Ancient Glass of South Asia: Archaeology, Ethnography and Global Connections [1 ed.] 9811636559, 9789811636554

Table of contents :
Foreword
Preface
Introduction
About This Book
Contents
Editors and Contributors
Glass Origin and Evolution
The Origin of Glass and the First Glass Industries
1 Introduction
2 The Origin of Glass
3 The First Glass Industries
3.1 The Primary Production of Late Bronze Age Glass
3.2 Historical Evidence for the Long-Distant Movement of Glass
3.3 The Working of Glass into Objects
4 Chemical Studies
4.1 Trade Between Egypt and Mesopotamia
4.2 The Origin of Glass Found in Late Bronze Age Greece
5 Conclusion
References
Glass in the Middle East and Western Europe at the End of the First Millennium CE, Transition from Natron to Plant Ash Soda or Forest Glasses
1 Introduction
2 Glass Transition in the Levant
3 Glass Transition in Egypt
4 Glass Transition in the Western Mediterranean
4.1 A Carolingian Glass Workshop in Meru
4.2 The Recycling of Lead Slag in the Area of Melle During the Carolingian Period
5 Conclusion
References
Glazed Steatite and Faience Technology at Harappa, Pakistan (>3700–1900 BCE): Technological and Experimental Studies of Production and Variation
1 Introduction
1.1 Analysis Methods
1.2 Experimental Replication
2 Indus Tradition Chronology and Terminology
2.1 Fired and Glazed Steatite
2.2 Faience—Powdered Quartz Body with Glaze
2.3 Major Faience Production Techniques
2.4 Steatite Paste and Steatite Faience
3 Glazed Steatite and Faience from Harappa
3.1 Period 1, Ravi Phase Glazed Steatite and Faience
3.2 Period 1, Ravi Phase Glaze Compositions
3.3 Period 2, Kot Diji Phase Glazed Steatite and Faience
3.4 Period 2, Kot Diji Phase Glazed Seals
3.5 Period 2, Kot Diji Phase Faience
3.6 Period 2, Faience Bangles and Vessels
3.7 Period 3, Harappa Phase Glazed Steatite and Faience
3.8 Period 3, Harappa Phase Glazed Steatite Beads and Pendants
3.9 Period 3 Glazed Steatite Button Seals and Intaglio Seals
3.10 Period 3 Faience Production
3.11 Period 3 Faience Beads
3.12 Period 3 Faience Bangles and Rings
3.13 Period 3 Faience Glazed Button Seals and Moulded Inscribed Tablets
3.14 Period 3 Faience Vessels
3.15 Period 3 Faience Figurines
3.16 Period 4/5 Late Harappa Steatite and Faience
4 Conclusion
References
Traditional Bead and Bangle Crafts in India
1 Introduction
2 Glass Beads
3 Bead Productions Cycle
3.1 Raw Material
3.2 Fuel
4 Ethnographic Survey of Bead–Bangle Production Cycle in PJAH Cluster
4.1 Firing the Furnace
4.2 Bead Making
4.3 Pendant Making
4.4 Eye Bead Making
4.5 Foiled Bead Making
4.6 Mould Bead Making
5 Bangle Making
6 Furnace
7 Traders Associated with Glass Trading
8 Mould-Makers
9 Stringing and Cleaning of Beads
10 Lamp Beads
11 Chevron/Millefiori
11.1 Chevron Making
11.2 Millefiori Beads
12 Zigzag-Patterned Bangle
13 Ethnographic Survey of the IP Bead Production Cycle at Papanaidupet
13.1 Drawing Furnace
13.2 Melting of Glass
13.3 Making the Glass Cone
13.4 Drawing the Tubes
14 Cutting the Tubes
15 Rounding Operation
16 Stringing the Beads
17 Signature of Furnace/Kiln
18 Observations
References
Scientific Study and Care of Glass
Elemental Compositions and Glass Recipes
1 Introduction
2 Ancient Glass Recipes
3 A Few Interesting Facts About the History of Glass Study
4 Analytical Techniques in General and LA-ICP-MS More Specifically
5 Case Study: Glass Beads from Kish, Iraq
6 Results
7 Conclusion
References
Isotope Analysis and Its Applications to the Study of Ancient Indian Glass
1 Introduction
2 General Considerations About Isotope Analysis
2.1 Radioisotopes
2.2 Instrumentation
3 Application of Isotope Analysis to Indian Glass
3.1 The M-Na-Al Glass
3.2 Samples
3.3 Sample Preparation
3.4 Results
3.5 Discussion
4 Conclusion
References
The Conservation of Glass
1 Introduction
2 Cleaning
3 Consolidation of Weathered Glasses
4 Repairing Broken Glass
5 Storage and Care of Glass Collections
6 ‘Crizzling’ or Atmospheric Deterioration of Glasses
7 Conclusion
References
Typology of Glass Beads: Techniques, Shapes, Colours and Dimensions
1 Introduction
2 Technique of Manufacture
2.1 Drawn Glass
2.2 Wound Glass
2.3 Coiled Glass
2.4 Folded Glass
2.5 Joined Glass
2.6 Rod-Pierced Glass
2.7 Moulded Glass
2.8 Drilled Glass
3 Shape and Dimensions
4 Length Ratio
5 Colour
6 Diaphaneity
7 Additional Information
8 Conclusion
References
Ethnography and Literature
Glass in Indian Archaeology, Ancient Literature, Historical Records and Colonial Accounts
1 Glaze
2 Faience
3 Early Glass in India
3.1 Distribution of Glass in Ancient India
3.2 Evidence for Glassmaking/Glass Working in Ancient India
3.3 Ancient Furnace
4 Ancient Indian Literature
5 Historical Records
6 Colonial Accounts
7 Furnace
8 Observations
References
Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts
1 Introduction
2 Pre-Iron Age
2.1 Bhagwanpura
2.2 Maski
2.3 Balathal
3 The Vitreous Product
3.1 Faience
3.2 Glaze, Glass and Enamel
4 Harinagar Evidence
5 Conclusion
References
History of Glass Ornaments in Tamil Nadu, South India: Cultural Perspectives
1 Introduction
2 Glass Beads from Arikamedu and Other Sites in Tamil Nadu
2.1 Glass Evidence from Kudikkadu/Karaikkadu
2.2 Comments
3 Glass Bangles in the Medieval Period
4 Glass in Literature
5 Glass in Inscriptions
6 Craft Specialization and Bangle Making/Selling Castes
6.1 Origin of the Balijas
7 Bead Production in Nagapattinam
8 Mounds of Bangle-Makers
9 Thiruvilayadal Puranam and Shiva as Bangle Seller
10 Sacred Centres and Bangle Merchants
11 Narikkuravars as Bead Sellers
12 Baby Shower (‘Bangle Protection’) Ceremony
13 Bangle Offering to River Deity
14 Gender and Ornament Use
15 Glass Artefacts as Social Markers: Ornaments of Common People and Subalterns
16 Competition in Glass Bangle Trade
17 Impact of Bangle Industry on the Society
18 Discussion and Conclusions
References
Traditional Glass Mirror Making in Kapadvanj, Gujarat, India and an Outline of the Use
1 Mirror Making at Kapadvanj
2 The Glass Works and the Furnace
3 Blowing the Glass Mirrors
4 New Knowledge
References
Glass Products in South Asia
Glass Beads of Eastern India (Early Historic Period)
1 Introduction
2 Problems of Tracing Ancient Distribution
3 Eastern Indian Glass Beads
3.1 Colour
3.2 Shape
3.3 Shapes Found in Different Colour Groups
4 Chronology
5 Situating Eastern Indian Beads in Indian Context
6 Concluding Remarks
References
A Review of Selected Glass Bead Types from the 2007–2009 Seasons of Excavation at Pattanam, India
1 Introduction
2 The Pattanam Glass Bead Corpus
2.1 Indo-Pacific Beads from Pattanam
2.2 Other Bead Types from Pattanam
3 Conclusion
References
Glass Bangles in South Asia: Production, Variability and Historicity
1 Introduction
2 Attending to Process: Different Paths to the Bangle Form in South Asia
2.1 The Khalbut: Enlargement on a Cone/Conical Mandrel
2.2 Balanced on Two-Mandrels: The Example of the Nepali Churihars
2.3 Round and Seamed: Joined Cane Bangles
3 From Antiquity to History: Interpreting the Glass Bangle
4 A Short Archaeological History of the Glass Bangle in South Asia
4.1 The Earliest Finds: New Evidence for Production and Exchange Networks
4.2 Regional Trends in the Early Historic: The North-West, Deccan and Beyond
4.3 Early Medieval Bangles
5 Conclusion: Enduring Questions in the Study of South Asian Glass Bangles
References
West Asian Glass in Early Medieval India as Seen from the Excavations of Sanjan
1 Introduction
2 The Evidence from Sanjan
3 The Glassware from Sanjan
3.1 Introduction to the Glass
3.2 Previous History of Research
3.3 Apart from the Vessels
3.4 The Vessels
4 Important/Significant Discoveries from Sanjan Include
5 Conclusion
References
Interrelations in Glass and Glazing Technologies in Mughal Tilework
1 Introduction
2 Mughal Tiling: Growth and Development
3 Glaze and Glass Technologies
4 Analytical Methods
5 Analysis
6 Results and Discussions
6.1 Plant Ash Glass and Glazes
6.2 Mineral-Soda Glass and Glazes
6.3 Colourants
7 Conclusion
References
The Diffusion of South Asian Glass
Indian Glass Beads in Western and North Europe in Early Middle Age
1 Introduction
1.1 Glass Compositions Encountered in Europe During the Early Middle Ages
1.2 Early Middle Ages Glass Beads in Western Europe
2 Method of Analysis
3 Small Drawn Merovingian Glass Beads Exhibiting Unusual High-Alumina Composition
3.1 Studied Corpus and Archaeological Contexts
3.2 Analytical Results
4 Large Opaque Red or Orange Barrel-Shaped Scandinavian Beads
4.1 Studied Corpus and Archaeological Contexts
4.2 Analytical Results
5 Discussion and Conclusion
References
Early Glass Trade Along the Maritime Silk Route (500 BCE–500 CE): An Archaeological Review
1 Scope of the Paper
2 Archaeological Evidence of Traded Glass Across the Indian Ocean/Maritime Silk Route
2.1 Western Indian Ocean Sphere
2.2 South Asia
2.3 Eastern Ocean Sphere
2.4 Beyond Southeast Asia: Southern China, Korea and Japan
3 Overview
References
Indian Glass in Southeast Asia
1 Introduction
2 Glass Analysis in South and Southeast Asia
2.1 Mineral Soda–High Alumina Glass
2.2 Potash Glass
2.3 Mineral Soda–Lime–Alumina Glass
2.4 Mineral Soda-Lime and Soda Plant Ash Glasses
3 Current State of the Research on Glass in Southeast Asia
3.1 Ornament Making at Khao Sam Kaeo and Khao Sek
3.2 The Bay of Bengal Bead Trade Network
3.3 The South India/Sri Lanka-Southeast Asia Connection
3.4 From the End of the 1st Millennium—to the Beginning of the 2nd Millennium CE
4 Summary and Discussion
4.1 Summary
4.2 Rewriting Francis’ Model
5 Conclusion
References
Indian Glass: Chronology and Distribution in Eastern Africa
1 Introduction
2 Background
3 Samples
4 Analytical Results
4.1 The M-Na-Al 6 Glass Group
4.2 Bead Colours
5 Discussion
5.1 Typology and Technology
5.2 Chronological Implications
5.3 Comparison with Southern Africa
6 Conclusion
References
Indian Glass Beads in Northeast Africa Between the First and Sixth Centuries CE
1 Introduction
2 Egypt
2.1 Quseir
2.2 Berenike
2.3 Marsa Nakari
2.4 Shenshef
2.5 Sikait
3 Nubia
4 Aksum
4.1 Adulis
4.2 Aksum
4.3 Maryam Anza
5 Conclusion
References

Citation preview

Alok Kumar Kanungo Laure Dussubieux   Editors

Ancient Glass of South Asia Archaeology, Ethnography and Global Connections

IITGN

Ancient Glass of South Asia

Alok Kumar Kanungo · Laure Dussubieux Editors

Ancient Glass of South Asia Archaeology, Ethnography and Global Connections

Editors Alok Kumar Kanungo Indian Institute of Technology Gandhinagar Gandhinagar, Gujarat, India

Laure Dussubieux Negaunee Integrative Research Center Field Museum of Natural History Chicago, IL, USA

ISBN 978-981-16-3655-4 ISBN 978-981-16-3656-1 (eBook) https://doi.org/10.1007/978-981-16-3656-1 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

Foreword

v

vi

Foreword

Preface

The Archaeological Sciences Centre (ASC) at the Indian Institute of Technology Gandhinagar (IITGN) has pursued a programme of organising short-term courses cum workshops/conferences that focus upon a selected archaeological artefact class or material. The aim of these events has been to expose a selected group of students with an acute sense of the specific problems and opportunities that are involved in the study of that material. This has taken shape in hosting a conversation between the leading experts of the field and equally to provide hands-on training in the ethnoarchaeological, experimental and scientific prospects of that particular field of archaeological research. With these objectives in mind, at these events, the resource persons are invited to speak for 45 min slots and allowing ample time for discussion. In these talks, they are required to address pre-assigned themes and topics that combine their research expertise and knowledge towards the planned content for a volume. These volumes themselves are envisaged not as the collection of papers or conference proceedings; they are intended to be valuable and useful resources on the subject. They are aimed at being useful for both students and researchers and thus to be useful for archaeological syllabi in South Asia and also globally. The first of these ‘conference cum workshops’ was held between 10 and 14 August 2015 and focused on stone beads. The proceedings of that conference have subsequently appeared as precisely such a volume meant to serve as a resource for both the teaching of stone beads and to aid further research into them (Kanungo, A.K. 2017. Ed. Stone Beads of South and Southeast Asia: Archaeology, Ethnography and Global Connections. Gandhinagar/New Delhi: IIT Gandhinagar/Aryan Books International). The second of such a conference was held from 21 to 25 January 2019 and focused on the History, Science and Technologies of Ancient Indian Glass. The book you are holding is the outcome of this second conference. The conference brought together a wide range of experts, which included archaeologists who have extensive experience of South Asian proto-glass, glass and archaeological chemists with expertise in the elemental analysis of glass. In addition, it included established ethnohistorians and ethnoarchaeologists of South Asian glass and vitreous materials, alongside craftspeople who brought their lifelong and inherited skill, expertise and knowledge. vii

viii

Preface

These five days of conference cum workshop included four days of academic presentations and two field trips, veritably covering all aspects of the study of glass. These ranged from the origin of glass and faience, to the manufacturing techniques developed at different times in South Asia and the regional distribution of key artefacts both within and as traded far outside the region. Valuably, the talks developed to papers for this book also included detailed introductions and extended examples of the analytical chemistry of ancient glasses. Finally, the field trips gave exposure to the contemporary traditional glass working at Kapadvanj and to the world famous archaeological heritage site of Vadnagar in Gujarat. The invited craftspeople at this workshop included glass bead-makers from Banaras, stone bead-makers from Khambhat and beading and mirror-stitching craftspeople from the Rabari and Mir communities of north Gujarat. An interesting experimental archaeology workshop on replication of Indus Valley faience technologies was conducted parallelly by Profs. Mark Kenoyer and Massimo Vidale. These diverse contributions brought together the challenges of studying the history, science and technology of ancient Indian glass in vivid detail. Considered together, they provided the best introduction to the complexities of regional diversity in glass traditions, the archaeometric challenges that stand before the field and the prospects of all we stand to learn from further investigations. The last major collective evaluation of the state of scientific interdisciplinary research on ancient Indian glass had been made in 1987 (Archaeometry of Glass: Proceedings of the Archaeometry Session of the XIV International Congress on Glass, 1986, New Delhi, India. Calcutta: Indian Ceramic Society). Similarly, the last monograph that had synthesised the available data on the history of Indian glass was written yet a generation earlier (Dikshit, M.G. 1969. History of Indian Glass. [Bombay]: University of Bombay). This book aims at filling precisely this gap. The description above has communicated the efforts made to provide as multifaceted, thorough and valuable an experience to the next generation of researchers, who will hopefully pose research questions and pursue methods of analysis that will build on, extend and exceed those reported here. The experts and participants at this truly international event were from eleven countries including USA, UK, France, Italy, Denmark, Cyprus, Poland, Malaysia, Thailand, Sri Lanka and India. It was gratifying to see that participants represented 54 universities, research institutes, laboratories, museums and state departments. The sixty student participants had been selected on the basis of prior interest in glass and/or ancient Indian technologies, and the conference-cum-workshop prepared them to embark on diverse research projects of their own. An ambitious series like this and workshop having a target to publish a time-bound reference manual, covering all related research areas of the topic, are not possible without the vision and support of the head of the Institute, trust of the authorities and tireless team effort of the unit in which we work. We are indebted to Prof. Sudhir Jain, Director of IITGN, for not only supporting the workshop at every stage but also giving his precious time in meeting the experts, participants and craftsmen who came for the workshop for the welfare of the centre; Prof. Amit Prashant, Dean of Research and Development, Prof. S. P. Mehrotra,

Preface

ix

Dean of External Affairs and Co-ordinator of ASC, and Prof. Michel Danino, Cocoordinator of ASC, for all the encouragement; Prof. D. P. Roy for the administrative support without entertaining any excuses yet making no reservations about the required paraphernalia; and Dr. Yadubirsingh Rawat for advocating that we develop a good publication of the outcome. The efforts put by our postdoctoral fellow and nodal officer for the conference, Dr. Oishi Roy, in organizing the event were tremendous. As always, Dr. Trupti More, Librarian of Deccan College Post-Graduate & Research Institute, played her role in promptly providing rare literature and references. Mr. Yashwant Chouhan and Mr. Shailesh Patani took care of all local logistical support, safety, local transport, campus accommodation and food. Mrs. Sunita Menon left no stone unturned to facilitate the smooth functioning of the workshop. Mr. Hatim A. Sham was the man behind the attractive posters, banners, brochures, invitations, conference tags and visuals. Mr. Devarsh Barbhaya captured the motions of the events. Ms. Shivangi Bhatt made the coordination with media look effortless. A special word of thanks to Dr. Neeldhara Misra for her professional best in managing the visual and media team. The field visits to Kapadvanj and Vadnagar were made under the guidance of Ahmed Basir Sisgar, proprietor of the Kapadvanj glass mirror workshop, and Dr. Abhijit Ambekar and his team from the Vadodara Excavation Branch of Archaeological Survey of India, respectively. Our gratitude to all the leading experts who came to IITGN presented their results, trained the students and made sure that the outcome of the workshop in book form was prepared with rigorous scientific and academic standard. Stone bead craftsmen Mr. Anwarhusain Shaikh and Mr. Pratap Bhai of Khambhat; glass bead craftsmen Nandlalji and Krishan-ji from Banaras Beads Limited; and glass beading craftswomen Meghaben Rabari, Ashaben Rabari, Sakinabe Miri, Madinabed Miri and Zanab Miri of Kutch made the workshop an experimental training reality for the participants. Prof. Sonal Mehta of CEPT University and Mrs. Niyati Kukadia of Eklavya Foundation, Ahmedabad, coordinated the beading works. Dr. Alok Kumar Kanungo would like to add a special thanks to Mrs. Banabasini Kanungo and Dr. Shahida Ansari that have been the two pillars of strength and inspiration for him while organising this conference leading to this book. He is grateful for the regular discussions and the affectionate waiting until he is done with the daily work during the six-month long preparation for the conference. We would also like to thank Scott Staszak for his help with some English languagerelated matter. If we have omitted anyone, we offer our deepest apologies. We once again express our heartfelt thanks to all who have helped in this huge endeavour. IITGN acknowledges financial support received from the Indian Council for Historical Research (ICHR), the Indian Council of Social Science Research (ICSSR), the National Science and Engineering Research Board (NSERB-DIA), the Gujarat Council on Science and Technology (GUJCOST) and the Directorate of Archaeology—Gujarat State. Gratitude is also expressed to the International Commission on Glass (ICG) and the Elemental Analysis Facility—Field Museum (FM) for timely

x

Preface

support for a few international travels, and to Banaras Beads Ltd. for logistic support for the live glass bead-making display during the conference.

