurence, 1994:150. Wallace-Hadrill, A. 1994. House and Society in Pompeii and Herculaneum. Princeton University Press. Watkinson, D and Neal, V. 1998. First Aid for Finds. Lavenham Press. Watling, L. 2004. Fuel Consumption of Roman Dye Vats. Undergraduate Dissertation, University of Bradford, 2004. White, T. D. 2000. Human Osteology. Academic Press, San Diego. 2nd Ed. Wild, J.P. 1970. Textile Manufacture in the Northern Roman Provinces. London, Cambridge. Wild, J.P. 1970. Textile Manufacture in the Northern Roman Provinces. London, Cambridge. Cited in: Moeller, W. 1976. The wool trade of ancient Pompeii. E.J.Bril:Leiden. Wild, J. Peter. 1977. “The Wool Trade of Ancient Pompeii” Review of Moeller, 1976. Textile History, vol8. Pasold Research Fund Ltd, Edington, Wilts. Pp180-181 Wilson, A. 2007, unpublished. Professor of Archaeology, Institute of Classics, University of Oxford. Wilson, D.R. 2000. Air Photo Interpretation for Archaeologists. Tempus, Stroud. 2nd Ed. Woolley, L. 2001. Medieval Mediterraneam Textiles in the Victoria and Albert Museum: Resist Dyed Linens from Egypt dating from the Fourth to the Seventh Centuries AD. Textile History 32 (1) 106-113, 2001. Eds Chapman, S & Levey, S. Pasold Research Fund. Wylie, A. 1989. Archaeology cables and tacking: the implications of practice for Bernstein’s ‘Options beyond Objectivism and Relativism’. Philosophy of Social Sciences. 19: 1-18. Cited in by Bray in Bray, P. 2004. Choosing and justifying our conclusions: ‘Inference to the Best Explanation’ in the context of the transition to iron based economics in Eurasia. MPhil dissertation, University of Bradford, 2004. Yellen, J.E. 1977. Introduction to “The Obtuse Angle as a Functional Edge” by Crabtree, D.E. Chapt 2 in Ingersoll, D. Yellen, J.E. Macdonald, W. 1977. New York. “Experimental Archeology” Pub: Columbia, New York

166

Appendices Appendix One: Coding Pompeii Coding Pompeii: The layout of the city and address description Location of dye works in Pompeii Location of Properties Appendix Two: Published Article, Hopkins et. al. 2005 Understanding the economic influence of the dyeing industry in Pompeii through the application of experimental archaeology and thermodynamics Introduction Dye Vat Design Methodology Implementation of Engineering Theory Results Conclusion Bibliography

Appendix Three: Dye Vat Recording Records Dye Vat Recording Sheets Pompeii Dye Vat Recording Sheet August 2002 (Vat)

Pompeii Dye Vat Recording Sheet August 2002 (Property)

Appendix Four: Deriving Lead data Assembling lead data for model The application of Finite Element modelling to the empirical data of Sahota and Riddington (2000) to determine the creep of lead at high temperatures for inclusion in the virtual model of the dyeing apparatus. Appendix Five: ABAQUS Input decks ABAQUS Input decks Input deck for the lead column at 20oC Input deck for the lead column at 40oC Input deck for the lead kettle

167

Investigations into the Dyeing Industry in Pompeii noted that each of the maps of Pompeii used in the Gazetteer are after Laurence, 1994.

Appendix One Coding Pompeii: The layout of the city and address description

Location of properties

The geographical numbering system used in Pompeii was assigned during excavation before the modern era. Pompeii was divided into districts each named a regio. These were then subdivided into insulae. The individual buildings were designated by the numbering of doorways. Regio and insulae were identified using Roman numerals, upper case for Regio, lower for insulae. The individual buildings were identified in Arabic (modern) numerals. The properties under investigation are: • • • • • •

The location of the dyeing properties is of interest as they are not distributed uniformly or predictably. This may be due to economic or legal reasons, reasons that are endemic entities, that is they are intangible reasons known only to the population in which they are located. However, it is possible to examine these reasons and explore the difference in location. Property I viii 19 This dye works was located in a courtyard and semiunderground stable attached to a residence. The dye works was separate within the residence, so it was possible to reside there without coming into direct contact with the dyeing. It appears that the dyeing would have been a source of income to the property owner in addition to any other income for another profession.

I viii 19 Vi4 Vi5 VII ii 11 VII xiv 17 IX iii 2

Location of dye works in Pompeii

Properties Vi4 and Vi5

Figure A1.1 is a map of Pompeii showing the location of the dye works (after Laurence, 1994). It may be noted that the properties are not located together and as such are located in areas of different land use. It should be

These two dyeing properties are located next to each other on a street that consists of small workshops. They take up the corner of a large residence. It is probable

Figure A1.1. Map of Pompeii showing the location of the dye workshops (after Laurence, 1994, amended by Author). The properties highlighted in green are those that withstood scrutiny and may be concluded to be dye works. The property highlighted in blue is the additional property identified by Moeller that has been shown not to be a dye works.

168

Appendices

that the location was originally part of the residence but was walled off and converted to provide additional income while maintaining the social distance from the process. It is not possible to enter Vi4 and Vi5 from the residence, but as the entrance is next to these two properties the location is not a barrier to communication.

2. Whereas VII ii 11 has been entirely taken over with dyeing, property IX iii 2 has more modest apparatus. However, its location is still advantageous. Property VII xiv 17 This property is a considerably sized residence with two manufacturing workshops within it. One is the dye works located at the back of the building. The other is a bakery located at the front of the building. The owners of the property would have had two industries to provide income.

Property VII ii 11 This property is of interest as the dyeing workshop was originally a villa, but has been converted to a dye works. It may be supposed that this occurred following the 62AD earthquake. The property of deceased landowners without legal heirs was seized and redistributed by the government. (Laurence, 1994). Some properties were abandoned altogether due to damage or the movement of owners away from the city, (Richardson, 1988). This left the properties open to reclamation by the remaining residents for either residential or business use. The presence of a dye works in a residential area may be related to the priority given to residential areas for water reconnection. (After Maiuri, 1942:138, cited in Allison, 1992; Hodge, 1992, Laurence, 1994). A dye works located near water will be operational quicker than a dye works without water as vast quantities of water is needed for the process. It is therefore preferable to locate a dyeing business in a residential area, relocating there if necessary. The property in question is of a large size and so would have been desirable.

It may be noted that each of these properties is located adjacent to or within a relatively large residence. This may be due to the owner converting part of the residence to dyeing, or the dyer enlarging parts of the property following business success. However, the priority given to residential areas following the earthquake of 62AD should not be overlooked, (after Richardson, 1988; Hodge, 1992; Laurence 1994). This priority would have enabled dye works located in residential areas to begin operation after the earthquake much sooner than those in non-residential areas, so that by the time of the eruption of 79 AD only those dye works that had had this advantageous location were still in business, (after Maiuri, 1942:138, cited in Allison, 1992). The location and significance of location of manufacturing facilities within the city of Pompeii was examined more fully by Robinson, (1999). Robinson’s work involved a study of the organisation of space at a neighbourhood, regional and citywide level and how this organisation related to social control and economic power. He concluded that it was a ‘complicated palimpsest of different and at times competing moral, politically and historical social landscapes’.

Property IX iii 2 This property is directly opposite property VII ii 11. Whether they are related is unknown. However, the advantages of water supply and location to property VII ii 11 would have been beneficial to property IX iii

169

Investigations into the Dyeing Industry in Pompeii Eventually it will be possible to determine the operating parameters of each of the dye vats, the quantities of consumables involved and the amount that could be produced. This should help answer the question as to the significance of the dye industry in Pompeii to the local economy.

Appendix Two H. Hopkins, L. Willimott, R. Janaway, D. Robinson, W. Seale. 2005. Understanding the economic influence of the dyeing industry in Pompeii through the application of experimental archaeology and thermodynamics. In Scientific Analysis of Ancient and Historic Textiles, Informing Preservation, Display and interpretation eds R. C. Janaway and P. Wyeth. Archetype Publications, London.

Key Words: Heat transfer, Reconstruction, Dye vat, Pompeii, Economy

Presented as a poster at AHRB Research Centre for Textile Conservation and Textile Studies First Annual Conference ‘Scientific Analysis of Ancient and Historic Textiles: Informing Preservation, Display and Interpretation’ in July 2004.

Introduction The role of the textile industry in Pompeii, in particular its contribution to the urban economy, has been the subject of debate since Moeller’s The Wool Trade in Ancient Pompeii, (Moeller, 1976). Moeller identified a number of structures throughout the city as pertaining to the textile producing industry, and concluded that ‘there must have been considerable surplus for export’. Jongman (Jongman, 1988) heavily criticised Moeller’s work claiming that the industry was in fact far smaller than had been previously thought, only capable of supplying Pompeii. Unfortunately, subsequent studies have been based on the same evidence as Moeller and Jongman’s, namely study of the remains in situ and review of the published literature. At present this alone is insufficient to allow a full understanding of the textile industry and its contribution to the economy of Pompeii and its Hinterland.

Understanding the economic influence of the dyeing industry in Pompeii through the application of experimental archaeology and thermodynamics Heather Hopkins,1 Lorna Watling,2 Rob Janaway,1 Damian Robinson,1 Bill Seale2 Abstract The influence of the dyeing industry in Pompeii on the local economy has been under discussion since the publication by Moeller in 1976. Since no absolute answer has emerged, the question was re-examined using two additional methods, experimental archaeology and the principles of thermodynamics.

This study aims to approach the question of the economic significance of the textile industry in a new and interdisciplinary way. Experimental reconstruction and the application of heat transfer principles from the field of engineering will test the assumptions of previous authors and provide new evidence of how the dye plants operated. Moeller’s identification of dyeing apparatus and workshops shall be re-evaluated and their operation reassessed. A recalculation of the capacity of each dye works shall be viewed in context with the population of Pompeii, to finally determine the scale of the economy. This finding shall withstand greater scrutiny, as it shall be based not purely on observation of the remains and published literature, but also on the operating parameters of a dye vat, determined through experimental reconstruction and application of heat transfer and other principles.

A full-scale replica of a dyeing apparatus from Pompeii was constructed and used to simulate repeated dye runs, and so determine operating parameters such as the times involved to heat and cool a vat and the consumables needed. This first replica also allowed a better understanding of how the apparatus was actually used. Thermodynamic principles, which were applied to understand the successes and failures within the experimental work, suggested that the vat operated in a predictable way and enabled the operational mechanics of the vat to be established. It is now possible to use both the experimental results and the thermodynamic modelling to determine not just the consumables used, but also the working environment needed for the vat to operate, allowing an understanding of the limitations to dyeing and to workers. Issues of practicality such as storage of consumables and disposal of exhaust gases may now be thoroughly examined.

To understand the operating parameters of a dye vat and the dyeing process, a copy of a small dye vat apparatus was constructed and used to replicate dyeing with madder. The replica was constructed from modern materials, their properties compared with Roman materials using the heat transfer principles. Heat transfer principles were used to model the vat. The replica vat and engineering theory had been developed independently, but were then combined to produce the fuel consumption model. Following this

Department of Archaeological Sciences, University of Bradford, Bradford, BD7 1DP 2  Department of Mechanical Engineering, University of Bradford, Bradford, BD7 1DP 1 

170

Appendices

the vat was amended to include a flue. The difference this made to the operating parameters was noted

Prior to dyeing the wool is cleaned (‘scoured’) and ‘premordanted’, (Frayn, 1984; Grierson, 1986). Mordanting is the process by which the wool is treated with a chemical (normally a metal salt) to allow the dye to stick. To dye the wool it is placed in the metal kettle of the dye vat once the dye within the vat has dissolved in the water and reached the correct temperature. The wool and ‘dye liquor’ are then simmered and the kettle left to cool naturally. This allows further dye to stick and a stronger, faster colour to result.

and the model was calibrated using the flued dye vat. Throughout this paper the first reconstructed vat (without the supplementary flue) will be referred to as the unflued vat and the modified vat (that includes the supplementary flue) will be referred to as the flued vat. Textiles do not tend to survive well in the archaeological record, (Watkinson and Neal, 1998:65; Harris, 1999:8), hence finds of Roman textiles are a rarity. The Masada textiles have been preserved to an extraordinary degree, the original dyes still being discernable, and represent the largest collection of Roman textiles discovered, (Sheffer and Granger-Taylor, 1994). Wool appears to have been the most commonly used and most commonly dyed material, (Sheffer and Granger-Taylor, 1994; Frayn, 1984:142-161). Madder is the most common dye found in the plant record. (Walton-Rogers, 1997). Madder was certainly in widespread use as a dye, (Ponting, 1980), and has been discovered in textiles throughout the Roman world. (Koren, 1994; Taylor, 1987). Of the dyes that were found in the Masada textiles, only madder could be identified from the dye back to the plant, (Koren, 1994), as the chemistry of madder is such that it may be unambiguously identified. Therefore it was decided that the vat would be used to replicate madder dyeing.

Control of the fire through the firebox is extremely important, as while the dye liquor must simmer, it must not be allowed to boil. Boiling ruins the fleece and causes changes to the dye possibly resulting in a different colour, (Storey, 1978). The fire (and therefore temperature) is influenced through the fuel amount and the airflow. Energy is released from the fuel during the process of combustion. The amount of energy per mass of fuel is quantified in the form of the calorific value of the fuel. The energy raises the temperature of the air above the fire, causing it to become less dense, the air rising upwards away from the fire. As the air leaves the fire it draws in new air containing more oxygen. This causes the fire to sustain itself. If the amount of air drawn in is insufficient the fire is extinguished, as there is insufficient oxygen to sustain combustion. (Rossotti, 1993). If a flue is present the change in density of the warm air (the density between the air at the base when compared to the air at the top of the flue) allows it to rise up the flue. (Fullick, 1994; Çengel and Boles, 1998). Therefore the draw is greater in a fire assisted by a flue. While the air travels up the flue it loses heat to the sides and so indirectly heats the vat. While the only thing driving the movement of the air is the heat generated

Dye vat design A dye vat is an apparatus used for dyeing wool, textile or yarn. It consists of a metal kettle, containing the material to be dyed, dye and water. This is supported in a brazier so that the kettle is held above a fire. The fire provides the heat for the dyeing reactions to take place and must be carefully monitored. Figure A2.1 shows the parts of a dye vat.

Lid Metal kettle Surround Dye liquor Firebox

Figure A2.1. Diagram showing the parts of a dyeing

Figure 1.The Diagram showing the parts of exists a apparatus. arrows indicate the natural ‘flue’ that an unflued vat. indicate the dyeing apparatus.in The arrows natural “flue” that exists in an unflued vat.