Delegates during the workshop on ‘History, Science and Technology of Ancient Indian Glass’, twenty-first January 2019 at IIT Gandhinagar

Gandhinagar, India Chicago, USA

Alok Kumar Kanungo Laure Dussubieux

Introduction

There has been a growing need for books with an international reach focusing on archaeological artefacts from ancient India and South Asia encompassing scientific applications. Glass objects are part of these artefacts that have been neglected by scholars. Glassmaking and glass working are among the early pyrotechnologies and chemistry processes applied by human. Beads and bangles, being small and easy to wear, carry and trade, have been transported thousands of kilometres, across both land and sea. They have a long history of production and use in the Indian subcontinent and are still produced in the present-day traditional craft centres. Their history is preserved in the archaeological record, epigraphy and ancient literature. They are represented in sculpture and paintings, as well. Bead and bangle production techniques encompass a wide range of technologies ranging from simple to highly complex. These technologies developed and evolved through time based on both the creative inspiration of individual craftspeople and the needs to meet the demands of both local and international consumers. Beginning with the earliest glass beads dating to more than 3400 years ago at Bhagwanpura in India, furnace wound glass beads have been mastered in North India for 3000 years and furnace drawn beads have been produced in South India for 2500 years. Quickly, they became most sought-after glass products in the ancient and modern world. These beads and bangles were used by all levels of society as a way to both integrate communities culturally through the use of important symbolic objects and differentiate people by the designs and complexities of production. This book on ancient glass is the second in a series of books dealing with various artefacts that started with a book on stone beads and demonstrates important continuities from past to present. Ancient artefacts help us to better understand the importance of the past for developing new technologies in future. The book is divided into five parts. The first of these parts is ‘Glass Origin and Evolution’ that included four chapters. Thilo Rehren in his chapter on ‘The Origin of Glass and the First Glass Industries’ introduces the chemistry of glass as a matter of three different components: the sand/quartz base to which a flux is added alongside the third component—a variety of ‘spices’ to colour, opacify and lend it special qualities. Professor Rehren’s paper provides an overview of the complexity involved in the study of trace element contributions from both the flux and colourants. His xi

xii

Introduction

paper also stresses the need to locate all archaeometric analysis within a sense of the contemporary glass cultures and elite networks of political economy that sustained them. The chapter on ‘Glass in the Middle East and Western Europe at the End of the First Millennium CE, Transition from Natron to Plant Ash Soda or Forest Glasses’ by Bernard Gratuze, Nadine Schibille and Inès Pacta addresses the issue of the specificities of the transition from natron glasses to plant ash glasses and ‘forest’ glasses in the connected spheres of the Middle East and Western Europe at the end of the first millennium. Dr. Gratuze’s paper shows what chemical analysis can reveal when are combined an innovative sample selection from well dated assemblages with the precision of Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS). Chapter ‘Glazed Steatite and Faience Technology at Harappa, Pakistan (> 3700–1900 BCE): Technological and Experimental Studies of Production and Variation’ by Jonathan Mark Kenoyer summarizes the results of more than twenty years of the study of Harappan glazed steatite and faience technologies. The paper provides a sense of the pyrotechnical virtuosity and playfulness with which Harappan craftsmen excelled at the manipulation of this material. Prof. Kenoyer summarizes results not only from the use of a range of instrumental techniques [ICPMS, scanning electron microscope (SEM) and others] but also from his extensive replication studies. The chapter on ‘Traditional Bead and Bangle Crafts in India’ by Alok Kumar Kanungo summarises key insights from his ethnoarchaeological work at sites like Papanaidupet, Purdilnagar, Jalesar, Akrabad and Hasayan, to ask Indian archaeologists to be more attentive to the skill, expertise and innovations with which South Asian glass crafts developed and diversified. In doing so, the paper highlights the need to be attentive to the ‘when and why’ of changes in Indian glass craft traditions, especially in the pre-colonial era, a task in which archaeology can contribute but hitherto has not. Turning to the evidence for production, he argued that the problem in Indian archaeology persisted on account of our expectations as far as the forms of evidence and a misunderstanding of the taphonomic processes that are involved. As a result, the distinctive debris of both glass production and glass working is likely often misrecognized. There are four chapters under the second part, i.e. ‘Scientific Study and Care of Glass’. The chapter on ‘Elemental Compositions and Glass Recipes’ by Laure Dussubieux provides a synoptic overview of the kinds of questions which can be chemically asked of glass artefacts. Dr. Dussubieux very usefully organises these into three kinds of questions. First are questions that can be asked of glass making: (who made glass, where, with what technology, which ingredients, and what was the organisation of primary production). Second are questions that we can ask of trade in glass: (who traded what, what trade in raw glass existed, how networks sustained varied trade) and finally questions of the use of glass. The chapter on ‘Isotope Analysis and Its Applications to the Study of Ancient Indian Glass’ by Laure Dussubieux, Christophe Cloquet and Thomas Oliver Pryce introduces new scientific approaches to understand glass production and trade. By looking at strontium, neodymium and lead isotope signatures of ancient glass, it is possible to determine the origin of the raw materials and understand the circulation of raw materials as well as finished products. The chapter on ‘The Conservation of Glass’ by Stephen P. Koob is an introduction

Introduction

xiii

to the kinds of care which are demanded in the handling of glass. It provides a very useful and detailed discussion of the preferred binders (Paraloid B-72) that should be used in the conservation of glass. The chapter on ‘Typology of Glass Beads: Techniques, Shapes, Colours and Dimensions’ by Joanna Then-Obłuska provides a tour-de-force survey of the issues, challenges and attention to detail which the typological study of ancient glass beads demands. The paper admirably summarises the different methods by which ancient glass beads were made and provides excellent illustrations of their visible traces on artefacts. The third part, ‘Ethnography and Literature’, covers four chapters. The chapter on ‘Glass in Indian Archaeology, Ancient Literature, Historical Records and Colonial Accounts’ by Alok Kumar Kanungo dismantles the unhelpful debates over the origins of glass, glassmaking and widespread use in South Asia. The paper examines a series of otherwise difficult to understand textual references (in the Satapatha Brahmana, the Arthashastra and other texts) and points to how the metaphorical and allusive use of glass and glassmaking must presume at least a few centuries of familiarity with the material. The latter texts and social customs which define the customs related to use of glass bangles are discussed. The colonial accounts that documented the native glass production and glass bangle making in different regions of India are dealt upon. The chapter on ‘Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts’ by Bhuvan Vikrama argues, contrary to the present knowledge, that there existed a possible knowledge of glass crafts in India from at least 2nd millennium BCE on the basis of the evidence revealed by recent finds from that site. The chapter on ‘History of Glass Ornaments in Tamil Nadu, South India: Cultural Perspectives’ by Veerasamy Selvakumar is a thorough and thought-provoking review of the evidence for the production, use and status of glass in Tamil Nadu. His paper also provides a very rich account of the historical evidence on glassmakers and especially the caste of bangle traders and makers known from Tamil inscriptions. The chapter on ‘Traditional Glass Mirror Making in Kapadvanj, Gujarat, India and an Outline of the Use’ by Jan Kock and Torben Sode presents a precis of their work over the last several decades on Indian glass crafts—of primary oval shaped hot lead-coated glass mirror-making, mirror work and mirror use. The fourth part, i.e. ‘Glass Products in South Asia’, deals with five chapters. The chapter on ‘Glass Beads of Eastern India (Early Historic Period)’ by Sharmi Chakraborty addresses the important issue of how we assess the scenario of glass beads and their use in a regional perspective using new methods such as cluster analysis in the case of early historic Bengal. The chapter on ‘A Review of Selected Glass Bead Types from the 2007–2009 Seasons of Excavation at Pattanam, India’ by Shinu Anna Abraham concentrates on the non-Indo-Pacific beads of Pattanam in an attempt to understand the complexity of the trade network the site was part of. The chapter on ‘Glass Bangles in South Asia: Production, Variability and Historicity’ by Mudit Trivedi revisits the questions of chronological change, typological diversity and cultural significance of the glass bangle as an artefact type of a much-neglected point of entry into the study of South Asian glass. The chapter on ‘West Asian Glass in Early Medieval India as Seen from the Excavations of Sanjan’ by Kurush F. Dalal and Rhea Mitra-Dalal details the range and density of tenth- to twelfth-century

xiv

Introduction

glass tableware that they recovered during excavations including bottles, vials, footed plates, distillation apparatus, goblets and other items such as buttons. The chapter on ‘Interrelations in Glass and Glazing Technologies in Mughal Tilework’ by Maninder Singh Gill presents the results of his study investigating early Mughal architectural tilework. This paper is a case study of the interaction of indigenous Indian glass tradition in the context of a cosmopolitan court culture, which drew equally in its political and material cultures on Central and South Asian traditions. The fifth and final part, ‘The Diffusion of South Asian Glass’, covers five chapters. The chapter on ‘Indian Glass Beads in Western and North Europe in Early Middle Age’ by Bernard Gratuze, Constantin Pion and Torben Sode summarizes recent discovery and identification of a range of Indian glass beads in early medieval Europe in two distinct clusters. The first group of finds were from Western Europe and France in the period between 500 and 800 CE and as recovered from Merovingian era elite burials. The second and more puzzling group was that as recovered from Northern Germany, Denmark and Sweden in the seventh and eighth centuries. The chapter on ‘Early Glass Trade Along the Maritime Silk Route (500 BCE–500 CE): An Archaeological Review’ by Sunil Gupta reviews the discovery of glass across most of the Old-World civilizations from mid-second millennium BCE till the BCE–CE transition when the maritime trade in raw and crafted glass becomes widespread, with long-distance networks active from the Red Sea to the South China Sea. This paper also provides the first review of the archaeological evidence of glass trade across the Silk Route in the broad period 500 BCE–500 CE. The chapter on ‘Indian Glass in Southeast Asia’ by Laure Dussubieux draws on her decade long study of the compositional groups of glass in Southeast Asia (especially sites in Thailand, Vietnam and Myanmar). The paper demonstrates how influential models such as the Arikamedu centric story advanced by Peter Francis Jr. of technology transfer and/or the movement of craftspeople are in need of re-evaluation in the light of the elemental analysis of glass from these sites. The chapter on ‘Indian Glass: Chronology and Distribution in Eastern Africa’ by Laure Dussubieux and Marilee Wood reports on recent research on Indian glass beads found on the western rim of the Indian Ocean, highlighting the chronological shifts of Indian production centres that fed the African bead market throughout the 2nd millennium CE. The final chapter on ‘Indian Glass Beads in Northeast Africa Between the First and Sixth Centuries CE’ by Joanna Then-Obłuska presents new evidence to the South Asian audience of Indian beads as traded to Northeast Africa in the period between the first and sixth century CE. The above-mentioned chapters, written by some of the best known authorities, make the book one of its kinds with holistic approach making it a reference work on the subject. Alok Kumar Kanungo Laure Dussubieux

About This Book

This book Ancient Glass of South Asia—Archaeology, Ethnography and Global Connections provides a comprehensive research on ancient Indian glass. The contributors include experienced archaeologists of South Asian glass, and archaeological chemists with expertise in the chemical analysis of glass, besides, established ethnohistorians and ethnoarchaeologists. It is comprised of five parts, and each part discusses different aspects of glass study: the origin of glass and its evolution, its scientific study and its care, ancient glass in literature and glass ethnography, glass in South Asia and the diffusion of glass in different parts of the world. The topic covered by different chapters ranges from the development of faience to the techniques developed for the manufacture of glass beads, glass bangles or glass mirrors at different times in South Asia, a major glass-producing region and the regional distribution of key artefacts both within India and outside the region, in Africa, Europe or Southeast Asia. Some chapters also include extended examples of the archaeometry of ancient glasses. It makes an important contribution to archaeological, anthropological and analytical aspects of glass in South Asia. As such, it represents an invaluable resource for students through academic and industry researchers working in archaeological sciences, ancient knowledge system, pyrotechnology, historical archaeology, social archaeology and student of anthropology and history with an interest in glass and the archaeology of South Asia.

xv

Contents

Glass Origin and Evolution The Origin of Glass and the First Glass Industries . . . . . . . . . . . . . . . . . . . . Thilo Rehren Glass in the Middle East and Western Europe at the End of the First Millennium CE, Transition from Natron to Plant Ash Soda or Forest Glasses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bernard Gratuze, Nadine Schibille, and Inès Pactat Glazed Steatite and Faience Technology at Harappa, Pakistan (> 3700–1900 BCE): Technological and Experimental Studies of Production and Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jonathan Mark Kenoyer

3

21

39

Traditional Bead and Bangle Crafts in India . . . . . . . . . . . . . . . . . . . . . . . . . 101 Alok Kumar Kanungo Scientific Study and Care of Glass Elemental Compositions and Glass Recipes . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Laure Dussubieux Isotope Analysis and Its Applications to the Study of Ancient Indian Glass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 Laure Dussubieux, Christophe Cloquet, and T. O. Pryce The Conservation of Glass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Stephen P. Koob Typology of Glass Beads: Techniques, Shapes, Colours and Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 Joanna Then-Obłuska

xvii

xviii

Contents

Ethnography and Literature Glass in Indian Archaeology, Ancient Literature, Historical Records and Colonial Accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Alok Kumar Kanungo Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts . . . . . . . . 259 Bhuvan Vikrama History of Glass Ornaments in Tamil Nadu, South India: Cultural Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 Veerasamy Selvakumar Traditional Glass Mirror Making in Kapadvanj, Gujarat, India and an Outline of the Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301 Jan Kock and Torben Sode Glass Products in South Asia Glass Beads of Eastern India (Early Historic Period) . . . . . . . . . . . . . . . . . . 325 Sharmi Chakraborty A Review of Selected Glass Bead Types from the 2007–2009 Seasons of Excavation at Pattanam, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345 Shinu Anna Abraham Glass Bangles in South Asia: Production, Variability and Historicity . . . 361 Mudit Trivedi West Asian Glass in Early Medieval India as Seen from the Excavations of Sanjan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383 Kurush F. Dalal and Rhea Mitra-Dalal Interrelations in Glass and Glazing Technologies in Mughal Tilework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405 Maninder Singh Gill The Diffusion of South Asian Glass Indian Glass Beads in Western and North Europe in Early Middle Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427 Bernard Gratuze, Constantin Pion, and Torben Sode Early Glass Trade Along the Maritime Silk Route (500 BCE–500 CE): An Archaeological Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451 Sunil Gupta

Contents

xix

Indian Glass in Southeast Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489 Laure Dussubieux Indian Glass: Chronology and Distribution in Eastern Africa . . . . . . . . . . 511 Laure Dussubieux and Marilee Wood Indian Glass Beads in Northeast Africa Between the First and Sixth Centuries CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533 Joanna Then-Obłuska

Editors and Contributors

About the Editors Alok Kumar Kanungo is a faculty at IIT Gandhinagar and an adjunct faculty at Flinders University. He was born in Odisha and grew up in close contact with many indigenous communities of eastern and north-eastern India. His early childhood experiences led him to eventually focus on archaeological and ethnographic studies of indigenous and ancient technology. For the last two decades, Dr. Kanungo has travelled and documented the rich heritage of the Nagas of northeast India, and the Bondos and Juangs of Odisha both in the field and in museums across Europe and the UK. He has worked in many areas where it is difficult to say where anthropology or history stops and archaeology begins. He has studied and published extensively on the subject of glass and glass-bead production and written or edited fifteen books and seventy research articles and book chapters. He has been the recipient of many prestigious awards including SPARC, Humboldt, Fulbright, Rakow and Homi Bhabha Fellowships. He has lectured at many universities and research institutes in Taiwan, England, USA, New Zealand, Bangladesh, Italy, France, Turkey, Malaysia, Germany and Thailand, besides India. Laure Dussubieux is a chemist specialized in the determination of the compositions of ancient artefacts made from synthesized or natural glass, metals and stones. She obtained her Ph.D. in Chemistry from the University of Orléans (France) in 2001 with a dissertation focused on the use of laser ablation—inductively coupled plasma— mass spectrometry (LA-ICP-MS) to study the provenance and the circulation of ancient glass beads around the Indian Ocean. Prior to her appointment at the Field Museum, she was a post-doctoral fellow at the Smithsonian Institution (Museum Support Center, Maryland, USA) where she developed the application of LA-ICPMS to the study of ancient gold and the use of portable X-Ray Fluorescence to survey cultural artefacts. Since 2004, she has managed the Elemental Analysis Facility (EAF) at the Field Museum and her current title is a research scientist. At the EAF, in a little more than a decade, in addition to her own research on ancient glass from South and Southeast Asia, she has facilitated more than 150 projects addressing questions related to the archaeology of cultural production, interaction and exchange. xxi

xxii

Editors and Contributors

Contributors Dr. Shinu Anna Abraham is Associate Professor at St. Lawrence University and Archaeologist. She has done fieldwork in Egypt, Israel, India and the USA. She has two ongoing research projects: the systematic survey of iron and glass production in southern Andhra Pradesh, India, and the investigation of South India glass beads to reconstruct both Indian Ocean exchange patterns and ancient South Indian craft production processes. She is interested in the archaeology of craft/technology, state formation and archaeological theory. As Senior Editor, she published Connections and Complexity: New Approaches to the Archaeology of South and Central Asia, by Left Coast Press in 2013. Dr. Sharmi Chakraborty works as Fellow, Centre for Archaeological Studies and Training, Eastern India. Her main interest is the archaeology of the early historic period of India in West Bengal. Her doctoral dissertation has been on the early historic site of Chandraketugarh (2000). She directed exploration in the Bakreswar River Valley and conducted excavation in Paharpur (historic to early medieval) and Kusumjatra (chalcolithic). Her ethnographic work was published as a monograph (Ceramic Variability: An Ethnographic Perspective, 2018). She is Editor of Pratna Samiksha (New Series), a peer-reviewed journal of archaeology. Dr. Christophe Cloquet works as Research Engineer in geochemistry at CRPG and Head Manager of a CNRS National Facility for rocks and mineral analysis (SARM). He did Ph.D. in geosciences at CRPG, University de Lorraine, France, and postdoctoral in Montréal, Canada, at the Geotop (UQAM) and Ghent University, Belgium. He specializes in isotopic geochemistry, developing and using clean room conditions and MC-IC-PMS instrument. He has a strong track record in using isotope ratios to understand sources and processes in the critical zone. He is focused on tracing anthropogenic or natural sources by using metal isotope ratios. Dr. Kurush F. Dalal received his Ph.D. in the Early Iron Age in Rajasthan from the Department of Archaeology, Deccan College, Pune. Later, his research interest shifted to the Early Medieval Period of the West Coast of India, and since then, he has excavated the sites of Sanjan and Chandore. As Assistant Professor (archaeology) in Centre for Extra Mural Studies, University of Mumbai, he co-directed the Salcette Explorations Project, an Urban Archaeology Project, documenting the Archaeology of Mumbai. His interests in archaeology include memorial stones and ass-curse stones in India, numismatics, defence archaeology and architecture, ethnoarchaeology, culinary anthropology, food archaeology and other related subjects. Dr. Maninder Singh Gill is Art Conservator and Archaeological Scientist based in Noida, India. He trained for a MA in conservation at the National Museum Institute (NMI) and was later conferred with a Ph.D. in archaeological science by the University College London (UCL). He has been working in the field of conservation in India

Editors and Contributors

xxiii

since 1999. His interests lie in the application of scientific methods for the analyses of art and archaeological materials. He has conducted research on a wide variety of artefacts and materials from the medieval to early modern period, including architectural glazed tiles, wall paintings, stucco work and painted decorations in historic interiors. Dr. Bernard Gratuze is Director of research at the French National Center for Scientific Research (CNRS), Institut de Recherche sur les Archéomatériaux, Centre Ernest-Babelon (IRAMAT-CEB), CNRS/Université d’Orléans, France. He received his Ph.D. and the Habilitation for the direction of Ph.D. from the Analytical Sciences Department of Orléans University. His current research interest includes the development of analytical protocols using laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) for glass (as well as for lithic materials) to study their production and trade from Protohistory to the Modern Period. Dr. Sunil Gupta is Director i/c at the Allahabad Museum. He completed his Ph.D. in archaeology from the Deccan College, Pune, in 1998. He has been Nehru Fellow at the Victoria and Albert Museum, UK (1997), and JSPS Postdoctoral Fellow at the International Research Center for Japanese Studies, Kyoto (1998–99). He has done archaeological fieldwork in Japan, China, East Africa and India at Kamrej (Gujarat) in 2003 and Bankat (Allahabad district) in 2008. He is Editor of the Journal of Indian Ocean Archaeology. His current focus is the archaeology of ‘trade and civilization’ in the context of the early Indian Ocean world. Dr. Jonathan Mark Kenoyer the George F. Dales Jr. and Barbara A. Dales Professor of Anthropology, has been teaching at the University of Wisconsin–Madison since 1985. He has worked on excavations and ethnoarchaeological studies in both Pakistan and India since 1975. He has served as Field Director and Co-Director of the Harappa Archaeological Research Project since 1986. He has a special interest in ancient technologies and crafts, socio-economic and political organization as well as religion. These interests have led him to study a broad range of cultural periods in South Asia as well as other regions of the world, including China, Japan, Korea, Oman and West Asia in general. Dr. Jan Kock teaches in the Department of Medieval and Renaissance Archaeology, Aarhus University, Denmark, since 1994. He served as Curator in the famous Aalborg Historical Museum in Denmark (1974–1994). His research involves medieval archaeology, ethnoarchaeology, study of glass and history of technology, and also warfare. He is today one of the foremost European authorities in the field of beads and glass studies in general and that of India in particular. Along with Torben Sode, he has been authoring a series of research publications on various crafts of India. He has travelled and documented in almost all corners of India where there is evidence of traditional glass technology.

xxiv

Editors and Contributors

Dr. Stephen P. Koob Chief Conservator in Corning Museum of Glass since 1998, was responsible for the care and preservation of all of the museum’s collections until his retirement in 2020. He also oversaw the maintenance and repair of objects in the museum’s conservation laboratory and provides documentation of such objects throughout their repair. He is an expert in dealing with ‘crizzling’, a condition that affects unstable glass. He has recently taken over the chairmanship of Technical Committee 17, which studies the Archaeometry and Conservation of Glass, as part of the International Congress on Glass. He is Author of the book, Conservation and Care of Glass Objects (2006). Mrs. Rhea Mitra-Dalal graduated from Deccan College, Pune, with a master’s in archaeology and has a special interest in food. She also has a great love for the English language. She is Blogger and Freelance Writer as well as Food Entrepreneur and Hand-crafter. Dr. Inès Pactat completed her Ph.D. in the ‘Production, distribution and consumption of glass in France between the eighth and the eleventh century’ from the University of Burgundy Franche-Comté. She is involved in French and Croatian archaeological fieldworks and, as Member of the Executive Board of the French Association of Glass Archaeology (AFAV), has co-organized the 8th International Congress on Medieval Glass in Western Europe at Besançon in 2016. She now holds a postdoctoral position at the IRAMAT-CEB in the framework of the ERC project GlassRoutes directed by Dr Nadine Schibille. Dr. Constantin Pion is Scientific Collaborator of the Belgian Royal Institute for Cultural Heritage (Brussels) and Professor at Free University of Brussels and Royal Institute for Art History and Archaeology where he teaches medieval art history and archaeology. He received his Ph.D. in art history and archaeology from the Free University of Brussels in 2014 with a dissertation focused on Merovingian glass beads in Western Europe. His research focuses on typology, chronology, glassmaking processes and uses. In collaboration with Bernard Gratuze and Orléans University, he studies glass recipes, identifying small Indian glass beads import in Western Europe during fifth and sixth century CE. Dr. T. O. Pryce is Senior Researcher at the French National Centre for Scientific Research since 2013. He received his Ph.D. in archaeometallurgy from the UCL Institute of Archaeology (2009). A three-year Leverhulme Trust Early Career Fellowship at Oxford University led to a one-year Senior Postdoctoral Fellowship with the French Institute of Research for Development, based in Laos in 2012. He is Director of the French Archaeological Mission in Myanmar since 2012 and obtained an ANR grant for the ‘BROGLASEA’ project in 2016. His research focuses on the hunter-gatherer to state transition in Mainland Southeast Asia, with a metallurgical specialization there and in surrounding regions.

Editors and Contributors

xxv

Dr. Thilo Rehren is A. G. Leventis Professor for archaeological sciences at the Cyprus Institute in Nicosia, Cyprus, where he leads the research centre for Science and Technology in Archaeology and Culture (STARC). He worked for nearly ten years as Research Scientist at the German Mining Museum in Bochum. In 1999, he joined the University College London (UCL) Institute of Archaeology as Professor for archaeological materials and technologies. From 2011 to 2016, he led the development of UCL Qatar as a centre of excellence for archaeology, museology and conservation based in Doha, Qatar. His research covers topics dealing with ancient glass and metal technology. Dr. Nadine Schibille obtained her Ph.D. in the history of art from the University of Sussex in 2004. During her doctoral research, she developed an interdisciplinary strategy to investigate the material and aesthetic aspects of light in the art and architecture of the Byzantine Empire. Following a M.Sc. from the Institute of Archaeology at UCL (London) in 2005, she has held postdoctoral positions at Stanford University, the Getty Institute (Research Fellowship), and the University of Oxford (Marie Skłodowska-Curie Intra-European Fellowship). She joined the CNRS in 2015 to lead an ERC project entitled GlassRoutes that traces Mediterranean-wide developments in the production, trade and consumption of glass. Dr. Veerasamy Selvakumar is Faculty Member in the Department of Maritime History and Marine Archaeology, Tamil University, Thanjavur. He completed his doctoral and postdoctoral researches from Deccan College, Pune. He was Nehru Trust for the Indian Collections at the Victoria and Albert Museum (NTICVAM) Visiting Researcher at the Centre for Maritime Archaeology, Southampton University, in 2004. With a NTICVAM UK Visiting Fellowship in 2018, he was trained in ceramic studies at UCL and the British Museum. His research interests include archaeology of India, prehistory, maritime history and archaeology, archaeological theory, heritage management, history of science and technology, ceramic studies, Indian Ocean cultural interactions and ecocriticism. Mr. Torben Sode is Proprietor of Glass Bead Trading, Denmark, and Independent Glass Researcher. He has travelled all around the world in search of questions related to glass in general and glassmakers in particular. His search for continuity in tradition has resulted in some of the well-referred works on glass bead and bangle productions at Purdalpur, glass production at Jalesar and glasswork at Kapadvanj. He is an expert in glass conservation and has worked on many traditional glass-working centres in Europe as well. Invariably, his publications include references to the people who work on glass. Dr. Joanna Then-Obłuska is Assistant Professor in the Polish Centre of Mediterranean Archaeology at the University of Warsaw and Research Associate in the Oriental Institute at the University of Chicago, specialising in the archaeology of Northeast Africa. After taking part in various excavations and surveys in Israel and Egypt (2003–2007), she wrote her Ph.D. on burial customs of nomadic people in

xxvi

Editors and Contributors

ancient Egyptian deserts. Since then, her projects focus primarily on issues of society and economy, looking at ancient Sudanese and Egyptian beads and jewellery, in terms of both their materials and techniques. Dr. Mudit Trivedi is Assistant Professor of anthropology at Stanford University. His research focuses upon the intertwined archaeology of religion, conversion and urbanism at the medieval fortified city of Indor in district Alwar, Rajasthan, where he conducted survey and excavations in collaboration with the Department of Archaeology and Museums, Government of Rajasthan. Previously, he worked on projects related to the South Indian Megalithic–Iron Age and Early Historic periods, the settlement history of the middle Ganga Plains (with special reference to Kausambi) and the North Indian Palaeolithic. Dr. Bhuvan Vikrama works as Superintending Archaeologist at the Archaeological Survey of India. He obtained his master’s and Ph.D. (Decline of the Indus Valley Civilization: Socio-Economic Factors) degrees in ancient Indian history, archaeology and culture from Ram Manohar Lohia Avadh University, Faizabad. His noteworthy contributions are in excavations at Lalkot, Humayun’s Tomb and Shalimar Bagh, Delhi; Siswania, Ayodhya, Ahichhatra, Harinagar and Sakatpur, Uttar Pradesh; and Dholavira, Gujarat. He has, so far, published more than two dozens of research papers on a variety of topics. Dr. Marilee Wood is Honorary Research Associate at the University of the Witwatersrand specialized in the archaeology of glass beads. She began bead studies while living in South Africa in the 1990s. She received an MA in archaeology from the University of the Witwatersrand in 2005 and then a Ph.D. from Uppsala University in 2012. She specialises in researching and writing about glass beads, mainly from pre-European contact sub-Saharan Africa.

Glass Origin and Evolution

The Origin of Glass and the First Glass Industries Thilo Rehren

Abstract Glass is unique among the archaeological materials of the Late Bronze Age, in its production, use and social meaning. Emerging as a regularly produced substance in the mid-second millennium BCE almost simultaneously in both Mesopotamia and Egypt, we still know surprisingly little about its origin and the organization of production and distribution to the elite workshops shaping it into colourful objects. However, over the past two decades, the combination of trace element chemical analyses of glass and a reassessment of earlier excavated production debris enabled us to made massive progress in our understanding of this industry, as is summarized in this paper.

1 Introduction Glass is unique among the artificial pre-modern materials. Ceramics and plasters, which first appear more than 10,000 years ago, are earth-based substances which have more or less the same chemical composition as their raw materials, clay and gypsum or limestone, respectively. Metals are a mono- or oligo-elemental material extracted from more complex ores, fundamentally different in appearance and properties of the raw material; first emerging as a man-made material some 7000 years ago (Radivojevi´c et al., 2010), from at least 4000 years ago, metal production leaves vast amounts of waste such as slag and furnace remains in the archaeological record. Glass, in contrast, only appears in regular quantities from c 3500 years ago (Nicholson & Henderson, 2000). Significantly, as an ‘additive’ material made from the fusion of several compositionally relatively simple raw materials into a single final product of more complex composition, it leaves almost no production waste. Despite being an artificial material with its own unique working properties, during the Late Bronze Age at least glass was still seen much more as a substitute of precious stone, on a Th. Rehren (B) The Cyprus Institute, Aglantzia, Nicosia, Cyprus e-mail: [email protected]; [email protected] University College London, London, UK © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_1

3

4

Th. Rehren

Fig. 1 a–b Two samples of intensely coloured glass from the glass workshops in Qantir/Pi-Ramesse in the eastern Nile Delta. The glass is imitating red carnelian and dark blue lapis lazuli. Photos B. Schoer

par with lapis lazuli, turquoise, amethyst, jasper, obsidian, amber and such like, and like those natural precious stones it was used mostly to make decorative items such as inlays or beads and small vessels. Thus, Late Bronze Age glass was typically intensely coloured (Fig. 1a–b), and while it was often worked into polychrome artefacts exploiting the unique working properties that glass afforded, it still maintained in the final product the impression of a natural precious stone. Glass has several major advantages compared to the precious stones mentioned above which it is meant to emulate; as an artificial material, it can be relatively easily produced in significant quantities, independent of access to a possibly obscure or remote mining area, and the range and shades of coloration can be actively manipulated rather than being limited to what is available in the geological deposits. However, producing large quantities of glass requires sufficient organization to procure the necessary raw materials, while creating different colours depends on the access to and knowledge of specific metal compounds or minerals. Evidently, both conditions were sufficiently fulfilled during much of the Late Bronze Age both in Egypt and in Mesopotamia, not least through the interregional connectivity which the elite exchange network of that period provided. The term ‘glass’ is about as specific as the word ‘metal’, with several chemically very distinct types of glass now being known (for an overview, see Rehren & Freestone, 2015, and literature therein). When describing an object, it is not really sufficient to state that it is made from ‘metal’; instead, for a proper understanding of its role it is well established to state whether this object is made from iron, or copper, or bronze, or silver, tin, gold, lead, etc. The same is true for glass, although the significance of different chemical glass types is different from the recognition of different types of metal and alloy. When a new metal or alloy emerges in the history of mankind, it is added to the list of previously known metals. In contrast, in most cases, different glass compositions are used in different geographical regions, and often, a new compositional glass type tends to replace the earlier used composition, and the

The Origin of Glass and the First Glass Industries

5

coexistence of different compositional groups is often restricted to relatively short transitional periods of a few generations. In the context of Near Eastern archaeology, three main glass types dominate the spectrum, namely the Late Bronze Age plant ash glass; the Iron Age to Byzantine mineral natron glass; and the Islamic plant ash glass. Other types include mixed-alkali glass, known particularly from the end of the Late Bronze Age and Early Iron Age of Italy; various types of aluminium-rich glass found occasionally in Byzantine and Islamic western Turkey as well as abundantly in India (Dussubieux et al., 2010); the Han Chinese lead-barium glass mimicking jade; lead glass, in use at least since the Roman period; and the various potash-based forest glass types of medieval Europe. Despite their distinct composition, and in contrast to the various metals, the three main Near Eastern glass types are visually almost indistinguishable; their appearance is not determined by the base glass chemistry, which almost inevitably produces a near-colourless transparent material, but by the range of minor impurities or additives that provide distinct colours and opacity. Instead, the distinct chemical signatures that identify one or the other compositional glass type are contained in the minor and trace elements and require appropriate chemical analysis to positively identify them.

2 The Origin of Glass There are different types of ‘origin’ of glass that need to be distinguished. The idea of producing a glassy substance has been around for several millennia before proper glassmaking emerged. The clearest evidence for this is the production of faience, consisting of a body of clean-crushed quartz coated by a thin layer of glass which formed during the firing of the object. Solid glass objects appear much later, around 1600 to 1500 BCE both in Mesopotamia and in Egypt, using the same basic raw materials. These are crushed quartz as the main material, plant ash to help make it melt during firing, and typically copper oxide to make it blue-green in colour. Also the type of objects made from glass is very similar to that made from faience. Does this mean that the idea of glassmaking has its origin in the earlier faience industry? This is quite possible, given the similarity of raw materials, but there does not seem to be a transitional stage where faience became more and more ‘glassy’ until it became glass proper. However, already Peltenburg (1987: 20) points out that the way how faience was worked, namely being shaped while cold before being fired to become rigid, is fundamentally different from the way how glass was worked while hot and states that ‘faience working may have played no more than a minor role in the establishment of glass-working’ (Peltenburg, 1987: 19). To this can be added that glass is made in a two-stage process where glassmaking and glass working are well separated, while faience has more in common with pottery production, where also a previously plastic raw material is shaped while cold and then fired into a rigid form. Others have explored potential cross-craft inspirations from metallurgy to glassmaking (most forcefully but unfounded Dayton (1978) who suggests that the LBA cobalt-blue glass of Egypt is in fact the waste of silver smelting in central

6

Th. Rehren

Europe; for more likely interactions, see also Peltenburg, 1987: 20, 22; Mass et al., 2002 and Rehren, 2003), but these are more related to shared use of specific materials than to the inception of glass as a new material more generally. While for the last half century or more it had been generally understood that glassmaking was invented in Mesopotamia, a recent reassessment of the dating evidence puts this into question (Shortland et al., 2018). In any case, it is noteworthy that physical evidence for the making of glass during the Late Bronze Age is only known from several sites in Egypt (see below), but not Mesopotamia. However, Shortland and co-workers have convincingly shown that the chemical signature of glass found in Mesopotamia is systematically different from that of glass found in Egypt, on the basis of several diagnostic trace elements such as zirconium, chromium, titanium and lanthanum (Shortland, 2005; Shortland et al., 2007), and that there were likely several separate production sites in the Near East, too (Shortland et al., 2018). Thus, it is reasonable to assume that there were indeed several sites in Mesopotamia making glass from its raw materials, which just have not yet been found in the archaeological record. The making of glass from its raw materials is quite different from the making of glass objects. It involves the collection and processing of the raw materials, usually silica either as finely crushed quartz pebbles or sand, a flux such as plant ash or mineral natron and specific additives to provide colour and/or opacity to the raw glass. Glassmaking fuses these very clean raw materials with little if any waste left behind, making it far less archaeologically visible than metal smelting, which generates large amounts of durable slag. Glass working or manufacturing, in contrast, is done by artists shaping the existing glass by heating it to a state of sufficient softness for drawing, moulding or other mechanical manipulation. This leaves slightly more archaeologically visible waste, typically fragments of coloured glass rods (Fig. 2), mis-shaped or broken fragments and discarded drips and blobs of glass not fit for recycling. The best example of this sort of material from a Late Bronze Age context has been excavated over a century ago by Flinders Petrie at Amarna (Petrie, 1894) and underpins to this day much of our understanding of LBA glass-working techniques (e.g. Lilyquist & Brill, 1993). Both unworked glass ingots and finished glass objects are known to have been transported long distances, as part of the Late Bronze Age network of elite gift exchange; the c. 175 glass ingots weighing around 2 kg each and several amphorae holding tens of thousands of glass beads from the Uluburun shipwreck (which sank c. 1300 BCE off the southern coast of Turkey) are the best archaeological example of this (e.g. Pulak, 2008), but iconographic and textual evidence also demonstrates long-distance movement of glass (Shortland, 2012). Thus, archaeological glass has at least four points of origin or provenance, beyond the still open question of the conceptual origin of glass as an artificial material mimicking precious stones: the geological origins of the various raw materials which can be traced by geochemical methods; the production origin, that is the furnace in which these raw materials were fused into glass; the stylistic origin reflecting the artisans working in the studio in which the raw glass was shaped into its desired form, typically involving glass of different colours and potentially different geological or furnace origins; and then

The Origin of Glass and the First Glass Industries

7

Fig. 2 Small selection of the countless glass rods excavated by Petrie at Tell el-Amarna in Egypt, indicating the former presence of a glass workshop there producing polychrome objects such as beads and core-formed vessels. UCL Petrie Museum of Egyptian Archaeology, London

the archaeological origin, that is the site where the object was finally discarded or deposited, and later retrieved by archaeologists or other individuals. Unfortunately, this latter origin is sometimes obscured by a further ‘provenance’, namely that of the arts market with its often missing, incomplete or outright false documentation of find spot. Historically, most science-based glass research focused on finished artefacts, taken from consumption sites such as tombs or elite buildings. While this produced a good understanding of the chemical compositions of this finished glass (e.g. Brill, 1999; Lilyquist & Brill, 1993), and the art historical trends and developments in its manufacturing (e.g. Schlick-Nolte, 1968; Stern & Schlick-Nolte, 1994; Tait, 2012, to mention just a few key works), it did little to reveal the ways of making glass, or to distinguish the original production sites of the raw glass. In this text, the focus will therefore be on the production and manufacturing evidence of glass in the Late Bronze Age of the Eastern Mediterranean, and what we now know—and do not know yet—about the organization of this industry.