171

Figure A2.2. Unflued dye vat from VII ii 11, replicated in reconstruction

Investigations into the Dyeing Industry in Pompeii by the fire, it should be noted that the combustion may only be sustained if the movement draws sufficient new air. A flue allows the increased movement of air as the exhaust air may be released in a greater volume thereby increasing draw. The arrows in Figure A2.1 show the natural ‘flue’ that exists in an unflued dye vat.

been heated in a vat containing the mordant, usually alum, (Grierson, 1986; Sheffer and Granger-Taylor, 1994) dissolved in water. Following this a vat containing the dyestuff dissolved in water (90 litres for a 2kg fleece) would have been heated, and the fleece added. The fleece would have been simmered for an hour and then be allowed to cool naturally. The heating allows the activation of the dye and cooling within the dye liquor allows the dye to stick to the fleece.

Methodology A replica vat was constructed based on measurements and photographs from Pompeii (see table 1). The vat was a replica based on the design of vats within property VII ii 11, (see Figure A2.6). The kettle was manufactured from stainless steel. The design was based on the kettles that remained in properties 1 viii 19 and VII xiv 17 (see Figure A2.6). The recipe that was used included just madder and alum (Story, 1978) and required the vat to contain 90 litres of water, which were then heated to 95oC and held at that temperature for 1 hour. The vat was then allowed to cool naturally with the fleece and water still in place. While no madder or alum was actually used their presence was allowed for. The effect of lead and madder on fleece was examined through laboratory work.

The dye kettles from Pompeii were originally manufactured from a lead-based alloy. Heat transfer calculations showing what was used in the replica was similar to the original, (Çengel and Boles, 1998). Lead was most suitable as a dye vat material as alternative metals, such as copper and bronze, are mordants and sadden (darken) or alter the dyestuff. Lead is also a mordant, but brightens the colour by increasing the uptake of dye by the cloth. The disadvantage of lead that its strength in relation to its weight means that it is fragile, especially after repeated heated and cooling (as creep may alter the shape). (Pers Comm. Wright, 2004). It is not possible to lift the vat or for it to support its own weight. This means that construction and repairs must take place in situ, as it would not be possible to lift a vat into a brazier.

The operating parameters that were tested included the time it took for the vat to heat and cool, and the amount of fuel that the vat required. The fuel used was pine, not because it was believed that pine had been used in antiquity, but because it provided a uniform fuel from which the calorific value could be calculated. Heating with charcoal was also attempted. Methods of emptying and cleaning the vat were also assessed. The flued vat was tested in the same way.

Implementation of Engineering Theory Part of the study into the economic impact of Pompeii’s dyeing industry required an assessment of the quantity of fuel used. Results from the experimental work were not directly transferable due to the use of modern materials in the reconstruction. However this was not an insurmountable problem given that the experiments gave actual results for a measurable system. The results from the experiments could be used to understand and quantify the heat transfer processes taking place within the system during operation. By accounting for the energy required by various parts of the system it is possible to calculate the quantity of fuel necessary to run the system, (Çengel and Boles, 1998).

There is little evidence for or against the use of a lid. As yet a lid has not been discovered through excavation. The pictorial evidence is unclear. The subject depicted in the wall painting outside Verecundus’s workshop in Pompeii (Wild, 1970) is still debated. The other notable picture is from a tombstone in Arlon, Belgium (Wild, 1970) and depicts the stirring of a dye vat, an activity that took place with the lid off, and so there is no lid depicted in this picture. This had led to the conclusion that a lid was not used in antiquity in conjunction with these dye vats. However calculations of heat lost and matter transfer (water evaporating) during the experiment demonstrated that a lid was required when heating the vat. A wooden lid of 2cm thickness would have halved the heat lost through the top of the vat. The water loss without the lid would have led to the ruin of the fleece.

It was also realised that a study of this nature could enable a better understanding of the system through analysis of the variables applied to the system and how great an affect on the fuel used changing those variables would have. This would require a model to be created for finding the fuel consumed so that variables could be tested without altering any of the other operating parameters. An understanding of which variables have the most influence on the fuel consumed will be invaluable in future reconstructions involving heat transfer.

The dyeing process in Pompeii would have been similar to processes carried out through history. The fleece would have first been pre-mordanted – it would have

Below is a simplified version of the system that was examined.

172

Appendices

Ambient air temperature = 20°C

Lid was 0.02 m thick and made of wood

Initial surround temperature = 20°C Final surround temperature = 95°C

Initial vat temperature = 20°C Final vat temperature = 95°C Initial water temperature = 20°C Final water temperature = 95°C Vat was made of stainless steel

Surround was made of brick

Average exhaust gas temperature = 200°C

Air flow rate through system was 1 m/s Fuel used was wood

Figure A2.3. Diagram showing initial parameters used for calculations.

It was noted through experimental work and heat loss calculations that energy was lost through the following ways (parts are labelled on Figure A2.1):

The variables examined for the affect they had on the fuel consumption were the surround material, vat material, initial water temperature, final water temperature and ambient air temperature. The effect of changing the quantity of water was also briefly examined. Prior to examining the materials it was important to establish which property of the materials that had been used in the calculations was causing the greatest change in the results for the fuel consumed (Callister, 2000; Çengel and Boles, 1998). The three properties of the materials that were examined were specific heat capacity, thermal conductivity and density. The specific heat capacity is the energy required to raise 1kg of a substance by 1 degree (when using the units kJ kg-1 K-1). The thermal conductivity is the rate of flow of energy through a material of thickness 1m and with a temperature difference of 1 degree between the two sides of the material (when using the units W m-1 K-1). Density is the mass per unit volume (kg m-1). (Callister, 2000; Çengel and Boles, 1998; Fullick, 1994). It was found that the specific heat capacity had the most significant affect and hence materials with relatively high, medium and low specific heat capacities were used for the material evaluation.

1. Heat was lost through the top of the vat 2. Energy was required to heat the vat to steady state temperature 3. Energy was required to heat the surround to steady state temperature 4. Energy was required to heat the water to steady state temperature 5. Energy was lost through the walls of the vat 6. Energy was lost to the air flow through the system 7. Energy was lost to the ground A set of calculations to account for the energy lost in the processes described above were established (Çengel and Boles, 1998; Diamant, 1986; Holman, 1981, 2002; Edwards et al, 2002) and this could be used as the basis of a model for the fuel consumption of the dye vat. The model created was calibrated using the data from the experiments done on the modified vat (flued version). This gave an error within the model of ± 0.5 kg. This quantity of fuel was arrived at using a calorific value of 15,800 kJ/kg (Cooper & Rose, 1977).

The results of the key factors assessment are shown below: 173

Ex pe r Pr ime op nt al er t un y Pr flu op VII e ii er 11 d (N ty Pr ) va op VII t7 ii er 1 (N Pr ty 1 1 v ) at op v ii 6 er 1 (N Pr ty V 9 v ) at op II ii 3 er 1 (N Pr ty V 1 v at ) op II ii er 11 4 (N ty ) va VI Ii t8 E Pr xpe i 11 (N op rim ) va er t9 e nt ty (N a VI ) I x l flu e Pr iv op 17 d (Y e Pr ) va op rty t5 V e Pr r (Y i op ty 1 5 v at ) er vi 1 Pr ty V i 19 (Y op ) II va x er t1 iv t y 17 Pr (Y op VII ) va xi er t v 4 t y 17 Pr (Y op VII ) va xi er t3 v ty 1 (Y 7 Pr VII ) va xi op t v 8 er 17 (Y ty ) va 1 v t Pr op ii 19 2 (Y er ) va ty t5 V (Y i4 ) va t8 (Y )

Mass (kg)

Vat material

Mass of water (kg)

Initial water temperature

Mass of surround (kg)

174 Ambient air temperature

2500 25

2000 20

1500 15

1000 10

500

5

0

0

Fuel consumption (kg)

Surround material 99°C

95°C

90°C

35°C

20°C

5°C

15°C

10°C

5°C

Stainless Steel (316)

Stainless Steel (405)

Lead (Chemical)

Sandstone

Brick

Granite

Fuel consumption (kg)

Investigations into the Dyeing Industry in Pompeii

Effect of different variables on fuel consumption

16

14

12

10

8

6

4

2

0

Final water temperature

Figure A2.4. Effect of different variables on fuel consumption

Fuel consumption of different sizes of vat

Fuel used (kg)

Figure A2.5. Effect on fuel consumption of different sizes of vat. (N) indicates that no external flue was present. (Y) indicates the presence of an external flue.