3 The First Glass Industries The term ‘industry’ here is used relatively loosely to refer to a wider system of production and distribution of a specific commodity that follows certain planned,

8

Th. Rehren

regulated and repeated processes, implying a degree of oversight and management, as well as a significant scale of production beyond what one would call subsistence level. Any discussion of the glass industry, therefore, depends on a minimum level of information regarding raw material procurement, technical production processes and distribution of the products. Until about 20 years ago, Late Bronze Age glass production was known almost exclusively only from the surviving products, and any discussion of the organization of their production had to be restricted to differences in stylistic features evident from the finished products (e.g. Schlick-Nolte, 1968), with only limited reporting of ‘glass slag’ and other putative glass-related waste from glassmaking from several sites in Egypt and Mesopotamia. A critical assessment of this evidence indicated that most of it was in fact either not related to glass at all or related to glass working, as in the workshops from Tell el Amarna and Lisht, both in Egypt, or indeed the storage and movement of glass in the form of ingots, such as those from a storage context in Tell Brak in Syria (Oates, 1987: 187; Shortland et al., 2007). Clearly, the presence of ingots at an archaeological site does not indicate local production since they were often transported, and thus likely out of their production context—as is patently obvious from the largest known find of Late Bronze Age glass ingots, from the Uluburun shipwreck. Since then, however, primary glassmaking has been unequivocally demonstrated in several sites in Egypt, including Qantir/Pi-Ramesse (Rehren & Pusch, 2005; Pusch & Rehren, 2007), Amarna (Smirniou & Rehren, 2011) and Lisht (Smirniou et al., 2018), and this, together with new compositional and other evidence, now facilitates at least an initial discussion of the Late Bronze Age glassmaking industry.

3.1 The Primary Production of Late Bronze Age Glass The rich evidence from Qantir enabled a detailed reconstruction of the glassmaking techniques used there (Pusch and Merkel & Rehren, 2007; Rehren, 2007; Rehren & Pusch, 2007, 2008), laying the groundwork to argue the case also for Amarna and Lisht, where the surviving evidence is far less immediately conclusive. In all cases, glass was being made by crushing and grinding white quartz pebbles into a flour-like powder, mixing it with plant ash and firing it initially at relatively modest temperatures (estimated to around 900 °C) in specific ovoid ‘beer’ jars reused for this purpose into an intermediate or ‘semi-finished’ glass, still rich in residual quartz grains and half-reacted lime-rich crystals (Fig. 3a–b). A second firing, probably with additional plant ash and at somewhat higher temperatures (estimated to be around 1050 °C), then led to the formation of a fully fused and clear glass to which the necessary colourant was then added. This second melting and colouration took place in purpose-made cylindrical crucibles, of which large numbers have been found in both Qantir and Amarna (Fig. 4a–b). These crucibles, representing the final stage of the primary glass production (Rehren, 1997), enabled the first glimpse into the organization of the industry (Rehren et al., 2001). In Qantir, the vast majority of glass colouring crucibles have remains of red glass in them, coloured by the addition of

The Origin of Glass and the First Glass Industries

9

Fig. 3 a Sherds of a reaction vessel from Qantir/Pi-Ramesse, used to produce semi-finished glass (bright white layer, top right and lower left corner). The glass is separated from the ceramic by a dull white parting layer made of lime and a small amount of clay. b Close-up of another fragment with adhering semi-finished glass, rich in residual quartz grains. Photos Excavation Qantir

small amounts of copper oxide, while the overwhelming evidence from the crucibles from Amarna is for the production of dark blue glass, coloured by the addition of minute amounts of cobalt oxide. Further differences between the two sites exist in the size of the crucibles, with those from Qantir generally being taller. Systematic differences in the size and shape of copper-blue and cobalt-blue glass ingots, respectively, have also been identified among the Uluburun ingots (Nicholson et al., 1997), similarly indicating that two separate workshops were making these (Rehren et al., 2001). Despite these differences in the colour, size and shape of the end product, there are overwhelming similarities which clearly place the glassmaking practice at the various sites into the same overall recipe and tradition. Most significantly, they share the same type of basic raw materials, and all vessels employed in this process, the reused domestic ovoid jars and the specially made technical crucibles, were fitted with a specific lime-rich ‘parting layer’ which prevented the liquid glass from sticking to the ceramic surface, thus facilitating the clean removal of the product from the vessels (Merkel & Rehren, 2007; Turner, 1954). There is some evidence also from Lisht, despite the much more limited amount of material available for study and the lack of contextual documentation, to demonstrate that the technology was the same throughout all three sites, and nothing has been found yet that would not be consistent with the model just outlined. Thus, it seems that in Egypt for several centuries glassmaking was an established and well-controlled activity, consistently producing material of almost identical quality and finish, but of different colours and with subtle systematic compositional differences, at several different sites. Unfortunately, as mentioned above no archaeological evidence has yet been found for glassmaking in Mesopotamia or elsewhere in the Near East, making it very difficult to even discuss whether the glassmaking practice outside Egypt used the same tools and techniques, or followed its own separate traditions. However, the chemical studies, particular of minor oxides and trace elements, do show that the same recipe was used in both Mesopotamia and Egypt, of quartz pebbles, plant ash

10

Th. Rehren

Fig. 4 a Four well-preserved cylindrical crucibles from Qantir/Pi-Ramesse. Photos Excavation Qantir. b Drawings of another four cylindrical crucibles, with a range of internal profiles. Drawings Excavation Qantir

The Origin of Glass and the First Glass Industries

11

and a range of metal oxides for colouring and opacification, with regionally specific differences that set the Mesopotamian glass clearly apart from the Egyptian glass (Shortland et al., 2007).

3.2 Historical Evidence for the Long-Distant Movement of Glass The recent reassessment of the date of the glass finds from Nuzi in modern-day Iraq, part of upper Mesopotamia, has thrown open again the question of who invented glassmaking (Shortland et al., 2018). Intriguingly, even though the archaeological evidence for Late Bronze Age glassmaking is entirely of Egyptian origin, we are left with the assertion of Egyptian Pharaohs, in particular Thutmose III, that as part of the bounty from their military campaigns into Syria they brought home Syrian glass workers, who then presumably established the Egyptian glass industry. The lop-sided archaeological evidence for glass production in Egypt only, when we can reasonably assume that glass had also been made in Mesopotamia, is balanced by the epigraphic and iconographic evidence for import of glass into Egypt from countries to the East. These mostly take the form of tomb paintings depicting crates of coloured lumps being offered as tribute to Pharaoh (e.g. Shortland, 2012, Fig. 7.5). The associated inscriptions speak of ‘true’ lapis lazuli and turquoise, and baskets next to them, holding round blue-coloured discs, are labelled by the tomb painters simply as unqualified lapis or turquoise, strongly suggesting that indeed these latter are ingots of coloured glass that are delivered together with real precious stones, i.e. those of geological origin (Shortland, 2012: 140). The import of glass into Egypt is further corroborated by a good number of the Amarna letters, where either Pharaoh Akhenaten requests local rulers in the Levant to provide glass, or the rulers humbly present their glass shipments to Pharaoh. Examples of these transactions read like this (Moran, 1992): EA 370, title: ‘From the Pharaoh to a vassal’ "Say to Idiya , the ruler of Ašqaluna: Thus the king. He herewith dispatches to you this tablet-letter, saying to you, Be on your guard. You are to guard the place of the king where you are. The king herewith sends to you Irimayašša … And know that the king is hale like the Sun in the sky. For his troops and his chariots in multitude, from the Upper Land to the Lower Land, the rising of the sun to the setting of the sun, all goes very well".

And in response (to a different but presumably similar letter): EA 235, ‘An order for glass’ “Say to the king, my lord, my Sun, my god, the Sun from the sky: Message of Sitatna, your servant, the dirt at your feet. (I pr)ostrate myself at the feet of the king, my lord, my Sun, my god, 7 times and 7 times. ([at the feet of the king, my lord])-(emphasis-?). [I] have obeyed the [or]ders of the king’s comm[issioner] to me, to guard the citie[s f]or the king, my lord. I have guarded very carefully. M[oreover], the king, my lord has wri[tten] to me for glass,

12

Th. Rehren [and] I herewith send 50 (units), [their] weight-(i.e. I herewith send: ‘50–weight’), to the king, my lord.”

Or, indeed: EA 323, title: ‘A royal order for glass’. “To the king, my lord, my god, my Sun, the Sun from the sky: Message of Yidya, your servant, the dirt at your feet, the groom of your horses. I indeed prostrate myself, on the back and on the stomach, at the feet of the king, my lord, 7 times and 7 times. I am indeed guarding the [pl]ace of the king, my lord, and the city of the king, in accordance with the command of the king, my lord, the Sun from the sky. As to the king, my lord’s, having ordered some glass, I [her]ewith send to the k[ing], my [l]ord, 30–(“pieces”) of glass. Moreover, who is the dog that would not obey the orders of the king, my lord, the Sun fr[o]m the sky, the son of the Sun, [wh]om the Sun loves?” -EA 323, lines 1–23 (complete).

Others are slightly less subservient, but still reflect a clear difference in hierarchy: EA 148, A letter from Tyre “The king, my lord, has written for glass. I give to the king, my lord, what I have on hand – 100 (units) in weight.” EA 331, Letter from Sipti-B(a)l, mayor of Lakisha. “And as to the king, my (l)ord’s, (having) ordered w(hate)ver glass (I) may have on hand, I herewith (s)end (it) to (the kin)g, my lord, my god...”

There are several key messages in these letters that are relevant to our discussion. Firstly, glass was important enough for Pharaoh’s direct official requests for it, suggesting that the production in Egypt was insufficient to satisfy local demand. Thus, significant quantities of glass are requested by Akhenaten despite the glassworks at Akhetaten, the ancient city at the site of Tell el-Amarna, producing their own glass. Secondly, glass is being sent in ‘pieces’, and in substantial quantities: 30, 50 and 100 units being mentioned in the letters cited here. Thirdly, the glass orders are indiscriminate: ‘having ordered some glass’, ‘whatever glass I may have on hand’. In contrast, the labels in the tomb paintings are more specific, referring to ‘true lapis lazuli’ or ‘true turquoise’ as well as just ‘lapis lazuli’ and ‘turquoise’, where the latter is thought to refer to artificial material of the appropriate colour, that is glass. Another aspect emerging from these letters, not often discussed so far in the literature, is what they say regarding the control over glass production. While traditionally, glassmaking in Egypt at least has been linked strongly to royal control, such as the ‘palace workshops’ in Amarna (Petrie, 1894), it seems here that multiple minor city states in the Levant had regular access to substantial quantities of glass, alongside the plethora of other commodities they were trading. The third strand, of archaeological evidence for the long-distance movement of glass, comes from the shipwreck found at Uluburun on the south Turkish coast. Among the rich cargo of this vessel were around 175 disc-shaped glass ingots, almost all of blue colour (Fig. 5a), and falling into two discrete groups: a slightly larger and thicker variety coloured by cobalt and a slightly smaller and thinner variety coloured by copper (Fig. 5b, based on data from Nicholson et al., 1997). The direction of this ship and its cargo is not entirely clear, but most assumptions are that it was travelling

The Origin of Glass and the First Glass Industries

13

Fig. 5 a Five of the c. 175 cylindrical glass ingots excavated from the Uluburun shipwreck, at an exhibition at the Deutsches Bergbau-Museum in Bochum (Germany) in 2005/2006. b Comparison of thickness and diameter of a selection of glass ingots from the Uluburun shipwreck. The 12 cobaltcoloured ingots are systematically larger than the five copper-coloured ones. Data from Nicholson et al. (1997)

14

Th. Rehren

westwards, from the Levant towards the Aegean (Goren, 2013; Pulak, 2008); thus, it may have had either Mesopotamian or Egyptian glass on board, or even both, in keeping with the multicultural nature of its cargo overall. The chemical trace element signature of all analysed glass ingots from Uluburun, however, strongly points to their Egyptian origin, regardless of colour.

3.3 The Working of Glass into Objects Glass working differs in several fundamental ways from glassmaking; firstly, it does not involve chemical operations nor access to a range of specific raw materials. It only requires the ability to soften existing glass sufficiently to shape it, a good sense for aesthetically pleasing forms and colour matches and of course access to finished coloured glass. It is, therefore, argued here that there is no reason why glass working should be done by the same people who made the glass, or in the same workshops. Instead, it is likely that the artisans who were producing objects were different from the craftsmen making the glass and probably worked in different workshops. These workshops may have been contiguous within the same (palatial) compound or may have been separated by very long distances; only excavation of relevant sites can answer this for individual cases. Archaeological evidence exists for artistic glass studios in Egypt, Greece and Mesopotamia (Petrie, 1894; Panagiotaki et al., 2005; Shortland, 2012), and it is reasonable to assume that more existed which have not been identified yet. One important difference though exists between glassmaking and glass-working sites: from the evidence so far is seems that glassmaking sites were specializing in the production of just a few colours, such as copper red in PiRamesse and cobalt blue in Amarna, both with just a little evidence for the production of copper-blue glass, too, while the glass-working centres evidently had access to a much wider range of colours, with copper blue as the by far dominating colour in terms of quantity worked (Rehren et al., 2001). In summary, we have the following (few) pieces of evidence for our jigsaw. Glassmaking may or may not have been a Mesopotamian invention, before arriving as a fully developed technology in Egypt. Evidence for glassmaking is so far restricted to Egypt, where now several sites are documented to have been making glass from raw materials and producing coloured glass ingots. There is a indication that these workshops specialized in just one or two more demanding colours while also making standard copper-blue glass (see Pusch & Rehren, 2007: 158–163 for a fuller discussion of this). Glass ingots were subject to long-distance trade and exchange, facilitating access to coloured glass for glass studios which did not have the ability to make (all) their own glass (Fig. 6). Glass working was practised in all major centres, and there are pretty clear stylistic differences between artefacts made in Mesopotamian, Egyptian and Greek workshops. The evidence for the location of such workshops demonstrates that in all three cultural spheres they are mostly set within elite contexts, such as palaces or temples, underlining the elite nature of the material. An interesting observation is that the colour spectrum differs between the three main glass-using

The Origin of Glass and the First Glass Industries

15

Fig. 6 Abstract representation of the Late Bronze Age glass industry arranged by colour, where light blue glass coloured by copper or bronze was produced widely in Mesopotamia and Egypt, respectively, while specific workshops specialized in the rarer colours, such as red in Qantir/Pi-Ramesse, cobalt blue in Amarna, and the antimony-based colours probably in northern Mesopotamia or the Caucasus region, near the Late Bronze Age antimony mines. From Pusch & Rehren (2007: 162, Fig. 113), with permission of the authors

regions; in particular, the Aegean seems to be restricted almost exclusively to light and dark blue with very little white glass, while in Mesopotamia and Egypt other strong colours, such as black, red, amethyst, yellow and white are rather common, in addition to the dominant light blue. Significantly, dark blue glass, coloured by cobalt, seems to be rather common in Egypt, but for much of the Bronze Age absent from Mesopotamia.

4 Chemical Studies Mesopotamian and Egyptian glass is in its essence chemically rather homogenous soda-lime silica glass with slightly elevated levels of potash and magnesia, with little variation beyond the specific additives such as copper, cobalt, lead, antimony and manganese needed to impart the colour. There is no systematic difference in composition between glass from the mid-second millennium BCE and that from the late second millennium BCE, nor between different sites and regions where the glass had been found. Over the last 10–15 years, through archaeological and analytical work, systematic differences in trace elements were identified between glass found in Mesopotamia and that found in Egypt (Shortland, 2005; Shortland et al., 2007). The main diagnostic criterion is the ratio of the two elements chromium and lanthanum, which for Mesopotamian glass is around 8 to 10, while for Egyptian glass it is mostly between 2 and 3 (see Fig. 7). This separation, with some limited overlap in the region of low trace element concentrations, is further supported by distinct differences in the

16

Th. Rehren

Fig. 7 Position of Uluburun glass ingots (blue squares) in the trace element diagram for lanthanum (La) vs chromium (Cr) shows that they are consistent with Egyptian glasses (circles), but differ from the Mesopotamian glasses (red triangles). In contrast, the analysed Mycenaean glass beads (green circles) match mostly the Egyptian pattern, but some also fit the Mesopotamian signature. Modified from Smirniou (unpubl.)

levels of both zirconium (Zr) and titanium (Ti), with Mesopotamian glass typically having less than 20 ppm Zr and less than 300 ppm Ti, while most Egyptian glass has higher concentrations in these elements, reaching 60 ppm Zr and 600 ppm Ti. It is reasonable to assume that the differences in glass found in these two regions reflect different production centres, with the stone tools used to crush and grind the hard quartz pebbles into fine flour-like powder likely playing a significant role in the origin of these trace elements (Rehren & Pusch, 2008). On this basis, it is now possible to revisit the whole situation concerning Late Bronze Age glass production and consumption, based on evidence rather than speculation. The emerging picture so far is bound to be subject to change as further analyses become available, but the following observations can be made at present.

4.1 Trade Between Egypt and Mesopotamia All available data point to a very strong separation between the two major glass-using regions, with almost all analysed objects from Mesopotamia having the same trace element signature, regardless of their archaeological find spot, being distinct from the signature of Egyptian-found glass. Thus, on the basis of this powerful tool and the increasing number of analyses, there is now some limited indication of regular trade in glass between the two regions. It has to be mentioned though that this picture is based on the analyses of glasses mostly from the second half of the Late Bronze Age. Most recently, Kemp et al., (2020) analysed some of the earliest glass beads found in Egypt and found them to be of typical Mesopotamian composition, consistent with the assumption of an early import of glass from Mesopotamia into Egypt. Still, the later period of apparently near-complete self-sufficiency of Egypt in her glassmaking coincides with the iconographic and textual reports of substantial glass imports into

The Origin of Glass and the First Glass Industries

17

Egypt mentioned above, which therefore form an interesting contrast to the analytical picture.

4.2 The Origin of Glass Found in Late Bronze Age Greece So far, there is now a picture emerging of an industry comprising two major glassproducing regions in Mesopotamia and Egypt, respectively, both following the same overall technology but with their own trace element signatures and distinct styles, demonstrating the existence of glassmaking and glass-working centres in both regions. Judging from the analytical evidence, the interaction between these two appears to be very limited. This is in contrast to the textual and pictorial evidence, leaving the door wide open for further research and modification of the current picture. An intriguing counterpoint to these two independent but coexisting industries is given by the Late Bronze Age glass from Greece. Countless glass objects have been found in excavations of Late Bronze Age sites there, mostly as beads and pendants which, stylistically, are clearly Mycenaean. A major study of such glass finds using electron microprobe analyses was done by Nikita and Henderson (2006), concluding that there is a strong likelihood that a discrete Greek glass production existed. However, the limited number of trace element analyses available to date does not support this; rather, it strongly indicates that all glass found in Greek Late Bronze Age contexts can be assigned to either an Egyptian or a Mesopotamian origin (Fig. 7) (Smirniou & Rehren, 2013). Thus, there is little reason at present to assume an independent Greek glass production, even though glass working was well developed there. This cautious view is further supported by the clear Egyptian signature of the analysed glass ingots from the Uluburun shipwreck (Jackson & Nicolson, 2010), suggesting that the large number of copper-blue and cobalt-blue glass ingots on board the Uluburun vessel was of Egyptian origin, and en route to a Greek destination. This view is also consistent with the very limited colour range available to Greek artisans, consisting almost exclusively of light and dark blue.

5 Conclusion Two different and partly contradictory pictures of the Late Bronze Age glass industry emerge from the foregoing observations. Ample textual and iconographic evidence points to a substantive flow of glass from the Levant and northern Mesopotamia into Egypt, mostly in the form of tribute or gift exchange. This contrasts with the lack of primary glassmaking evidence outside Egypt, and the consistent Egyptian trace element signature for all but the earliest analysed glass finds from Egypt, and the similarly consistent but different Mesopotamian signature for all analysed glass finds from Mesopotamia. Thus, there is extensive talk of glass trade, but no hard evidence for it.

18

Th. Rehren

Archaeometric research, meanwhile, has demonstrated that there were several distinct glassmaking centres in Egypt and most likely also in Mesopotamia. At least three Egyptian glassmaking sites have so far been identified archaeologically, and the evidence available indicates that all three workshops followed closely the same traditions and practices when making glass. The observed differences are at the level of individual ‘hands’ doing the same thing in a slightly different manner, such as making the same cylindrical crucibles in a slightly larger or smaller diameter, or filling them to different heights in different workshops. While this level of detail is not yet available from Mesopotamian glass workshops, it is already clear from the chemical analyses that the glass workers there followed the same basic recipe of using crushed quartz pebbles and plant ash as a flux—just that the stone tools they used for grinding the quartz into dust were made from their local more basaltic rocks, while the Egyptians apparently used more granitic stone tools reflecting the abundance of this rock type in the Nile Valley. These, in turn, left different geochemical signatures in the glass as a result of contamination of the ground quartz dust. Similarly, the plant ash in Mesopotamia had a slightly different composition to the one in Egypt, reflecting the different soil chemistry in the two regions. In contrast, there is no evidence, textual or archaeological, of primary glassmaking in Greece, while all analysed Mycenaean-style glass artefacts have either a Mesopotamian or an Egyptian trace element signature. This is consistent with the analytical work done so far on the Uluburun glass ingots, which all seem to be of Egyptian origin, and may well have been on their way to the Aegean. Several major lacunae come immediately to mind when looking at this picture. Firstly, there is a clear need for more trace element analyses of the earliest glasses found in Egypt, in particular those from the three foreign wives of Thutmose: Is their trace element signature Mesopotamian or Egyptian? The work done by Kemp et al. (2020) on the glass beads from Gurob in Egypt, demonstrating their Mesopotamian origin, shows the potential of this approach. Secondly, where is the archaeological evidence for glassmaking in Mesopotamia? Here, a rather pessimistic comment by Petrie, made nearly a century ago, may still be valid: ‘[…] a good deal more knowledge of the history of glass might become available when Asia was really opened up for proper investigation and study, but for this to happen it might be necessary to wait for another cycle of civilization. […] and then – and probably not until then – one might really get to know, for the first time, what was the actual origin of glass-making’ (Petrie, 1926: 233). Thirdly, is it reasonable to assume that the glass sent to Egypt from the Levantine city states such as Ashkelon and Gaza was really made there, or may they in turn have obtained the glass from elsewhere—and possibly not from further East in Mesopotamia, but from a production site somewhere in Egypt, in some sort of circular trade? The latter hypothesis would resolve the apparent discrepancy between a purely Egyptian trace element signature found in all analysed Egyptian glass artefacts and the large number of glass ingots mentioned as imports into Egypt. What does this then say about the control over glass production and trade in Late Bronze Age Egypt, when these minor city states have such ample access to glass, and even Pharaoh has to ask for them to deliver it (back) to Egypt?

The Origin of Glass and the First Glass Industries

19

These may be some still rather fundamental questions; however, it is a major step forward compared to twenty years ago that we are now able to ask these very specific questions rather than debating whether or not Egypt was at all making her own glass (Newton, 1980: 176 writes: ‘the Egyptians could only melt other people’s glass even though they could fabricate the most exquisite items from it … the Egyptian court depended for their basic raw material, or for an essential ingredient thereof, on imports from Asia’). Furthermore, the newly available analytical methods, particularly laser ablation ICP-MS with their minimally invasive sampling and very low detection limits for trace elements, make it feasible to generate the data necessary to address and possibly even answer some of these questions in the not too distant future.

References Brill, R. H. (1999). Chemical analyses of early glasses (Vol. I and II). Corning Museum of Glass. Dayton, J. (1978). Minerals, metals, glazing and man, or who was Sesostris I? Harrap. Dussubieux, L., Gratuze, B., & Blet-Lemarquand, M. (2010). Mineral soda alumina glass: Occurrence and meaning. Journal of Archaeological Science, 37, 1646–1655. Goren, Y. (2013). International exchange during the late second millennium B.C.: microarchaeological study of finds from the Uluburun ship. In J. Aruz, S. B. Graff, Y. Rakic (Eds.), Cultures in contact, from Mesopotamia to the Mediterranean in the second millennium B.C. (pp. 54–61). Jackson, C. M., & Nicholson, P. T. (2010). The provenance of some glass ingots from the Uluburun shipwreck. Journal of Archaeological Science, 37, 295–301. Kemp, V., McDonald, A., Brock, F., & Shortland, A. J. (2020). LA-ICP-MS analysis of Late Bronze Age blue glass beads from Gurob, Egypt. Archaeometry, 62, 42–53. Lilyquist, C., & Brill, R. H. (1993). Studies in early Egyptian glass. Metropolitan Museum of Art. Mass, J. L., Wypyski, M. T., & Stone, R. E. (2002). Malkata and Lisht glassmaking technologies: Towards a specific link between second millennium BC metallurgists and glassmakers. Archaeometry, 44, 67–82. Merkel, S., & Rehren, Th. (2007). Parting layers, ashtrays, and Ramesside glassmaking: An experimental study. In E.B. Pusch, & Th. Rehren (Eds.), Hochtemperatur-Technologie in der Ramses-Stadt—Rubinglas für den Pharao (pp. 201–221). Gerstenberg. Moran, W. L. (1992). The Amarna letters. Baltimore (Md.). Newton, R. G. (1980). Glass, including stained glass. Conservation of Materials Series. Nicholson, P. T., & Henderson, J. (2000). Glass. In P. T. Nicholson & I. Shaw (Eds.), Ancient Egyptian materials and technology (pp. 195–224). Cambridge University Press. Nicholson, P. T., Jackson, C. M., & Trott, K. M. (1997). The Ulu Burun glass ingots, cylindrical vessels and Egyptian glass. Journal of Egyptian Archaeology, 83, 143–153. Nikita, K., & Henderson, J. (2006). Glass analyses from Mycenaean Thebes and Elateia: Compositional evidence for a Mycenaean glass industry. Journal of Glass Studies, 48, 71–120. Oates, D. (1987). Excavations at Tell Brak 1985–86. Iraq, 49, 175–191. Panagiotaki, M., Papazoglou-Manioudaki, L., Chatzi-Spiliopoulou, G., Andreopoulou-Mangou E., Maniatis Y., Tite, M. S., & Shortland, A. (2005). A glass workshop at the Mycenaean citadel of Tiryns in Greece. In Association Internationale pour l’Histoire du Verre (Eds.), Annales du 16 e Congrès de l’Association Internationale pour l’Histoire du Verre (pp. 14–18). Peltenburg, E. (1987). Early faience: Recent studies, origins and relations with glass, In M. Bimson, & I. C. Freestone (Eds.) Early vitreous materials (pp. 5–29). British Museum Occasional Paper, no. 56, British Museum.

20

Th. Rehren

Petrie, W. M. F. (1894). Tell el Amarna. Methuen. Petrie, W. M. F. (1926). Glass in the early ages. Journal of the Society of Glass Technology, 10, 229–234. Pulak, C. (2008). The Uluburun shipwreck and Late Bronze Age trade, 289–310. In J. Aruz, K. Benzel, & J. M. Evans (Eds.), Beyond Babylon: Art, trade, and diplomacy in the second millennium B.C. (pp. 289–310). Pusch, E. B., & Rehren, Th. (2007). Hochtemperatur-Technologie in der Ramses-Stadt—Rubinglas für den Pharao, Forschungen in der Ramsesstadt 6. Gerstenberg. Radivojevi´c, M., Rehren, Th., Pernicka, E., Sljivar, D., Brauns, M., & Boric, D. (2010). On the origins of extractive metallurgy: New evidence from Europe. Journal of Archaeological Science, 37, 2775–2787. Rehren, Th. (1997). Ramesside glass colouring crucibles. Archaeometry, 39, 355–368. Rehren, Th. (2003). Comment I on J.L. Mass, M.T. Wypyski and R.E. Stone ‘Malkata and Lisht glassmaking technologies: Towards a specific link between second millennium BC metallurgists and glassmakers.’ Archaeometry, 45, 185–190. Rehren, Th., & Pusch, E. B. (2005). Late Bronze Age Egyptian glass production at QantirPiramesses. Science, 308, 1756–1759. Rehren, Th., & Pusch, E. B. (2007). Glas für den Pharao—Glasherstellung in der Spätbronzezeit des Nahen Ostens. In G. Wagner (Ed.), Einführung in die Archäometrie (pp. 215–235). Springer. Rehren, Th., & Pusch, E. B. (2008). Crushed rock and molten salt? Some aspects of the primary glass production at Qantir/Pi-Ramesse. In C. Jackson & E. Wager (Eds.), Vitreous materials in the Late Bronze Age Aegean: A window to the east Mediterranean world (pp. 14–33). Oxbow Books. Rehren, Th., & Freestone, I. C. (2015). Ancient glass: From kaleidoscope to crystal ball. Journal of Archaeological Science, 56, 233–241. Rehren, Th., Pusch, E. B., & Herold, A. (2001). Qantir-Piramesses and the organisation of the Egyptian glass industry. In A. J. Shortland (Ed.), The social context of technological change: Egypt and the Near East, 1650–1550 BC (pp. 223–238). Oxbow Books. Schlick-Nolte, B. (1968). Die Glasgefässe im alten Ägypten. Hessling. Shortland, A. (2005). The raw materials of early glasses: The implications of new LA-ICPMS analyses. In Association Internationale pour l’Histoire du Verre (Eds.), Annales du 16 e Congrès de l’Association Internationale pour l’Histoire du Verre (pp. 1–5). Shortland, A. (2012). Lapis Lazuli from the Kiln: Glass and Glassmaking in the Late Bronze Age. Studies in Archaeological Sciences 2. Shortland, A. J., Rogers, N., & Eremin, K. (2007). Trace element discriminants between Egyptian and Mesopotamian Late Bronze Age glasses. Journal of Archaeological Science, 34, 781–789. Shortland, A. J., Kirk, S., Eremin, K., Degryse, P., & Walton, M. (2018). The analysis of Late Bronze Age glass from Nuzi and the question of the origin of glass-making. Archaeometry, 60, 764–783. Smirniou, M., & Rehren, Th. (2011). Direct evidence of primary glass production in Late Bronze Age Amarna. Archaeometry, 53, 58–80. Smirniou, M., & Rehren, Th. (2013). Shades of blue—cobalt-copper coloured blue glass from New Kingdom Egypt and the Mycenaean world: A matter of production or colourant source? Journal of Archaeological Science, 40, 4731–4743. Smirniou, M., Rehren, Th., & Gratuze, B. (2018). Lisht as a New Kingdom glassmaking site with its own chemical signature. Archaeometry, 60, 502–516. Stern, E. M. & Schlick-Nolte, B. (1994). Early glass of the Ancient World, 1600 BC to AD 50: Ernesto Wolf Collection. Gerd Hatje. Tait, H. (2012). 5000 years of glass. British Museum. Turner, W. E. S. (1954). Studies of ancient glass and glass-making processes, Part I: Crucibles and melting temperatures employed in Ancient Egypt at about 1370 BC. Journal of the Society of Glass Technology, 38, 436T-444T.

Glass in the Middle East and Western Europe at the End of the First Millennium CE, Transition from Natron to Plant Ash Soda or Forest Glasses Bernard Gratuze, Nadine Schibille, and Inès Pactat

Abstract The production of natron glass started at the beginning of the first millennium BCE and prevailed in the Mediterranean world for almost two thousand years. This production seems to cease progressively from the end of the eighth century CE onwards, with a different timing based on the region (e.g. Syria, Egypt). A recent study of Islamic glass weights and stamps, which provide a fairly continuous chronology of glass compositions from the reign of Abd al-Malik (685–705 CE) to the reign of the Fatimid caliph al-Hakim (996–1020 CE), shows that natron glass was produced in Egypt at least until the middle of the ninth century. Thus, in the Mediterranean world, a radical change is observed in glass recipes between the end of the eighth century and the tenth century CE resulting in the systematic use of soda plant ashes instead of natron. This recipe was then adopted all around the Mediterranean and became predominant in that area by the end of the twelfth century. In Western Europe, the halt of the import of natron glass at the end of the eighth century also induced a period of transition that led to the emergence of the production of new glass types. There, imported natron glass is progressively replaced by locally made glass produced with fluxes containing potash, lime or lead. Two case studies of glass workshops are used to illustrate some original solutions developed to maintain glass manufacturing in these regions.

1 Introduction This paper deals with the transition between antique natron glass and medieval glass which occurs around the Mediterranean Sea at the end of the first millennium CE (c. 800 CE). The history of glass production in the Mediterranean world, defined here as encompassing the Near East, the Middle East and Egypt, North Africa and Europe, can be roughly divided into three main phases based on the types of fluxing agents used to lower the melting point of silica (Fig. 1). The first phase starts in the middle of the third millennium BCE with the invention of glass. For the first thousand B. Gratuze (B) · N. Schibille · I. Pactat IRAMAT/CEB, National Centre for Scientific Research, Orléans, France e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_2

21

22

B. Gratuze et al.

Fig. 1 Timeline of the main glass-making recipes based on alkaline fluxes used in the Mediterranean and European world between the second half of the second millennium BCE and the first half of the second millennium CE. Recipes based on lead oxide fluxes are not illustrated

years of its history, glass production is not very well documented. It is only in the middle of the second millennium BCE that archaeological finds reveal a flourishing industry organized in two steps. The raw glass was first produced in primary workshops, localized probably in Egypt and Mesopotamia. Glass recipes involved the use of soda-rich ashes obtained from halophytic (or salt-resistant) plants mixed with crushed quartz pebbles (see Rehren, this volume). Objects were then manufactured in secondary workshops which were more widely distributed in the eastern part of the Mediterranean Sea. This model is supported by different archaeological finds such as the Uluburun shipwreck (c. 1400 BCE), with its cargo of glass ingots and finished objects (Pulak, 2008), the Amarna glass workshop (Shortland et al., 2007) or the Egyptian relief of the Annals of Thutmosis III, on the walls of the Karnak Temple. This first phase ends at the beginning of the first millennium BCE (c. 900–800 BCE). These archaeological finds reveal glassmaking in specific workshops and its trade for glass working from the middle of the second millennium BCE. From 1200 BCE onwards, for a short period of 200 or 300 years, glass was made in northern Italy at Frattesina (Veneto region, on the river Po, Angelini et al., 2004; Bellintani, 1997; Biavati & Verità, 1989). Two types of glass production have been, identified in this region. The glass most widely found is a mixed soda-potash glass. The other glass, less common and more subject to corrosion, is a potash glass. These

Glass in the Middle East and Western Europe …

23

two types of glass are characterized by high silica and low lime contents. Their production seems to cease at the beginning of the first millennium BCE. This event is the only documented evidence of glass production in the western Mediterranean region until the end of the eighth century CE. For the next two thousand years (1000/800 BCE—800 CE), the history of raw glass production around the Mediterranean Sea seems to be restricted mainly to the Levant and Egypt. During this period, a new type of fluxing agent taken from mineral deposits formed in dried-up lakes (natron, a natural sodium carbonate) starts being used (Shortland et al., 2006). It progressively replaced soda obtained from plant ashes to become the prevailing type of fluxing agent used for glass production in the coastal zones of the east Mediterranean area. However, even if it is at a lower scale, the production of glass using soda from plant ashes still continues throughout this period in some areas located in Mesopotamia and perhaps in southern Egypt (Mirti et al., 2008, 2009). The beginning of the end of this second phase is at the end of the eighth century in the Levant, with the gradual abandonment of natron and the reintroduction of soda obtained from the ashes of halophytic plants. As a result of this change, the glass workers in the Western Mediterranean basin, which have been dependent on raw glass supplies from the Eastern Mediterranean for more than a millennium, are developing new strategies. There is both an intensification of recycling practices and the emergence of new glass recipes that differ according to the location of the production areas. Although soda plant ash glass becomes the dominant glass in the west of the Mediterranean basin by the twelfth century, new types of glass, such as potash-lime glass, develop from the end of the eighth century in Western and Continental Europe. The main flux is no longer soda but wood ash, a variable mix of potash and lime with high magnesia, phosphorus and manganese concentrations. This glass becomes dominant in Continental Europe at the end of the tenth century. In parallel, recycled natron glass seems to still be used until the end of the twelfth century for specific productions: these include cobalt-blue vessels decorated with white opaque trails and spots, such as the Saint Savin bowl, and cobalt-blue-stained glass panels used in cathedral windows (Brill, 1999; Foy, 2001; Gratuze, 2020; Simon-Hiernard & Gratuze, 2011; Sterpenich & Libourel, 1997). Two other types of glass where the lime-alkaline flux is partly or totally replaced by lead oxide are also encountered from the end of the eighth century. • The first glass, with mainly lead oxide (PbO generally greater than 60%) and silica, and characterized by the absence of alkaline elements, is the lead glass sensu stricto. In the ninth century, this high lead glass seems mostly to be used for bead making and is mainly distributed in the Muslim world, in Eastern Europe (Poland, Russia) and in the Caucasus region (Bezborodov, 1975; Siemianowska et al., 2019). In the Islamic world, around the late tenth and early eleventh century, vessels were also made from silica lead glass. Earliest examples of an emerald green glass made of more than 60% lead oxide and of a deep emerald green colour, dated to the ninth–twelfth century, are found from Samarra and Nishapur (Krueger, 2014; Schibille et al., 2018; Wypyski, 2015).