Appendices

The fleece broke up considerably in the vat and the madder tangled with the fleece. This would have resulted in the drains on the plumbed-in vats blocking and the fleece to become irregularly coloured or extra time and effort being taken to untangle the fleece.

In each of the cases only the variable being examined was changed. The actual values for the fuel consumption are relative. This is because only the variable being examined was changed. The important part of this graph is the scale of changes in the fuel consumption for likely changes in the variables. The affect of changing the surround material has the greatest effect on the fuel consumed. This means that it will make little difference to the results acquired if say the air temperature was 20°C different to the temperature in Pompeii or if the vat material was made of stainless steel instead of lead.

The average time taken for heating the vat was two hours to heat, one hour to simmer and then at least four hours to cool. From this it was seen that the vats could not have been heated more than once a day. The water must cool naturally with the dye and fleece within it otherwise the dye will not combine with the fabric, (Storey, 1978). Therefore the vats may only have been heated once a day.

The next part of the study was to examine output of the other vats in Pompeii. The results for the flued vats did not follow the expected trend and it is believed that this is due to the assumption that all the flues had the same dimensions as the one of the replica. This was a necessary assumption because no other information on the flues attached to the vats in question was available. The results are shown in the graph below:

It has been assumed that each fleece needed 90 litres of water to allow successful dyeing (a lesser amount would not enable the fleece to move freely). The amounts of fleece that each dye vat could contain was calculated. As the vats were not constructed according to metric, it was difficult to determine the amount of fleece some of the vats could contain. Therefore each vat was given a minimum and maximum amount that it could have contained. This may be viewed in table A2.1. The maximum and the minimum calculations give a boundary to the maximum amount of dyed fleece that it was possible to produce in Pompeii each working day.

During the previous assessment it was assumed that the volume of water in the vat was constant. Whilst examining the results for other dye vats found in Pompeii it became apparent that the effect of water was of a similar order of magnitude to that of the surround material. This was investigated further and it was found that the water has a more significant impact on the fuel used than the surround material. Therefore if the dye vat was to be run with say 130 litres of water instead of 90 litres the fuel used would be significantly greater even if you removed 40 litres worth of bricks from the surround structure.

Minimum: 15 + 28 + 13 + 40 + 39 + 19 = 154 fleeces per day Maximum: 17 + 32 + 16 + 41 + 42 + 19 = 167 fleeces per day It would have been possible to dye between 154 and 167 fleeces a day in Pompeii. There were 318 working days in Pompeii.

This work has provided the basis for further study into the understanding of the working of the system.

Minimum: 154 x 318 = 48, 972 Maximum: 167 x 318 = 53, 106

Results

Between 48, 972 and 53,106 fleeces could be dyed in Pompeii per year.

It was noted that despite various attempts and methods charcoal would not light and sustain combustion in the unflued vat, and would light but then extinguish shortly after ignition in the flued vat. Wood was able to burn easily in either vat, with no external source of ventilation or through draft. On average the unflued vat would take two hours to heat, would be held at temperature for one hour and then would take in excess of four hours to cool. The unflued vat behaved similarly, but the flame control was greater and less flame and smoke came out of the front of the vat. Between six and seven kilograms of wood was required to heat either vat, a quantity that translates to a modern dustbin bag.

It was assumed that the population of Pompeii was 12,000 people, (Storey, 1997). If these totals were to be divided evenly between the population of Pompeii, the average would give an estimation as to the size of the industry in terms of its capacity to provide the populace with all of their textile needs. If the number produced per person was extremely large, this would immediately suggest export. If the number produced was not excessively large, this would result in a need for extra research into the scale of the industry.

This is equivalent to 120,000 kJ, if the calorific value of the wood is taken as 15,800 kj/kg. (Cooper and Rose, 1977).

48,972 12,000

175

= 4.081

This may be rounded down to 4 fleeces per person.

Investigations into the Dyeing Industry in Pompeii

53,106 12,000

flow did not appear to make a significant difference to ventilation, although it did alter the controllability of the flame.

= 4.4225 This may be rounded down to 4.4 fleeces per person.

Ryder states that the closest fleece to that of a Roman fleece (in terms of size, weight and make-up) is that of a Shetland sheep, (Ryder 1990). Therefore a fleece was taken as weighing approximately 2kg. 4.081 x 2kg = 8.162 kg

Conclusion Following the experimental work, engineering work and further literature review the following conclusions could be made about how the vat operated.

4.42 x 2kg = 8.84 kg

Conclusions that may be drawn regarding the operation of the vat:

Therefore between 8.162 kg and 8.84 kg dyed fleece was produced annually per person in Pompeii. This may sound like a lot. However, in modern terms this is a ‘washing machine load’ (albeit a large one). It could therefore be reasoned that the dyeing industry of Pompeii was extremely small, possibly specialist, and not of a scale large enough to export, (as Moeller, 1976, had suggested).

• The fuel was wood or a fuel that had similar calorific value per weight. It was not charcoal. • The quantity of fuel used may have been a limiting factor. • The presence of a flue in the original vat indicated a difficulty in causing adequate airflow of that vat. • The dye works may have been operated by a small number of people, between one and three. However this does not exclude a larger workforce. • All of the vats were usable by an average Roman. Vats that at present would not be usable show signs of having been altered. The original heights of all of the vats are discernable and these were usable. This would mean that height would not have restricted the workforce. • A lid was used during dyeing. This was probably constructed from wood. • The vats were heated and cooled once a day. It is most likely that they cooled overnight and were emptied in the morning, as they would have taken at least fours hours to cool naturally. • The quantity of water used had the greatest effect on the amount of fuel required to heat the vat. If the water quantity remained the same, the material used for construction of the brazier had the greatest effect on the quantity of fuel required.

This issue is confused by the use of dyed textile as a decoration and not the main piece of some garments. This custom would mean that the amount produced could be an underestimate. However, production of some items, such as curtains, would require such an amount that a larger quantity would be required than it appears. It may also be supposed, following the practical experiment, that to operate the dyeing apparatus alone would have only taken a relatively small amount of labour. It took two people minimal effort to run the single apparatus used in the reconstruction once it was alight and functioning. Although it may be seen that the smallest number of dye vats was three in a property, and that the properties usually had more than just dyeing apparatus, it may be supposed that the actual dyeing itself was not a labour-intensive task. Following review of the skeletal data (Capasso, 2001), it was noted that two of the authors were exactly the height of the average male and female Roman. Following closer examination of the original vats it was determined that both authors were able to use each one, and would even have been able to clean the base of each one.

Conclusions that may be drawn regarding the influence of the dyeing industry on the economy of Pompeii: The industry was not large enough to export. However, it was large enough to supply Pompeii and possibly the hinterland. The use of dyed material as a decoration and not as the main body of some garments may have led to an underestimation of the figure produced, but the need to scour the textile and the use of scouring plants leads to the conclusion that this is not a vast underestimate. Moeller’s findings are still an overestimation.