24

B. Gratuze et al.

Lead-alkali glass with various proportions of lead oxide, silica, alumina, alkali oxides and lime is divided into lead-soda-lime glass, lead-potash-lime glass and lead-soda-potash-lime glass. Although in minor quantities, these types of glass are encountered throughout the Europe from the end of the eighth century (Duckworth et al., 2015; Mecking, 2013; Pactat et al., 2017; Wedepohl & Baumann, 1997). From a chemical perspective, the contents of potash and magnesia are used to distinguish natron soda glass and plant ash soda glass (Fig. 2). In natron glass, the contents of these oxides are usually between 0.3 and 1% (and always below 1.5%), while in soda plant ash glasses, the contents of these oxides are well above 1.5%. Phosphorus pentoxide contents are also usually higher in soda plant ash glasses. The major stages of the transition from natron to plant ashes and its consequences will be presented below starting with the eastern part of the Mediterranean region, dealing separately with the Levant and Egypt, and then the western part of the Mediterranean region.

Fig. 2 Ternary diagram (Na2 O-MgO + K2 O-CaO) for the main types of lime-alkaline glasses encountered in the Mediterranean and European worlds between the second half of the second millennium BCE and the first half of the second millennium CE. Glasses made with lead oxide fluxes are not illustrated

Glass in the Middle East and Western Europe …

25

2 Glass Transition in the Levant To date, the transition from natron soda glass to vegetable soda glass in the Levant is probably the most studied and best documented. Different works have pointed out the progressive decrease of soda content in natron glass (Fig. 3) produced in the Levant and to a lesser extent in those produced in Egypt starting around the sixth century (Henderson, 2002, 2013; Chap. 4; Tite et al., 2006; Freestone forthcoming). Other studies have shown that between the eighth and tenth centuries, natron soda glass was gradually supplanted by plant ash soda glass across the Islamic world (Brill, 1995; Freestone & Gorin-Rosen, 1999; Freestone et al., 2000; Gratuze & Barrandon, 1990; Henderson, 1995, 1999, 2002; Matson, 1948; Sayre, 1965). Henderson et al. (2004) analysed glass artefacts from al Raqqa originating from both consumption and production contexts and dating from the eighth to the eleventh centuries. The results demonstrate the presence of four main groups of glass: one group of natron soda glass and three groups of plant ash soda glass. They also showed that one of the plant ash soda glass groups likely resulted from mixing natron and soda plant ash glasses. The first occurrence of plant ash glass is shown to occur in the late eighth/early ninth centuries, while the near disappearance of natron glass is only observed in the eleventh century. Similar results are depicted by Phelps in two recent papers (Phelps, 2018; Phelps et al., 2016). In the first one, results obtained from the analyses of 271 glasses from 17 excavations of ten different sites located in Israel, dated from the seventh to twelfth centuries are discussed. Although only the results obtained for the 133 natron glass samples are given in the paper, the relative percentage of natron glass and plant ash glass have been estimated by period of 50 years. Results indicate that a relatively abrupt compositional change took place in the late seventh–early eighth century. This period covers the reforming reigns of al-Malik and al-Walid, which marks the end of the Byzantine glass production and the establishment of the furnaces at Bet Eli’ezer (Freestone et al., 2000). During this period, an influx of natron glass with an Egyptian composition is visible. It is also worth mentioning that the production of natron glass at Bet Eli’ezer appears to have been limited to a short period before natron glass production ceased entirely in Palestine (Phelps et al., 2016). At Israeli archaeological sites, plant ash glasses first appear in the late eighth-early ninth century, probably as Fig. 3 Decline in soda contents with time for Levantine and Egyptian glasses (data from Freestone forthcoming)

26

B. Gratuze et al.

a result of a reduced local natron glass production creating the conditions in which plant ash glass technology was re-adopted. According to their MgO/CaO and K2 O/P2 O5 ratios, Al2 O3 and SiO2 concentrations, Phelps (2018) proposed to classify these new plant ash glasses into four groups. Two of them are related to Syro-Palestinian productions, while the two others are related to Mesopotamian productions: • group P-1, identified as Tyre-type is found from the beginning of the ninth century until the twelfth century. It corresponds with the Tyre-type defined by Freestone (2002); • group P-2, related to a single sample dated from the end of the eighth century corresponds probably to a Syrian type; • group P-3 is a Nishapur colourless type which first appears in Ramla at the end of the eighth century and is present until the mid-eleventh century. It corresponds to groups Sasanian 2 (Mirti et al., 2008, 2009), Nishapur colourless (Brill, 1995; Wypyski, 2015). This glass type has since been defined as Samarra group 1 produced in or near Samarra itself (Schibille et al., 2018); • group P-4, a cobalt-blue group found from the ninth to mid-eleventh centuries corresponds to the groups Nishapur coloured type (Brill, 1995) and Sasanian 1b (Mirti et al., 2008, 2009).

3 Glass Transition in Egypt Over the past few decades, first millennium natron glass types have been increasingly refined into at least nine primary production groups with distinct compositional characteristics corresponding to specific chronological ranges and geographical origins (Picon & Vichy, 2003; Phelps et al., 2016; Schibille et al., 2016, 2017; Freestone et al., 2018). Among these nine groups, the Levantine ones are now well-defined. The situation is different for the Egyptian glasses, despite Egypt being one of the major producers of glass throughout the first millennium CE. Indeed, Egyptian groupings still rely on very limited data obtained more than 30 years ago, owing to restrictions on the export of archaeological materials from Egypt for study. As a consequence, the Egyptian glass categories and the transition from natron to plant ash glass are less documented in that region compared to the Levant. It is worth noticing, however, that the first evidences of a major transition in glass production at the end of the first millennium were observed in Egyptian artefacts (Matson, 1948; Sayre, 1965). These first investigations carried out on early Islamic glass weights and stamps (Gratuze, 1988; Gratuze & Barrandon, 1990; Sayre & Smith, 1974) revealed four successive compositional groups that have been widely used as a benchmark, even though all the analytical data have not been published until recently. However, these first results on Egyptian glass groups were no longer sufficient for a detailed socio-economic and geopolitical analysis. This lack of data about the evolution of Egyptian glass production is probably due to the difficulties in obtaining a comprehensive set of well-dated samples from

Glass in the Middle East and Western Europe …

27

Egypt, and that until the early 2000s the most common analytical methods were destructive. Thanks to the rapid development over the last two decades of laser ablation—inductively coupled plasma—mass spectrometry (LA-ICP-MS), an analytical technique considered as nearly non-destructive, and the possibility offered by this method to study large populations of objects, a major study of Islamic glass weights and stamps from two major French public collections [the Bibliothèque nationale et universitaire (BNU) in Strasbourg and the Department of Islamic Arts in the Musée du Louvre in Paris] has recently been undertaken (Schibille et al., 2019). Islamic glass weights and stamps are small glass discs with Arabic inscriptions which present an important record of the early Islamic economic system as well as the organization of the glass industry (Balog, 1976; Morton, 1985). Although the real function of these objects, made in Egypt between the Umayyad (685– 705 CE) and Mamluk (1250–1517 CE) periods, has long been debated, it is now commonly accepted that the Umayyad and Abbasid glass weights served as coin weights while their function during the Fatimid and later caliphates can be more accurately described as money weights (Bates, 1981). The use of coin weights was probably restricted to minting workshops for checking the weight of coins produced under the control of the issuing authority, while money weights were probably more widely used by merchants for exchange control or for weighing the precious metal top-up in payments for goods. Based on the inscriptions which often carry the name of the official sponsor, director of finances or caliph, many of these glass weights can be dated with relative precision and are therefore well-suited to trace the chronological developments of glass compositions. Although no glass workshops are known, it is thought that the glass weights and stamps originated ultimately from Egypt, since they bear the names of Egyptian officials and most of the samples with known provenance were in fact retrieved from Fustat (ancient Cairo) (Ollivier, 2019). Although there is no evidence as to how or by whom these glass weights, vessel stamps and large commodity weights were produced, it is commonly accepted that their manufacture was strictly controlled by the official mint and officials of the treasury (Noujaim-Le Garrec, 2004). To follow the evolution of Egyptian glass chemical compositions, with a high temporal resolution, we have analysed 171 glass weights and stamps ranging from the reign of the Umayyad caliph Abd al-Malik (685–705 CE) to the reign of the Fatimid caliph al-Hakim (996–1020 CE). Although a significant hiatus in the production of glass weights between the last quarter of the ninth and the middle of the tenth century occurred, this period covers several radical changes in Islamic glassmaking, and the results allow the construction of a complete and precise temporal model for the transition from natron to plant ash glass in Egypt. Based on the strontium to calcium ratios in relation to magnesia, alumina, titanium, zirconium, lanthanum and thorium contents and using principal component analysis (PCA, calculated with Na, Mg, Al, Ca, Ti and Zr for natron glasses and adding P, K and Cr for plant glasses), we have identified eight different base glasses: four Egyptian natron glass types, dating to the eighth and ninth centuries (referred as Egypt 1A, 1B, 1C and Egypt 2) and three plant ash glass groups (one from the Levant—Lev—and two from Egypt—E1 and E2) that date to the second half of the tenth and the early eleventh centuries. A fourth

28

B. Gratuze et al.

plant ash glass group ascribed to a Mesopotamian provenance based on its high MgO (>3.5%) and low P2 O5 (3700–2800 BCE) and the development of more complex glazing and faience technologies during the subsequent Kot Diji Phase (2800–2600 BCE). The main focus will be on the diverse range of glazed steatite and faience production during the Harappa Phase (2600–1900 BCE). During the Late Harappa Phase (1900–1700 BCE) at the site, there is evidence for continued production of both glazed steatite and faience. The chemical composition and technology of archaeological examples are compared with experimental replicas to better understand the possible stages of production and recipes used to make both glazes and faience. The implications of Harappan glazing and faience for later developments of glass are discussed.

1 Introduction The study of glazing and vitreous materials of the Indus Tradition has been the focus of considerable research ever since the initial excavations of the ancient cities of Harappa and Mohenjo-daro in the 1920s and 30s (Fig. 1). The main focus of these early studies was to define and characterize the technologies found in the Indus and to compare them with discoveries from other early cultures in Mesopotamia and Egypt (Marshall, 1931; Ullah, 1931; Mackay, 1931, 1938a, 1938b; Beck, 1940). Although the early research on glazed steatite and faience was limited by the scientific techniques available at the time, the analyses did reveal some important aspects of glazing technology that is confirmed by more detailed analyses using the most recent techniques. Furthermore, the intimate understanding of materials and rigorous recording by E.J.H. Mackay, H.C. Beck, Sana Ullah and others has made it possible to compare their interpretations with more recent analyses. These early scholars J. M. Kenoyer (B) University of Wisconsin-Madison, Madison, USA e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_3

39

40

J. M. Kenoyer

Fig. 1 Major Sites of the Indus Tradition, Integration Era

provided a solid foundation for later studies, and this study attempts to go even further to develop a more precise understanding of glazed steatite and faience technology based on the analysis of samples from a broad chronological span of time from the site of Harappa. Major advances in our understanding of the production of steatite and faience ornaments have been made by numerous studies that are discussed in more detail below and through the work on artifacts recovered through new excavations at the site of Harappa itself. One of the major goals of the new excavations at Harappa, Pakistan, begun by the late Dr George F. Dales and me in 1986 was to develop a better understanding of the process of urban development at this major urban centre (Dales & Kenoyer, 1989). A specific focus was to learn more about technologies that played an important role as symbols of wealth and identity in the process of urban development (Kenoyer, 1989, 1992a, 1994a, 2000). The long-term goal in these studies has been to more precisely define the local and regional trajectories of each of the major technologies used at Harappa and other Indus sites to better articulate the contributions of specific regions to different technological processes that were characteristic of the Indus Tradition. Studies carried out by other members of the Harappa team over the years have added greatly to the knowledge of specific craft technologies, including pottery (Dales, 1991; Wright, 1991, 1993; Kenoyer, 1994a; Clark, 2007; Clark & Kenoyer, 2017), pyrotechnology, kilns, faience and steatite firing (Miller, 1996, 1997a, 1997b,

Glazed Steatite and Faience Technology at Harappa, Pakistan …

41

1999, 2008a, 2008b; Miller et al., 1996; Kenoyer, 2005c; Miller & Kenoyer, 2018), copper metallurgy (Miller, 1994; Kenoyer & Miller, 1999; Hoffman & Miller, 2009; Hoffman, 2019), lithics (Davis et al., 2016), and especially the identification of rock and metal raw materials and their trade (Kenoyer, 1997; Law et al., 2013; Law, 2002, 2006, 2011a, 2013a, 2013b, 2014). Additional studies of fibres are relevant for the study of beads and ornaments (Good et al., 2009, 2011; Kenoyer, 2017b; Wright et al., 2012), but will not be discussed in this paper, though they played an equally important role in the process of urban development. Subsistence studies are also closely linked to specialized crafts in terms of provisioning of urban centres (Belcher, 1998, 2000; Meadow & Patel, 2003; Miller, 2003; Weber & Belcher, 2003; Chase et al., 2014, 2016) and the economic foundation supporting specialized crafts, but they cannot be addressed in this paper. However, eventually it will be necessary to articulate the complex interplay of these different technologies in social, economic and ideological patterns that are distinctive of the Indus Tradition during its long and complex development (Kenoyer & Miller, 2007). The site of Harappa, Pakistan (Fig. 2), has been the focus of major excavations since the 1930s (Mughal, 1968; Possehl, 1991; Vats, 1940; Wheeler, 1947) and more recently by the Harappa Archaeological Research Project (HARP) (Dales and Kenoyer, 1993; Meadow & Kenoyer, 2001, 2005, 2008; Kenoyer & Meadow, 2016).

Fig. 2 Harappa Site map, 1986–2001, 2007

42

J. M. Kenoyer

General studies of the early discoveries and more detailed analyses of samples recovered from the HARP excavations provide a robust dataset for describing and interpreting the changing patterns of steatite glazing and faience production at a major Indus urban centre. Glazed steatite and faience production share important aspects of technology that make it useful to study both technologies together, yet each technology involves important differences that relate to the production sequences and the nature of the finished objects. Similar types of analyses can be carried out on both sets of artifacts, and this allows for important comparisons in production processes, waste materials and finished objects. In the following paper, the research conducted on both steatite and faience at the site of Harappa is presented in order to establish a broad foundation on which future studies can be undertaken. Due to the fact that this research is ongoing, some of the data presented are preliminary in nature and could change as our analyses continue. Regardless, the Harappa data can be used as a comparative dataset against which data from other urban and rural centres can be compared so that more detailed interpretations of technological development and variation can be proposed.

1.1 Analysis Methods The analysis of steatite and faience objects has been done using multiple scientific quantitative methods as well as basic qualitative observation. All steatite and faience objects from the HARP excavations have been tabulated and recorded, with most being photographed or scanned using a flatbed scanner. Measurements of select objects and groups of objects have been made using digital callipers to document morphological features such as maximum length, width, thickness and hole diameter (Kenoyer, 2017a). Select samples have been documented using a digital microscope (Dinolite™) that allows for more precise measurements (Kenoyer, 2017c). Samples selected for more detailed scientific analysis were exported to the USA with permission from the Department of Archaeology and Museums, Government of Pakistan. The types of analyses conducted on the faience samples included preparation of polished sections under the guidance of Dr Pamela Vandiver during an internship with her at the Smithsonian in 1993. Additional polished samples were prepared at the University of Wisconsin-Madison, Department of Anthropology and the Department of Geosciences. These polished cross sections of the objects provide information on the structure of the faience paste that was documented using a petrographic point-counting protocol developed in consultation with Stoltman (1991). Using SEM backscatter imaging, different areas of the object cross section were photographed at 700× for determining the percentage of unmelted quartz, glass, voids and other inclusions. Using a grid positioned at different angles, these images were analysed and between 350 and 500 points were tabulated to generate data that could be plotted in a ternary plot (see Fig. 32). This is the first time such data have been collected for Indus faience objects, and it provides an important source of data for comparing the preparation of different types of objects.

Glazed Steatite and Faience Technology at Harappa, Pakistan …

43

The composition of glass and the identification of other components including quartz, feldspar, zircon, titanium, etc., were analysed using the scanning electron microscopes (SEM) with backscatter imaging and with energy dispersive spectrometry (EDS) (Kenoyer, 2017c). The Department of Geosciences has had several different SEM–EDS units, and the most recent instrument is a Hitachi S3400 VPSEM. The SEM in the Department of Anthropology is a Hitachi 3870S that was previously located in the Department of Animal Sciences. Although EDS does not provide extremely accurate quantification, it does allow for the overall percentages of different minerals that are useful for determining the identity of inclusions as well as the general composition of the glass. Selected samples were sent to the Elemental Analysis Facility at the Field Museum, Chicago, for compositional analysis of the glass using Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LAICP-MS) by Dr Laure Dussubieux (Dussubieux et al., 2009). Additionally, some samples were analysed with X-ray diffraction (XRD) by Randall Law using a Rigaku Rapid II diffractometer. The complete data from all of these different analyses will be presented in a future publication, and some general observations are included in specific sections below.

1.2 Experimental Replication In conjunction with the analysis of archaeological samples, I have been involved in a long-term experimental program of replication of both fired and glazed steatite, as well as many varieties of faience. These experiments have been undertaken individually and in collaboration with many colleagues and students in the USA, Pakistan and India. In the experimental replications, I used commercial quartz sand and other components as well as quartz rocks and quartz sand collected from various rivers and geological locations in Pakistan. The replication of Indus faience undertaken in conjunction with the conference-cum-workshop on ‘History, Science and Technology of Ancient Indian Glass’ at IIT Gandhinagar (Kanungo & Trivedi, 2019: 357) was undertaken with many participants using local sand from the Sabarmati River. The fluxes that I have used include commercial sodium carbonate and sodium bicarbonate, as well as traditionally produced plant ash (sajji or khar) obtained in different parts of Pakistan, India and Afghanistan. Colourants such as copper oxide or iron oxide have been prepared from pure metals, as well as from natural minerals, such as malachite, cuprite, ochre and crushed and calcined bone. Different types of binders such as clay, honey, mustard oil and acacia gum were also used in order to determine possible variations in processing. In earlier experiments that I conducted, I incorrectly stated that I had used ‘shisham gum’ Dalbergia latifolia (or more correctly Dalbergia sissoo) (Kenoyer, 1994b: 38) when in fact I was using acacia or ‘kikar gum’. I have not yet experimented with ‘shisham gum’ though according to early Chitrasutra texts the ‘juice’ of this tree is included as one of the ingredients in the preparation of wall plaster for painting (Nardi, 2006: 121). Firing was carried out in a muffle furnace with precise temperature control, as well as in a reconstructed wood

44

J. M. Kenoyer

firing process based on the faience bead and tablet workshop discovered at the site of Harappa (Kenoyer, 2005c). In all of these experiments, the documentation of the processes and quantities as well as the firing conditions and temperatures has been recorded to gain insights into the results achieved using different materials. Preliminary observations from these experiments are included in the discussion below, but the details will be presented in a future publication.

2 Indus Tradition Chronology and Terminology The term Indus Tradition is used for the overall phenomenon of cultural development that is often referred to as the Indus Civilization or Harappan Civilization in other publications (see Meadow & Kenoyer, 2005 for full discussion). The chronological framework is based on the radiocarbon dates from Harappa as well as information from other sites, but it must be emphasized that each individual site in the Indus region needs to be dated and the precise location and context for the carbon should be published along with any published dates. Carbon that is collected from secondary contexts should not be considered as a valid date without detailed discussion of the post-depositional processes that were involved in the final location of the carbon. The development of steatite bead making begins during the Early Food Producing Era, Mehrgarh Phase (+7000–5500 BCE). During the Regionalization Era, Early Harappa Phase (5500–2800 BCE), we see the first appearance of blue green as well as white glazed steatite and faience. In the Harappa Phase (2600–1900 BCE), there are major developments in both steatite firing and glazing as well as the forming and colouring of a wide range of faience objects. There is much less primary context data on steatite and faience during the Late Harappa Phase (1900–1000 BCE), but both technologies continued, with some changes in styles of objects being produced (Table 1).

2.1 Fired and Glazed Steatite Steatite, or talc [hydrous magnesium silicate, Mg3 Si4 O10 (OH)2 ], is a metamorphic rock that has a Mohs hardness of 1–2.5 (Law, 2011b: 178) (Fig. 3e–h, j). The presence of talc makes it feel slippery or smooth, and it is commonly referred to as soapstone. It is found in many different colours, including white, grey, green, black and reddishbrown (Law, 2011b: 182, Fig. 7.4), but the Indus craftspeople usually fired and glazed the steatite objects to harden them and change the colour to white or blue green when coloured with copper (Miller, 2008a). When fired at 900–1000 °C, the talc is transformed into enstatite (magnesium silicate, MgSiO3 ), which has a Mohs hardness of 5–6. If steatite is fired over 1100 °C, the amorphous silica from the talc recrystallizes as cristobalite (a polymorph of quartz, SiO2 ), which has a Mohs hardness of 6–7 (Law, 2011b: 256–257).

Glazed Steatite and Faience Technology at Harappa, Pakistan … Table 1 Indus Tradition chronology: Harappa and early Mehrgarh/Nausharo

45

Localization Era Late Harappan phase c. 1900–1300 BCE Harappa: periods 4 and 5

1900–1700 BCE

Integration Era Harappan phase

2600–1900 BCE

Harappa: period 3C, Final

2200–1900 BCE

Harappa: period 3B, Middle

2450–2200 BCE

Harappa: period 3A, Initial

2600–2450 BCE

Regionalization Era Early Harappan (several phases)

c. 5500–2600 BCE

Harappa: period 2, Kot Diji phase

2800–2600 BCE

Harappa: period 1, A & B, Ravi/Hakra Phase

> 3700–2800 BCE

Mehrgarh, period III

4800–3500 BCE

Mehrgarh, period II

5500–4800 BCE

Early Food Producing Era Neolithic—Mehrgarh phase

c. 7000–5500 BCE

Mehrgarh, period 1, non-ceramic

7000–5500 BCE

Through a systematic analysis of archaeological samples from excavations at Harappa and other sites, compared with geological samples, Law was able to determine that two major varieties of steatite deposits are found around the greater Indus region, ultramafic igneous rock and dolomitic sedimentary rock (Law, 2011b: 190, Fig. 7.9). Dolomitic sources in northern Pakistan (Hazara, Parachinar, Landi Kotal) and northern India (Jammu) are the major sources for raw and fired steatite from Harappa. Less than 5% come from sources outside that region and possibly some come from the dolomitic sources from the northern Aravalli range in Rajasthan, India. However, the Rajasthan dolomitic steatite does not fire white, and Law argues that the Indus craftspeople appear to have selected steatite primarily from those dolomitic sources that can easily be fired to a white colour (Law, 2011b: 257–259) (Law personal communication 2019). The earliest evidence of unfired steatite used to make short cylindrical beads comes from Period I at the site of Mehrgarh, Baluchistan, and dates to around 7000 BCE though it is possible that this date may be revised in the future (Richard Meadow personal communication). During Period I and IIa, craftspeople at Mehrgarh were beginning to experiment with the heating of steatite to harden and whiten it (Barthélemy De Saizieu & Bouquillon, 1994: 51; Miller, 2008a; Law, 2011b). During the Regionalization Era, the practice of firing and glazing steatite beads becomes more common at Mehrgarh (Periods IIB c. 5500–5000 BCE and Period III 5000– 3500 BCE), as well as at the nearby site of Nausharo (Period 1, 3000–2500 BCE)

46

J. M. Kenoyer

Fig. 3 Harappa, Period 3, Faience and Steatite Workshop Debris, H2000 (a chert flake and blade; b1 and 3. moulded faience tablets, b2 and 4. bicolour moulded faience tablets; c blackened over-fired faience beads; d blue-green and yellow glazed faience beads; e white steatite lump; f white glazed steatite bead; g black steatite lump; h various colours of steatite fragments; i faience gaming piece and leech-shaped bicolour bead; j steatite tablet blanks, some partly inscribed; k steatite faience tablet mould that matches the moulded faience tablet 3b3)

(Barthélemy De Saizieu & Bouquillon, 1994: 52; Miller, 2008a; Law, 2011b) (dates based on Jarrige, 2008). At the site of Harappa, unfired steatite beads along with glazed steatite beads and glazed faience beads are documented beginning in Period I, the Ravi Phase occupation (c. >3800–3300 BCE) and continuing on through Period 2, the Kot Diji Phase (2800–2600 BCE) (Kenoyer & Meadow, 2000). Although earlier studies of the Ravi Phase steatite suggested that they might have originated in sources in Jammu (330 km northeast) and Rajasthan (500 km south-east) (Law, 2011b: 255), new analyses indicate that these samples are in fact sourced to northern Pakistan (Parachinar) and northern India (Uttaranchal) (Law personal communication 2019). By the Kot Diji Phase, there is good evidence for steatite from Sherwan area of Hazara (400 km north-west) (Law, 2011b: 255). In the Kot Diji Phase, carved and glazed button seals were produced for the first time using a glaze that was either white or blue green in colour. Most of the glaze on these seals has eroded, and it is rare to find traces of the glaze remaining, but we can assume that all fired steatite seals were originally glazed. Since the seals were glazed, the glassy surface may have made the design less distinct, and it is possible that they may have been used more for decorative purposes

Glazed Steatite and Faience Technology at Harappa, Pakistan …

47

than for stamping. However, one example of a seal impression from Lewan with a geometric design (Allchin et al., 1986; Shah & Parpola, 1991, Lwn-1) indicates that some of the seals were used to stamp clay sealings. Steatite beads continued to be made with white and sometimes blue-green glazes during the Kot Diji Phase. Fired and glazed steatite technology was widespread during the Regionalization Era, and steatite beads are common at most Early Harappan sites. Geometric button seals made of fired steatite or seal impressions are reported from sites throughout the northern and central Indus River Valley regions, as well as in the Ghaggar-HakraNara River Valley to the east. None of the original excavators have reported that these seals were originally glazed, but examination of a few button seals from Rehman Dheri as well as Kunal shows that they were originally glazed, and I would assume that all the seals from this time period were also glazed. These sites include Tarakai Qila (Allchin & Knox, 1981; Thomas, 1983; Shah & Parpola, 1991, Trq-1 to 4), Lewan (Allchin et al., 1986; Shah & Parpola, 1991, Lwn-1), Gumla (Dani, 1970–71; Shah & Parpola, 1991, G-11), Rehman Dheri (Durrani et al., 1995a, 1995b), Baror (Sant et al., 2005), Tarkhanewala Dera (Joshi & Parpola, 1987, Tkwd-1; Trivedi, 2009) and Kunal (Khatri & Acharya, 1997, 2005). So far, no sites in Gujarat dating to the Regionalization Era have reported glazed steatite beads or seals, but these may be discovered in future excavations. The wide distribution of Early Harappan glazed seals carved with geometric designs has been studied by Konosukawa (2013: 37ff), but more detailed studies are needed to determine if these were all made in the same workshop or by craftspeople trained in the same manufacturing tradition or workshop style (Jamison, 2017; Kenoyer, 1997; Ludvik, 2018). Sourcing of the steatite used to make the Early Harappan seals is also something that still remains to be done to determine the area from which the steatite raw material was procured. During the Harappa Phase, glazed steatite is most commonly seen in the form of carved seals that were fired and glazed with a thin white glassy surface. The use of blue-green glazes did however continue on button seals and ornaments. Most fired steatite beads were made with a white glaze, and this is seen on both larger ‘disc’ beads (defined as ‘very very short cylindrical’) as well as the famous micro beads (short cylindrical) that have been found at most Indus sites. The microbeads are often found stuck together because they were fused together in the process of firing, which will be discussed below in more detail. Although the production of glazed seals was discontinued during and/or after the Late Harappa Phase (1900–1000 BCE), glazed steatite beads and other ornaments continued to be produced using some of the same shapes and styles of the earlier period as well as new shapes and surface treatments. The glaze colours during the Late Harappa Phase range from white and blue green to dark greenish black.

2.2 Faience—Powdered Quartz Body with Glaze The archaeological use of the term ‘faience’ refers to any object made from sintered or partially fused powdered quartz that has a vitreous or glassy external surface

48

J. M. Kenoyer

(Fig. 3b–d, i). Quartz can be obtained from pebbles of white rock quartz, which has relatively few inclusions or from quartz sand that may have various other inclusions (Tite & Bimson, 1986: 69). Analysis of one faience bead from the Ravi Phase at Harappa shows the presence of quartz, as well as potassium feldspar that could come from a specific sand source or from crushed rock. Analysis of many faience bangles from the Harappa Phase indicates that the main source of raw material was probably sand from the Ravi River. Comparisons of modern Ravi River sand with the ancient faience as well as some actual quartz sand recovered from the Harappa excavations show the presence of a wide range of identical mineral inclusions, such as zircon, monzanite, potassium feldspar and even some traces of gold and silver. Unlike glazed steatite, which is made using a solid material that is carved and shaped, the finely ground quartz powder must be mixed with a binder and flux and then moistened with water to make a paste to shape or mould specific objects. Quartz can be ground using traditional grinding stones or mortar to a fine powder (30–50 µm), but compared to clay, it is generally relatively non-plastic and is difficult to shape without the aid of a binder. Binders that can be added include a small portion of clay or a plant gum that will burn out during the firing (e.g. acacia gum or kikar gum, Acacia arabica). Additionally, a flux is needed to help lower the melting temperature of quartz (1726 °C) (Ringdalen & Tangstad, 2016) so that the quartz grains will fuse together with a layer of glass. Based on the analysis of faience samples from Harappa, the flux used in the Indus appears to have been a plant ash, sajji or khar, made from one of many desert plants (see below for detailed discussion). It is possible that some faience was made with natron (sodium carbonate and sodium bicarbonate) or reh mitti, but so far, this is not documented in the faience samples studied from Harappa. Finally, some form of colourant is added to the mixture, and depending on the percentage and chemical composition of the colourant, it can also serve as a flux to contribute to reducing the overall melting temperature of silica. In most of the archaeological literature, three major methods of fusing and glazing the quartz body have been identified for the production of faience objects, namely application, cementation and efflorescence (Vandiver, 1982, 1983; Tite & Bimson, 1986). In all three methods, a binder and flux are needed during the forming, firing and glazing stages, and colourants are added either in the applied glaze, in the ash used for cementation or in the overall mixture in the case of efflorescence. A fourth technique, called compact faience, has been documented for the Harappa Phase in the Indus Tradition, and a fifth technique can be identified that combines efflorescence and application. The early analyses of faience objects by Ullah and Mackay were not focused on differentiating these techniques, and they often used overlapping terms or described Indus faience glazing techniques in ways that cannot be confirmed today. A detailed reading of their reports shows that they noted that some objects had frit mixed with the paste, which would indicate a process of efflorescence, and in other instances, the glaze was applied (Mackay, 1931: 365–367). Ullah discounted the possibility of the use of clay as a binder because he found no chemical evidence for clay in the samples he studied. He also discounted the use of gum as a binder because ‘… gum or any organic matter, would be consumed in the course of firing long before the flux

Glazed Steatite and Faience Technology at Harappa, Pakistan …

49

underwent fusion’ (Ullah, 1931: 686–687). This is actually not true as experiments that I have conducted show that acacia gum does work quite well as a binder and that the shape of the object does hold together through the firing. Beck noted that some faience beads from Harappa were made with relatively coarse quartz particles and others were much finer (Beck, 1940: 407). Mackay went even further and referred to ‘vitreous paste’, a form of faience that is glassy and has lots of small bubbles (Mackay, 1931: 574–576). In my experimental replications, the use of acacia gum tends to result in faience with lots of small bubbles, but further studies are need to determine if the bubbles result from gases released by the organic acacia gum or if they are the product of the melting quartz and flux. While the earlier observations contributed to our understanding of the main techniques, additional analyses and experimental replication provide a more accurate picture of the different techniques that will be presented below.

2.3 Major Faience Production Techniques (1).

(2).

The use of application for glazing steatite is well documented from all periods at Harappa, but this technique does not appear to have been used to glaze faience except in combination with efflorescence (see below). The application method of glazing requires the preparation of frit by melting silica with higher degrees of flux along with some form of colourant. According to Ullah, ‘The prevailing colours of the faience glaze are bluish-green or greenish-blue, although indigo blue, apple green, maroon, black, and colourless examples have also been found. The blue shades owe their colour to copper oxide, while the green contains iron oxide, in addition. The black or dark maroon glaze contains an excess of manganese oxide’ (Ullah, 1931: 687). All of the main colourants except cobalt were identified by the earlier analyses, but cobalt was proposed for the darker blue colour found on some ornaments and vessels (Mackay, 1931: 572) although he did not confirm this with analysis. The frit for making the glaze would need to be reground and applied as a slip to the exterior of the dry quartz body that can be either raw or bisque fired prior to glazing (Tite et al., 2008). Ullah argues that faience objects were probably fired twice, first to fuse it and then with a glaze applied to the exterior (Ullah, 1931: 686–687). The cementation process was probably used for glazing steatite beads at Harappa and other sites, but so far there is no conclusive evidence that it was used for the glazing of faience beads. In the cementation process, the object is shaped and dried prior to packing in a firing container filled with an ashy fluxing agent. The ashy mixture includes the colourant and fluxes but also contains non-fusible material such as lime that will keep the ashy mixture from sticking to the object (Tite et al., 2008; Wulff et al., 1968). The earliest glazed steatite beads from Harappa during the Ravi Phase are sometimes fused

50

(3).

(4).

(5).