It was noted that vats in enclosed properties had flues, despite their size, where as vats in more open properties did not, unless the vat was particularly large. Following the practical work and the examination of the property, it is believed that this is linked to the ventilation of each vat and the relative airflow. The finding is further supported as in an entirely open environment the

176

Appendices

Table A2.1, showing the possible minimum and maximum outputs of each dye works per day. The amount of fleeces were determined assuming that each fleece require 90 litres of water.

Y N Y Y Y

Internal Diameter (cm) 98 Circ 2.64m 74 97 40-50

Internal depth (cm) 51+ 54+ 63 75 50

4 vat 5 vat 6 vat 7 vat 8 vat

Y ? Y ? ?

97 61 120 74 120

80? 52+ 80 78+ 81

5

1 vat 2 vat 3 vat 4 vat

Y N N Y

95 50 99 85

60 under 91 68 69+

Ii

11

1 vat 2 vat 3 vat 4 vat 5 vat 6 vat 7 vat 8 vat 9 vat

N N N N N N N N N

106 106 96 76 50? 86 55 99 94

69? 74? 60? 56? 43? 54? 55 38+ 55+

VII

Xiv

17

1 vat 5 vat 6 vat 7 vat 8 vat 9 vat 10 vat 11 vat 12 vat 13 vat

Y Y Y Y Y Y Y Y Y Y

68 117 100 99 60 109 68 95 51 94

50 65+ 76 68 43+ 48? 51+ 63 43 60

Ix

Iii

2

1 vat 2 vat 3 vat

N N N

73 83 90

133 130 135

Regio

Insula

No.

Type

Flue

I

Vii

19

1 vat 2 vat 3 vat 5 vat 7 vat

V

I

4

V

I

VII

177

Vol. (l) 384.69 299.26 270.95 554.23 80

Total: 229+(591?) 151.97 904.77 335.47 916.09 Total: 425.29 178.68 523.44 319.54 Total: 608.91 653.03 434.29 254.04 84.43 313.68 130.67 292.51 381.69 Total: 181.58 698.84 596.90 523.44 121.58 447.91 185.22 446.56 87.84 416.39 Total: 556.66 703.37 858.83 Total:

New maximum 4 3 3 6 1 17 6 2 10 4 10 32 5 2 6 3 16 7 7 5 3 1 3 1 7 7 41 2 8 7 6 1 5 2 5 1 5 42 5 (new 6) 6 (new 7) 8 (new 9) 19

New minimum 3 3 3 5 1 15 5 1 10 3 10 28 4 1 5 3 13 7 7 4 2 1 3 1 7 7 40 2 7 6 5 1 5 2 5 1 5 39 5 (new 6) 6 (new 7) 8 (new 9) 19

Investigations into the Dyeing Industry in Pompeii Bibliography

Israel Exploration Society The Hebrew University of Jerusalem. pp257-264. Moeller, W. 1976. The wool trade of ancient Pompeii. E.J.Bril:Leiden. Ponting, K. 1980. A Dictionary of dyes and dyeing. Mills and Boon Ltd, London. Rossotti, 1993. Fire Oxford University Press Ryder, M. 1990. The Natural Pigmentation of Animal Fibres. Textile History, 21 (2), 135-48, 1990. Sheffer and Granger-Taylor, H. 1994. Masada IV The Yigael Yadin Excavations 1963-1965 Final Reports. Israel Exploration Society The Hebrew University of Jerusalem. Storey, G. 1997. The Population of Ancient Rome. Antiquity. 71:966-978. Storey, J. 1978. Dyes and Fabrics. Thames and Hudson, London. Taylor, G. 1987. Reds and Purples from the Classical World to Pre-Conquest Britain. Textiles in Northern Archaeology. NESAT III textiles symposium in York. Walton- Rogers, P. 1997. Textile Production at 16-22 Coppergate. British Council for Archaeology. Watkinson, D and Neal, V. 1998. First Aid for Finds. Lavenham Press. Wild, J.P. 1970. Textile Manufacture in the Northern Roman Provinces, Cambridge University Press. Wright, C.S. Pers Comms. 2004. Lecturer in materials, School of Engineering, University of Bradford.

Capasso, L. 2001. Palaeobiologia delle vittime dell’eruzione vesuviana del 79 d.c. I Fuggiaschi di ercolano. di Bretschneider Callister, W.D. 2000 Materials Science and Engineering: An Introduction, John Wiley & Sons, Inc. New York. Çengel, Y.A., & Boles, M.A. 1998 Thermodynamics: An Engineering Approach, McGraw Hill Cooper, J.R. & Rose, J.W. 1977, Technical Data on Fuel, British National Committee World Energy Conference, pp 304-30 Diamant, R.M.E., 1986, Thermal and Acoustic Insulation, Butterworths, Edwards, D.K., Denny, V.E., Mills, A.F., 1973 Transfer Processes. An Introduction to Diffusion, Convection and Radiation, Holt, Rinehart and Winston Inc. p139 Frayn, J. 1984. Sheep-Rearing and the Wool Trade. Francis Cairns. Liverpool. Fullick, P. 1994. Physics, Heinemann Education Publishers, Oxford Grierson, S. 1986. The Colour Cauldron Mill Brooks, Perth, Scotland. Harris, J. 1999. 5000 years of Textiles. British Museum Press. Holman, J.P. 1981, Heat Transfer, McGrawHill Holman, J.P. 2002, Heat Transfer, McGraw Hill Jongman, W. 1988. The Economy and Society of Pompeii. J.C.Gieblen. Amsterdam. Keenan-Jones, D. Hellstrom, J. Drysdale, R. 2011. Lead Contamination in the Drinking Water of Pompeii. In: Pompeii Art, Industry and Infrastructure. Eds: Poehler, E. Flohr, M. Cole, K. Oxbow, Oxford. Chapt 9. pp131-148. Koren, Z. 1994. Analysis of the Masada Textile Dyes. In, A. Sheffer and H. Granger-Taylor. Masada IV The Yigael Yadin Excavations 1963-1965 Final Reports.

Principle author: Heather Hopkins [email protected] Department of Archaeological Sciences, University of Bradford, Bradford, BD7 1DP

178

Appendices

Figure A2.6. Unflued dye vat from VII ii 11, replicated in reconstruction

179

Investigations into the Dyeing Industry in Pompeii Appendix Three

Pompeii Dye Vat Recording Sheet August 2002 PROPERTY NUMBER

Front elevation

Side elevation

Plan

Max external height

VAT NUMBER

Max external diam. Max internal height Max internal diam. Firebox opening ht. Firebox opening width Wall thickness Kettle support ht. Presumed kettle depth Step dimensions Attributes Complete? Evidence of Restoration? Flue?

Present? If not present: restoration, incompleteness or truly absent?

Step?

Present? If not present: restoration, incompleteness or truly absent?

Metal kettle?

Details

180

I

Appendices

Vat Construction Construction Material

Brick, Stone %

Dimensions of brick/stone Construction technique Mortar type Wall uniform in width? Firebox lintel Firebox lintel dimensions Orientation of Firebox Step Material

Evidence of Repair/Refurbishment Modern restoration? Ancient repair? Further Observations

181

Investigations into the Dyeing Industry in Pompeii

Pompeii Dye Vat Recording Sheet August 2002 PROPERTY NUMBER

Property type/layout Building type Number of Vats All in same room/area? Evidence of upper floor Inside/outside/under cover? Direction of floor slope Evidence of drains Drainage capacity? Level of Ventilation Evidence for Water supply Storage areas?

Water supply Water source Distance of vats from, water Pipe diameter Pipe construction Water storage, tanks, cisterns etc. Public or private water supply?

182

Appendices

Checklist for plan annotation: 1. Vat distance from wall 2. Orientation of firebox 3. Orientation of flue 4. What flue leads out onto 5. Doors- dimensions including sills and heights 6. Direction of floor slope 7. Ventilation type and size 8. Drains 9. Water supply 10. Possible storage areas (dimensions) 11. Movement between rooms and levels

Ventilation type and Checklist for elements 1. Height of each wall including incomplete walls 2. Ventilation- windows, doors, chimneys, floors (note where these lead to)

Notes

Other equipment Storage locations

183

Investigations into the Dyeing Industry in Pompeii temperatures have not and so have to be determined. Determination of missing data through extrapolation of the known values is the method by which design was originally tested before the advent of Finite Element Analysis, (Greenfield, 1972; Pomeroy, 1978; Boyle and Spence, 1983).

Appendix Four: Assembling lead data for model To create a simulation of the dyeing apparatus that showed the effect of the thermal cycle on the apparatus, it is necessary to include the creep values for lead. It is necessary to include the rate at which the lead would creep at a specific temperature under a specific load stress to allow the lead kettle to creep accurately. However, there is insufficient published data on the creep values of lead, so before this inclusion the values had to be determined.

The illustration below (Figure 4.1) is of the creep testing apparatus used by Sahota and Riddington, (2000). The lead was subjected to differing levels of stress at differing temperatures to determine the amount by which it would creep.

To determine these creep values it is necessary to build a model simulation of lead itself. However, the values to be included in this model also had to be determined. The values for the strain of the lead under the application of varying stresses at low temperatures have been published in the literature, (Sahota and Riddington, 2000). However, the values for the strain under higher

The values for creep strain at 20oC under varying load stresses were determined using this apparatus. When this data was used to construct a Finite Element simulation of a lead block of the same dimensions at differing stresses at 20oC the creep strain was shown to be similar. This may be viewed in Figures 4.2 and 4.3.

Figure A4.1. The diagram of the creep testing apparatus used by Sahota and Riddington, 200. From Sahota and Riddington, 2000

184

Appendices

Figure A4.2. Graph showing the creep of lead under differing stresses over ten days at 20°C

Figure A4.3. Graph showing the creep of lead under differing stresses over ten days at 20°C. 864,000 seconds is equivalent to ten days. This graph was constructed as a result of a simulation in FE Analysis of the data presented by Sahota and Riddington, (200)

185

Investigations into the Dyeing Industry in Pompeii When a second Finite Element model was constructed to simulate the lead block under the same stresses but at 40oC, this too matched the creep strain determined by Sahota and Riddington (2000). This may be viewed in Figures 4.4. and 4.5.

been used in the creep tests by Sahota and Riddington, (2000). This was then subjected to 7.5 MPa at differing temperatures to determine the rate of creep within the lead. Node 5001 (top left) was flagged to allow its change of location to be recorded as the creep took place. Figures 4.6 and 4.7 show the creep of the lead at 20oC. Friction coefficient between the platens and the lead was taken to be 0.25.

The values determined were used to create a geometric axi-symmetric simulation of the lead column that had

Figure A4.4. Graph showing the creep of lead under differing stresses over ten days at 40°C

Figure A4.5. Graph showing the creep of lead under differing stresses over ten days at 40°C. 864,000 seconds is equivalent to ten days. This graph was constructed as a result of a simulation in FE Analysis of the data presented by Sahota and Riddington, (200)

186

Appendices

Figure A4.6. FE simulation of the lead creep test undertaken at 20°C using creep data from Sahota and Riddington, 2000

Figure A4.7. Displacement of the top platen after 10 days at 7.5MPa at 20°C. The platen displaces by 10.66mm, the sample is 80mm in length, 10.66mm / 80mm = 0.133 = 13.3% strain

187

Investigations into the Dyeing Industry in Pompeii Figure 4.8 and 4.9 show the effect of creep at 40oC.

concluded that the model was accurate for modelling purposes and could be used to determine the missing data. The lead tests had not been undertaken to temperatures above 40oC in the literature and so it was necessary to determine values above this for inclusion into the model dyeing apparatus (which would have been heated to 120oC). These are shown in Figure 4.10.

It was determined that at 20oC the lead would creep 13.3%, while at 40oC the lead would creep 26%. These results may be compared with the findings of Sahota and Riddington (2000) in Figures 4.2 and 4.4. Anecdotally, the shape of the column after creep has taken place may be seen to match the column shown in the plate in Figure 4.1.

The results gained from this modelling shall be used to describe the creep behaviour of the lead kettle in the model simulation of the dyeing apparatus.

As the Finite Element model had been shown to match the original results from the lead creep test it was

Figure A4.8. FE simulation of the lead creep test undertaken at 40°C using creep data from Sahota and Riddington, 2000

188

Appendices

Figure A4.9. Displacement of the top platen after 10 days at 7.5 MPa at 40°C. The platen displaces by 20.57mm, the sample is 80mm in length, 20.57mm / 80mm = 0.26 = 26% strain

Figure A4.10. Creep strain of lead at 5 MPa at differing temperatures over ten days. The higher data for higher temperatures was extrapolated from the successful matching of the FE simulation and data for the lower temperatures.

189

Investigations into the Dyeing Industry in Pompeii ** ** -----------------------------------** Rigid surface lower and upper plate ** -----------------------------------** *node,nset=ref1 5000,0,-1 *node,nset=ref2 5001,0,81 ** ** lower plate ** *rigid body,ref node=ref1 *surface, type=segments, name=lplate, ref node=ref1 start, -1.0, 0.0 line, 21.0, 0.0 ** ** upper plate ** *rigid body,ref node=ref2 *surface, type=segments, name=uplate, ref node=ref2 start, 21.0, 80.2 line, -1.0, 80.2 ** ** --------------------------------------------** contact properties between sample and platens ** --------------------------------------------** *contact pair,interaction=intlow lowint,lplate *surface interaction,name=intlow *surface behavior,pressure-overclosure=exponential 0.1,1.0 *friction 0.25 ** *contact pair,interaction=intup upint,uplate *surface interaction,name=intup *surface behavior,pressure-overclosure=exponential 0.1,1.0 *friction 0.25 ** ** ------------------** Material properties ** -------------------

Appendix Five: Abaqus input decks Input deck for the lead column at 20oC, used to gain the data for the input deck for lead kettle. *heading Compressive creep Sahota, Riddington 7.5MPa ** ** Cylindrical test specimen compression tested ** Diameter = 40mm, length = 80mm ** 99.99% pure lead ** Stress = 5 MPa ** Temperature =20C ** ** -----------------------** Sample geometry topology ** -----------------------*node 1, 0,0.1 20,20,0.1 ** 1601, 0,80.1 1620,20,80.1 ** *ngen,nset=l1 1,20,1 *ngen,nset=l2 1601,1620,1 *nfill,nset=nall l1,l2,80,20 *element, type = cax4rt, elset = vat 1,1,2,22,21 *elgen,elset=vat 1,19,1,1,80,20,20 ** ** --------------------------------------------** upper lower elements in contact with rig ** --------------------------------------------** *elset,elset=lower,gen 1,19,1 *elset,elset=upper,gen 1581,1599,1 ** *surface,name=lowint,type=element lower,s1 *surface,name=upint,type=element upper,s3 190

Appendices

*CREEP,LAW=TIME 1.1E-4,2.0,-0.70,20 2.8e-4,2.0,-0.75,40 5.0E-4,2.1,-0.78,80 8.0E-4,2.1,-0.80,100 1.5E-3,2.1,-0.85,120 ** ** ------------------** Boundary conditions ** ------------------** *nset,nset=symm 1,1601,20 ** *boundary symm,1 ref1,1,2 ref1,6 ref2,1 ref2,6 ** ** ** -------------------------** contact stabilising spring ** -------------------------** *element,type=spring1,elset=sprg **9001,1 9002,5001 *spring,elset=sprg 2 10.0 ** ** -------------------------------** node and element sets for output ** -------------------------------** *nset,nset=n1,gen 20,1620,20 *nset,nset=nout ref1,ref2 ** ** -----------------** initial conditions ** ------------------** *initial conditions,type=temperature

** *solid section, elset = vat, material = lead ** *material, name = lead ** *elastic 16e3,0.3,20 14e3,0.3,60 12e3,0.3,100 10e3,0.3,150 8e3,0.3,200 ** *plastic,hardening=isotropic 3.92,0,20 7.84,0.01,20 11.76,0.04,20 15.68,0.1,20 3.136,0,60 6.272,0.01,60 9.408,0.04,60 12.544,0.1,60 2.5088,0,100 5.0176,0.01,100 7.5264,0.04,100 10.0352,0.1,100 2.00704,0,150 4.01408,0.01,150 6.02112,0.04,150 8.02816,0.1,150 1.605632,0,200 3.211264,0.01,200 4.816896,0.04,200 6.422528,0.1,200 *density 1.134e-8 *conductivity 35.3e-3,20 34.0e-3,127 31.4e-3,327 *specific heat 0.16,20 0.16,127 0.16,327 *expansion 24.1e-6 ** 191

Investigations into the Dyeing Industry in Pompeii ** ------------------------------------** *step,inc=10 self weight *static 0.25,1.0 ** ** ---------------------** load to 7.5MPa = 9.4kN ** ---------------------** *cload ref2,2,-9426 ** ** -----------** gravity load ** ----------** *dload vat,grav,9.81e-3,0,-1,0 ** *end step ** ** -----------------------------------------------** step 3 - load up to 7.5 MPa & hold for 10 days ** -----------------------------------------------** *step,inc=1000,amplitude=ramp hold for 10 days 864000 seconds 7.5MPa 20C *coupled temperature-displacement,cetol=5.0e-5 10.0,864.0 ** *end step

nall,20 ** ** ** ------------------------------------** step 1 - partial load and self weight ** ------------------------------------** *step,inc=10 self weight *static 0.25,1.0 ** ** ---------------------** load to 7.5MPa = 9.4kN ** ---------------------** *cload ref2,2,-100 ** ** -----------** gravity load ** ----------** *dload vat,grav,9.81e-3,0,-1,0 ** ** -----------------------** odb database output data ** -----------------------** *output,history,freq=10,variable=preselect *node output,nset=nout u *output,field,freq=10,variable=preselect *node output,nset=nall u *node output,nset=nout rf *node output, nset=n1 u,coord *element output,elset=vat s,e ** *end step ** ** ------------------------------------** step 2 - full load and self weight

Input deck for the lead column at 40oC, used to gain the data for the input deck for lead kettle. *heading Compressive creep Sahota, Riddington 7.5MPa ** ** Cylindrical test specimen compression tested ** Diameter = 40mm, length = 80mm ** 99.99% pure lead ** Stress = 5 MPa ** Temperature =40C ** 192

Appendices

*surface, type=segments, name=lplate, ref node=ref1 start, -1.0, 0.0 line, 25.0, 0.0 ** ** upper plate ** *rigid body,ref node=ref2 *surface, type=segments, name=uplate, ref node=ref2 start, 25.0, 80.2 line, -1.0, 80.2 ** ** --------------------------------------------** contact properties between sample and platens ** --------------------------------------------** *contact pair,interaction=intlow lowint,lplate *surface interaction,name=intlow *surface behavior,pressure-overclosure=exponential 0.1,1.0 *friction 0.25 ** *contact pair,interaction=intup upint,uplate *surface interaction,name=intup *surface behavior,pressure-overclosure=exponential 0.1,1.0 *friction 0.25 ** ** ------------------** Material properties ** ------------------** *solid section, elset = vat, material = lead ** *material, name = lead ** *elastic 16e3,0.3,20 14e3,0.3,60 12e3,0.3,100 10e3,0.3,150 8e3,0.3,200 **

** -----------------------** Sample geometry topology ** -----------------------*node 1, 0,0.1 20,20,0.1 ** 1601, 0,80.1 1620,20,80.1 ** *ngen,nset=l1 1,20,1 *ngen,nset=l2 1601,1620,1 *nfill,nset=nall l1,l2,80,20 *element, type = cax4rt, elset = vat 1,1,2,22,21 *elgen,elset=vat 1,19,1,1,80,20,20 ** ** --------------------------------------------** upper lower elements in contact with rig ** --------------------------------------------** *elset,elset=lower,gen 1,19,1 *elset,elset=upper,gen 1581,1599,1 ** *surface,name=lowint,type=element lower,s1 *surface,name=upint,type=element upper,s3 ** ** -----------------------------------** Rigid surface lower and uppper plate ** -----------------------------------** *node,nset=ref1 5000,0,-1 *node,nset=ref2 5001,0,81 ** ** lower plate ** *rigid body,ref node=ref1 193

Investigations into the Dyeing Industry in Pompeii 1,1601,20 ** *boundary symm,1 ref1,1,2 ref1,6 ref2,1 ref2,6 ** ** ** -------------------------** contact stabilising spring ** -------------------------** *element,type=spring1,elset=sprg **9001,1 9002,5001 *spring,elset=sprg 2 10.0 ** ** -------------------------------** node and element sets for output ** -------------------------------** *nset,nset=n1,gen 20,1620,20 *nset,nset=nout ref1,ref2 ** ** -----------------** initial conditions ** ------------------** *initial conditions,type=temperature nall,40 ** ** ** ------------------------------------** step 1 - partial load and self weight ** ------------------------------------** *step,inc=10 self weight *static 0.25,1.0 **

*plastic,hardening=isotropic 3.92,0,20 7.84,0.01,20 11.76,0.04,20 15.68,0.1,20 3.136,0,60 6.272,0.01,60 9.408,0.04,60 12.544,0.1,60 2.5088,0,100 5.0176,0.01,100 7.5264,0.04,100 10.0352,0.1,100 2.00704,0,150 4.01408,0.01,150 6.02112,0.04,150 8.02816,0.1,150 1.605632,0,200 3.211264,0.01,200 4.816896,0.04,200 6.422528,0.1,200 *density 1.134e-8 *conductivity 35.3e-3,20 34.0e-3,127 31.4e-3,327 *specific heat 0.16,20 0.16,127 0.16,327 *expansion 24.1e-6 ** *CREEP,LAW=TIME 1.1E-4,2.0,-0.70,20 2.8e-4,2.0,-0.75,40 5.0E-4,2.1,-0.78,80 8.0E-4,2.1,-0.80,100 1.5E-3,2.1,-0.85,120 ** ** ------------------** Boundary conditions ** ------------------** *nset,nset=symm 194

Appendices

ref2,2,-9426 ** ** -----------** gravity load ** ----------** *dload vat,grav,9.81e-3,0,-1,0 ** *end step ** ** -----------------------------------------------** step 3 - load up to 7.5 MPa & hold for 10 days ** -----------------------------------------------** *step,inc=1000,amplitude=ramp hold for 10 days 864000 seconds 7.5MPa 40C *coupled temperature-displacement,cetol=5.0e-5 10.0,864.0 ** *end step

** ---------------------** load to 7.5MPa = 9.4kN ** ---------------------** *cload ref2,2,-100 ** ** -----------** gravity load ** ----------** *dload vat,grav,9.81e-3,0,-1,0 ** ** -----------------------** odb database output data ** -----------------------** *output,history,freq=10,variable=preselect *node output,nset=nout u *output,field,freq=10,variable=preselect *node output,nset=nall u *node output,nset=nout rf *node output, nset=n1 u,coord *element output,elset=vat s,e ** *end step ** ** ------------------------------------** step 2 - full load and self weight ** ------------------------------------** *step,inc=10 self weight *static 0.25,1.0 ** ** ---------------------** load to 7.5MPa = 9.4kN ** ---------------------** *cload

Input deck for the lead kettle *heading Self weight + water, plastic material, temp dependant, creep ** ----------** Base of vat ** ----------*node 1,0,0 6,0,5 2691,270,0 2696,270,5 2741,275,0 2746,275,5 *ngen,nset=l1 1,6,1 *ngen,nset=l2 2691,2696,1 *ngen, nset=l3 2741,2746,1 *nfill,bias=1.001 l1,l2,269,10 *nfill 195

Investigations into the Dyeing Industry in Pompeii *elset,elset=vbase,gen 1,1611,10 ** *surface,name=vbase,type=element vbase,s1 ** ** -------------------------------** Rigid surface representing brick ** -------------------------------** *node,nset=ref1 9999,275,-1.0 ** *rigid body,ref node=ref1 ** *surface, type=segments, name=brkrs, ref node=ref1 start, 210.0, -5.1 line, 210.0, -1.1 circl,211.0, -0.1, 211.0, -1.1 line, 285.0, -0.1 ** ** -------------------------------** Rigid surface representing floor ** -------------------------------** *node,nset=ref2 9998,-5,-175.0 ** *rigid body,ref node=ref2 ** *surface, type=segments, name=floor, ref node=ref2 start, -5.0, -175.0 line, 210.0, -175.0 ** ** --------------------------------------------** contact properties between vat base and brick ** --------------------------------------------** *contact pair,interaction=vatbrk vatint,brkrs ** *surface interaction,name=vatbrk *surface behavior,pressure-overclosure=exponential 0.1,1.0 *friction 0.25 **

l2,l3,5,10 *element, type = cax4rt, elset = vat 1,1,11,12,2 *element, type= cax4rt, elset=vat 2691,2691,2701,2702,2692 *elgen, elset = vat 1,5,1,1,269,10,10 *elgen,elset=vat 2691,5,1,1,5,10,10 ** ----------** Side of vat ** ----------*node 2801,270,5 2806,275,5 4651,270,550 4656,275,550 *ngen, nset=l4 2801,2806,1 *ngen, nset=l5 4651,4656,1 *nfill,bias=0.99 l4,l5,185,10 *element, type = cax4rt, elset = vat 2801, 2696, 2706, 2812, 2811 2802, 2706, 2716, 2813, 2812 2803, 2716, 2726, 2814, 2813 2804, 2726, 2736, 2815, 2814 2805, 2736, 2746, 2816, 2815 2811, 2811, 2812, 2822, 2821 *elgen, elset = vat 2811,5,1,1,184,10,10 ** ** --------------------------------------------** lower vat base elements in contact with brick ** --------------------------------------------** *elset,elset=vatint,gen 1631,2731,10 ** *surface,name=vatint,type=element vatint,s1 ** ** --------------------------------------------** vat base elements in contact with floor ** --------------------------------------------** 196

Appendices

1.134e-8 *conductivity 35.3e-3,20 34.0e-3,127 31.4e-3,327 *specific heat 0.16,20 0.16,127 0.16,327 *expansion 24.1e-6 ** *CREEP,LAW=TIME 1.1E-4,2.0,-0.70,20 2.8e-4,2.0,-0.75,40 5.0E-4,2.1,-0.78,80 8.0E-4,2.1,-0.80,100 1.5E-3,2.1,-0.85,120 ** ** ------------------** Boundary conditions ** ------------------** *boundary l1,1 ref1,1,2 ref1,6 ref2,1,2 ref2,6 ** ** ---------------------------------** Elsets for simple pressure loading ** ---------------------------------** *elset,elset=vatbase,generate 5,2685,10 *elset,elset=vatside,generate 2801,4641,10 ** ** -------------------------** contact stabilising spring ** -------------------------** *element,type=spring1,elset=sprg 10000,4654 *spring,elset=sprg 2

** --------------------------------------------** contact properties between vat and floor ** --------------------------------------------** *contact pair,interaction=vatfloor vbase,floor ** *surface interaction,name=vatfloor ** ** ------------------** Material properties ** ------------------** *solid section, elset = vat, material = lead ** *material, name = lead ** *elastic 16e3,0.3,20 14e3,0.3,60 12e3,0.3,100 10e3,0.3,150 8e3,0.3,200 ** *plastic,hardening=isotropic 3.92,0,20 7.84,0.01,20 11.76,0.04,20 15.68,0.1,20 3.136,0,60 6.272,0.01,60 9.408,0.04,60 12.544,0.1,60 2.5088,0,100 5.0176,0.01,100 7.5264,0.04,100 10.0352,0.1,100 2.00704,0,150 4.01408,0.01,150 6.02112,0.04,150 8.02816,0.1,150 1.605632,0,200 3.211264,0.01,200 4.816896,0.04,200 6.422528,0.1,200 *density 197

Investigations into the Dyeing Industry in Pompeii vatbase,p3,0.005 vatside,hp4,0.005,550.0,5.0 ** ** ----------**gravity load ** ----------** *dload vat,grav,9.81e-3,0,-1,0 ** ** -----------------------** odb database output data ** -----------------------** *output,history,freq=10,variable=preselect *node output,nset=nout u *output,field,freq=10,variable=preselect *node output,nset=nall u *node output,nset=ref1 rf *node output, nset=n1 u,coord *node output, nset=n2 u,coord *element output,elset=vat s,e ** ** ---------------------------------------------------** print output to msg file for further post processing ** ---------------------------------------------------** *node print, nset=n1, freq=1000 coord u *node print, nset=n2,freq=1000 coord u *node print, nset=ref1,freq=1000 rf ** *end step ** ** --- 160 heat cool cycles using 1000 sec timeframe --** ** --------------------------------------------

10.0 ** ** -------------------------------** node and element sets for output ** -------------------------------** *nset,nset=n1,gen 1,2741,10 *nset,nset=n2,gen 6,2696,10 *nset,nset=n3,gen 2811,4651,10 *nset,nset=n4,gen 2816,4656,10 *nset,nset=nall,gen 1,4656,1 *nset,nset=nout 1 ** ** -----------------** initial conditions ** ------------------** *initial conditions,type=temperature nall,20 ** ** -------------------------** heat cool cycle amplitude ** -------------------------** *amplitude,name=cycle 0.0,20, 0.6,120, 4.2,120, 5.4,20 ** ** ------------------------------------** step 1 - self weight + water pressure ** ------------------------------------** *step,inc=10 self weight *static 0.25,1.0 ** ** -----------------** water presure load ** -----------------** *dload 198

*boundary,amplitude=cycle nall,11,,1 ** *end step ** --------------------------------------------

*step,inc=1000,amplitude=ramp heat cycle 1 *coupled temperature-displacement,cetol=5.0e-5 0.025,5.4 **

Step “Heat Cycle One” was repeated a further 159 times to allow the simulation of 160 cycles.

199

Online Content https://doi.org/10.32028/9781789697421-online The online content contains: Animation showing the effect on the lead kettle of 160 dyeing cycles, equivalent to six months use. Photograph of crushed lead pipe embedded in the wall of Vat 3, property Vi5

200