J. M. Kenoyer

together with a glaze that could be the result of either the application or cementation process. In addition, the discovery of glazed steatite beads fused with frothy slag at Harappa and also at the site of Lakhan-jo-daro suggests that the cementation process was probably used for glazing large numbers of beads in terracotta firing containers or saggars during the Harappa Phase. Efflorescence appears to have been the major method for glazing faience used at Harappa and most other Indus sites. The third method of efflorescence involves making a quartz paste that includes the flux as well as the colourant and some material to make the paste plastic. During the drying process, the evaporation of water draws most of the flux to the surface where it is highly concentrated. When fired, the exterior layer is melted to form the glaze, while the interior that now has less flux is fused but not melted and therefore maintains the intended form. This last process is the most common technique used in the Indus, while the other techniques were more common in other regions (Tite et al., 2008). Faience objects, such as seals, large beads or figurines, that have coarse quartz particles but with traces of coloured glaze extending from the interior to the exterior would be examples of this type of production. A fourth category of faience production, called compact faience, can be defined as a variation of the efflorescence process (McCarthy & Vandiver, 1990; Kenoyer, 1994b). As noted above, this technique was first identified as ‘vitreous paste’ by Mackay (1931: 574–576). This technique which is now referred to as ‘compact faience’ is documented at Harappa for one faience bangle from the Kot Diji Phase (2800–2600 BCE) as well as most bangles, beads, vessels and figurines dating to the Harappa Phase (2600–1900 BCE). Further studies need to be conducted to determine how widespread the production was and how long it lasted. In this process, a frit of partially sintered quartz with colourant was made and then reground to a very fine powder that has quartz particles that are around 30–50 µm or less in size. This extremely fine powder has the feel of talcum powder with no gritty texture (Kenoyer, 1994b). It was then formed into objects using a little additional flux and probably a binder to make the paste more plastic. After drying and a second phase of efflorescence, the items were fired in specially made saggars to protect them from discolouration from carbon, resulting in a smooth even glazed surface and a very strong compact body. This fifth technique involves the combination of efflorescence and/or compact faience techniques and the application of decorative colours on the surface of a previously formed but unfired vessel or bead. This technique is documented at Harappa and other large sites. This process involves the manufacture of an object using either process 3 or 4 and then the application of different colours of glazes in the form of applied decorative lines or patterns on the surface of the object. This technique was used during the Harappa Phase and the Late Harappa Phase to make decorated vessels and beads (see below for more discussion). This application is not the same as an applied glaze but more in the form of a paint or appliqué line of cobalt-coloured faience. One container for holding crushed galena (surma) from Harappa was decorated with a fish

Glazed Steatite and Faience Technology at Harappa, Pakistan …

51

scale design using yellowish white colour on a reddish body (Fig. 30g). The colours have not been analysed yet as this object was not available for more detailed study. The earliest faience beads at Harappa are found during the Ravi Phase occupation, dating to around >3700–2800 BCE. The only form found so far is long cylindrical, and based on the analysis of one fragment, they were coloured with copper and glazed with the efflorescence technique (Fig. 8f). Most of the faience beads studied from the sites of Mehrgarh and Nausharo appear to have been made of faience made from quartz, but a few beads showed evidence of steatite fragments that were mixed in with the quartz body (Barthélemy De Saizieu & Bouquillon, 2000: 97–99). The Ravi Phase also has evidence for the production of faience bangles, and one fragment was recovered from debris in a hearth that is associated with the final levels of the Ravi Phase based on the pottery and stratigraphy. This bangle has not yet been analysed, but it has an irregular blue-green glazed surface (Fig. 16a). Additional incised faience bangles and broken faience vessel were recovered from the Kot Diji occupation levels (Fig. 16d–g). During the Harappa Phase (2600–1900 BCE), faience was used to produce a wide range of objects, including beads, pendants, geometric seals, small figurines, larger composite figurines and monochrome as well as polychrome vessels (Mackay, 1931, 1938a; Kenoyer, 1994b, 2005c; Barthélemy De Saizieu, 2003; Miller, 2008a). The techniques used to produce these diverse objects demonstrate the immense virtuosity of Indus craftspeople and their ability to meet the challenges of modelling and glazing faience. The Late Harappan Period (1900–1000 BCE) sees the continuity of faience production to make beads and possibly bangles, but other types of objects were no longer produced.

2.4 Steatite Paste and Steatite Faience Two terms that are commonly found in the literature on Indus beads and ornaments are ‘steatite paste’ or simply ‘paste’ beads (1938a; Mackay, 1931) and ‘steatite faience’ (Barthélemy De Saizieu & Bouquillon, 2000), which is also sometimes called ‘talcose faience’ (Miller, 2008a). Both terms have resulted in considerable confusion and require some clarification. According to Mackay, ‘Steatite easily takes first place in importance; it is no exaggeration to say that three-quarters of the beads found at Mohenjo-daro are either cut out of steatite or moulded from a paste made from the ground-up stone. The latter is very common and unless optical means are employed is frequently very difficult to distinguish from the natural stone. I can find no evidence that the beads made of this steatite paste were shaped in a mould. We have not as yet found a single mould for bead-making, nor does the appearance of any of the beads suggest that they were made in this way. It seems that the steatite paste was made up into blocks, from which the beads were carved in the same way as they would have been from natural stone. I do not know if great pressure would make powdered

52

J. M. Kenoyer

steatite adhere together, but under a glass of moderate power I can see no binding substance, and if one was employed it is quite transparent’. (Mackay, 1938b: 495). The term ‘steatite faience’ was used by Ullah to describe two objects, one object was the base of a vessel, and the other a fragment of a seated human figurine neither of which were illustrated by photographs or drawings. The body of the material of this peculiar class of ware is powdered steatite, which is bound together with a flux. Originally objects of this material must have been glazed, but they have undergone much decomposition, and not a vestige of the glaze has been left on them. The material is soft and cream white; but the presence of a little copper oxide in it, revealed by chemical analysis (Table 1, p. 689), leaves no doubt that they were originally coloured blue or green like ordinary faience. In fact, it is highly probable that their whole technique was identical with that of the faience described above, and that steatite was introduced to replace quartz, in order perhaps to get over the difficulty experienced in crushing this very hard mineral (Ullah, 1931: 687).

These objects had around 57% silica, but 27% magnesia, and both objects had traces of copper oxide (Ullah, 1931, Table 1, p. 689), which would suggest that they were made with the efflorescence technique. Other details of the composition and structure were not published, so we do not know if the steatite was finely ground or coarse particles mixed with the silica. We can assume that these two objects date to the Indus occupation and not from a later period since Mackay or Marshall would have certainly made a point of this. However, since there are only two such objects reported to date, further examples are needed before proposing that the Indus had a well-developed ‘steatite faience’ technology. In fact, I would argue that all of the beads referred to as made of ‘steatite paste’ or ‘paste’ were in fact fired and glazed steatite that was weathered and became friable due to post-depositional processes. Examination of fired and glazed steatite beads from the excavations at Harappa has provided many examples of necklaces that include hard, well-preserved beads alongside identical ones that are very fragile and powdery. The weathering of fired steatite and its decomposition due to variable soil chemistry is something that still needs to be studied more precisely. Earlier archaeologists were not successful in replicating steatite ‘paste’ beads because of the fact that steatite does not melt and fuse like silica. The proposition that paste beads were made using steatite powder gained some notoriety through the untested hypothesis of K.T. M. Hegde regarding the manufacture of Indus microbeads. Hegde argued that the beads could have been made of steatite paste by using an extrusion process to produce hollow tubes that were then sliced to produce the tiny beads (Hegde et al., 1982). Unfortunately, Hegde never published any photos of this process, and there is no record that he actually ever made such beads. Another study of steatite and faience beads from Mehrgarh/Nausharo by Vidale discussed the presence of steatite ‘paste’ beads (Vidale, 1989a), but here too the analysis was inconclusive. Vidale has argued that it is difficult or impossible to differentiate weathered and decomposed fired massive steatite from steatite that was fired as a paste (Vidale & Miller, 2000: 64–66). In none of these earlier studies has there been any claim of successful replication of Indus style steatite paste beads, let alone the very small microbeads. After considerable analyses of steatite beads (Law, 2011b) and through

Glazed Steatite and Faience Technology at Harappa, Pakistan …

53

experimental attempts at creating steatite paste beads (Law, 2018), Law has clearly shown that steatite short cylindrical beads and microbeads could not have been made from steatite paste. Based on the evidence found to date, Indus steatite beads of all sizes appear to have been made from massive steatite that was sawn, drilled, shaped, fired and glazed. Evidence of the entire manufacturing process was first discussed and published by Mackay (1938a). Workshops with evidence for steatite short cylindrical and microbead manufacturing have also been reported from Harappa (Kenoyer, 2005c), Lakhan-jo-daro (Mallah & Rajput Shafiq, 2016), Mehrgarh/Nausharo and Mohenjo-daro (Vidale, 1989b, 1995). The widespread discussion about steatite paste beads at Mohenjo-daro and other sites (Mackay, 1931: 576–577, 1937: 10–12, 1938b: 350, 495, 1943: 405; Ullah, 1931; Vidale, 1989a: 293–294) may have provided some basis for the use of the term ‘steatite faience’ in describing a small sample of beads found at the sites of Mehrgarh and Nausharo, Baluchistan (Barthélemy De Saizieu & Bouquillon, 2000: 99–101). Five beads from Mehrgarh dating to the Pre-Indus Chalcolithic levels (Mehrgarh Period VII, Early Harappa Phase, Regionalization Era) and one from the Indus Period at Nausharo show evidence that some steatite fragments were mixed into the quartz powder to create a form of ‘steatite faience’ (Bouquillon & Barthélémy De Saizieu, 1994, 1995, Bouquillon et al., 1995; Barthélemy De Saizieu & Bouquillon, 2000: 97–99, Plate II, 6, 7) or ‘talcose faience’ (Miller, 2008a). Based on the description in their published accounts, the steatite or talc does not appear to be uniformly distributed in the faience paste, but appears as scattered particles imbedded in a vitreous quartz matrix. ‘They are made up of fragments of steatite embedded in a complex matrix of Si, Mg and Cu. Each fragment of steatite is surrounded by a ring of reaction comparable to the felty zone observed in the glazes of the steatite beads …’ (Bouquillon & Barthélémy De Saizieu, 1995: 531). Normally if you are preparing a paste for making quartz faience, the quartz is ground to a fine powder in order to make it easier to shape. This uniform grinding would also apply to any steatite powder that would be added to the quartz paste. However, the sizes of the steatite particles seen in the published SEM photos of two of the bead samples are very irregular and do not appear to represent finely ground steatite powder that would have been added intentionally to a faience paste. The steatite particles in the bead from Mehrgarh are approximately 100 µm to more than 400 µm (Barthélemy De Saizieu & Bouquillon, 2000, Pl III, 7), and the steatite particle identified in the bead from the Indus period at Nausharo is around 100 µm with possibly some larger particles that were not identified in the published image (Barthélemy De Saizieu & Bouquillon, 2000, Pl III, 8). I would suggest that it is possible that the steatite particles in these beads were the result of contamination, possibly from nearby steatite bead working area, and not an intentional addition to create ‘steatite faience’. Regardless of the intentionality of steatite particles in the ancient samples, I began some experimental replication of faience using different amounts of added ground steatite or talc powder. The control sample was made with quartz, flux (sajji) and a colourant (copper oxide) (Fig. 4a). In one sample, a small amount (9%) of very finely ground steatite or talc powder was added, and this amount does make the paste more plastic without major changes to the surface texture or glaze colour. Further studies

54

J. M. Kenoyer

Fig. 4 Experimental faience with steatite/talc (a Plain faience, no steatite/talc; b Faience with 9% steatite/talc, c Faience with 33% steatite/talc; d Faience with 50% steatite/talc)

were carried out in collaboration with Dr Heather M-L Miller to determine if different amounts of steatite/talc combined with quartz might have distinctive functional or visual properties (Miller & Kenoyer, 2018). SEM analysis of the experimental faience with 9% fine ground steatite/talc clearly shows the particles of steatite/talc mixed with the unmelted quartz and all are embedded in glass (Fig. 4b). Smaller particles of steatite/talc appear in lighter patches in the shape of small linear particles with irregular edges. In two other samples, 33 and 50% of ground steatite/talc was added to the paste (Fig. 4c and d, respectively). In these samples, the overall structure was less compact and more friable, and the steatite/talc particles are very visible throughout the body and surface of the faience object. The colour of the glaze and the surface texture were also altered significantly in these later samples. No examples of ancient faience containing steatite/talc have been found at Harappa. Two samples from Mohenjo-daro analysed by Ullah had around 27% magnesia which he associated with steatite/talc, and based on the experimental samples that I made in the lab, this amount of steatite/talc would result in a very friable and less compact form of faience. In summary, based on my current analysis of objects from Harappa, there is no evidence for the use of steatite/talc powder in any of the faience objects, and ‘steatite faience’ or ‘talcose faience’ does not seem to have become an important aspect of faience technology at the site or to my knowledge, any other Indus site.

Glazed Steatite and Faience Technology at Harappa, Pakistan …

55

3 Glazed Steatite and Faience from Harappa 3.1 Period 1, Ravi Phase Glazed Steatite and Faience Excavations of the Ravi Phase deposits during the 1996 season revealed a small room with a hearth and scattered pottery in situ on the floor that had a wide variety of beads made of terracotta, unfired steatite, fired and glazed steatite and faience (Kenoyer & Meadow, 2000) (Figs. 5a and 6). In the northeast part of the excavation area, a concentration of unfired steatite beads was found that was originally assumed to represent manufacturing waste from a bead workshop. Detailed examination using a digital microscope and also the SEM has shown that some of these beads were heavily worn and were probably part of a necklace or bracelet and not discards or inadvertently lost during the manufacturing process. However, fragments of sawn steatite were found in these same deposits along with pieces of faience slag or glazing slag, so there is no question that the glazed steatite beads and faience beads were being made somewhere in the vicinity, but just not in this domestic area (Kenoyer, 2012). Not far from these unfired steatite beads a necklace fragment made with fired and white glazed steatite beads was found in situ. Additional loose beads of fired and glazed steatite as well as one example of a faience bead were found on the floor and in the course of sifting. All of the beads were recorded at the site, and the best-preserved examples were put on display in the Harappa Museum. Analyses of some of selected

Fig. 5 a Harappa, Ravi Phase, Period 1, steatite beads, Harappa 1996, b Kot Diji Phase, Period 2, steatite beads, Harappa 1996

56

J. M. Kenoyer

Fig. 6 Harappa 1996, Trench 39S, Excavation plan showing locations of Ravi Phase steatite beads

beads are presented below to provide detailed information about the manufacture and glazing of these early examples from Harappa. Based on measurements of 364 complete glazed steatite beads from the Ravi levels, it is possible to see that there are two main types of beads, short cylindrical and long cylindrical with some examples of other shapes (Fig. 5a). The short cylindrical beads are from 0.4 to 3.5 mm in length with the majority grouped around 1.2– 2.2 mm. Longer beads range from 4 to 16 mm in lengths with some clusters of beads that appear to have been made in the same length (Fig. 7a). The overall bead diameters cluster in the range from 1.8 to 2.4 mm (Fig. 7b). The unfired steatite beads found along with the fired and glazed beads fall within the same range as the short cylindrical beads and do not appear to represent a different manufacturing tradition. The maximum drill hole diameters are between 0.8 and 1.2 mm in diameter (Fig. 7c). The analysis of the Ravi steatite drill holes indicates that most of them were perforated using a thin copper or bronze drill with a bevelled tip that was wider than the shaft, resulting in a relatively straight cylindrical drill hole (Kenoyer, 1997, 2005a: 162).

Glazed Steatite and Faience Technology at Harappa, Pakistan …

57

Fig. 7 Harappa, Period 1, steatite beads, a maximum length frequency, b maximum diameter frequency, c maximum drill hole diameter frequency

58

J. M. Kenoyer

Fig. 8 Harappa, Period 1, Ravi Phase steatite and faience beads, a Unfired steatite bead, H96 75143, b Fired steatite with red slip, H96 7511, c Fired and white glazed steatite bead, H96 7512-30a, d Fired and green glazed steatite bead, H96 7530-5, e Fired and blue-green glazed steatite bead, H96 7528-38, f Faience bead with blue-green glaze, H96 7512-11

Some steatite beads were unfired. They had a smooth worn exterior surface, and the rounded edges of the drill holes show that they were strung and worn either as an ornament or sewn onto clothing (Fig. 8a). In contrast to these unfired examples, most steatite beads from the excavations were fired and decorated with some form of slip or glaze. The main types of decorative surfaces documented so far are red slip (Fig. 8b), white glaze (Fig. 8c) or a coloured glaze of green (Fig. 8d) or blue-green (Fig. 8e). The one faience bead that has been examined had a blue-green glaze (Fig. 8f), and the other examples of faience beads (not illustrated) also have traces of blue-green glaze. The Ravi steatite beads were made by sawing and shaping bead roughouts or sawing thin sheets of steatite and then chipping or grinding them to the desired shape and drilling them. A large number of the very short cylindrical beads would have been strung together onto a strong thread, probably cotton and either ground on a flat grinding stone or rounded with a handheld, grooved stone. The rough grinding striae were not removed by polishing, but appear to have been intentionally left rough to allow the slip or glaze to adhere more firmly to the exterior surfaces of the bead. Some of the Ravi Phase steatite beads are made in the form of short bicones with a rough unpolished surface that has traces of a red slip. The slip is made from an iron-rich clay but since it cannot really fuse with the underlying talc, it is relatively fugitive and does not have a uniform red surface (Fig. 8b). In the case of glazed beads, sometimes the glaze is so thick that two beads became fused together and the glaze covers most of the area where they were touching during the firing process (Fig. 9a).

Glazed Steatite and Faience Technology at Harappa, Pakistan …

59

Fig. 9 Harappa, Period 1, Ravi Phase, a SEM image of two Harappa Period 1, glazed steatite beads stuck together covered with a white glaze (H96 7512-34), b SEM image of two glazed steatite beads stuck together with glaze, but the exterior glaze is eroded, revealing grinding striae (H96 7512-32)

Due to weathering and differential shrinkage, glazes do not adhere well to steatite, and they usually flake off (Fig. 9b). When the glaze has flaked off, the linear grinding striae or saw marks from manufacture are clearly visible. An excellent example from the Ravi levels shows the rough longitudinal grinding striae on two beads that are still fused together with some glaze though the external glaze has weathered away (Fig. 9b). The rough alignment of the grinding striae indicates that these beads were ground together on a string, fired and glazed together and then strung together on a string. The firing process has not been documented for the Ravi Phase, but probably involved placing strings of unfired beads into a firing saggar protected by layers of ash mixed with crushed limestone or crushed bone to keep them from sticking together, a process that has been documented in cementation firing of beads in Iran (Wulff et al., 1968). It is possible that the glazing of large quantities of short cylindrical beads was achieved using either the application method or cementation process, or possibly a

60

J. M. Kenoyer

combination of the two. The long cylindrical glazed steatite beads were coated with a glaze made of a flux mixed with ground silica and a colourant. These beads appear to have been fired in saggars that were lined with crushed bone, as many beads have setting marks with traces of bone adhering to one surface. The crushed bone results in a layer of sharp edges and points that allow the faience to rest above the flat surface of the container and avoids the glaze sticking to the terracotta. The tiny pieces of bone are easily removed after firing.

3.2 Period 1, Ravi Phase Glaze Compositions The glaze on Ravi steatite beads was analysed using SEM EDS (Hitachi S3400 VPSEM, at the Department of Geosciences, University of Wisconsin-Madison). Further studies are planned, but preliminary results indicate that the main components of the glaze besides SiO2 are the fluxing agents such as Na, Ca, K and Mg that probably derive from plant ash (Tite et al., 2006) and traces of colourants. White glazed beads have minor traces of Fe, P, Pb and S. The green glaze has larger amounts of Fe and some Cu, and the blue-green glaze has larger amounts of Cu and negligible Fe. In areas where bone setter marks are present, the amounts of P and Ca are quite high, but these are probably from the bone itself and not part of the glaze. Only one sample of a faience bead (Figs. 8f and 10a, b) was available for study using SEM EDS. This bead had traces of cotton fibre on the broken end, and this is the earliest evidence for cotton fibre at the site of Harappa. The analysis of the core and exterior glaze using SEM EDS was done on the broken section of the bead as well as on a polished section of the bead. Based on these analyses, the quartz body has some glassy phases that include Na, K, Ca, Mg that would indicate a plant ash for the flux (Tite et al., 2006). The colourants of Cu and Fe are found in the glassy part of the core as well as on the exterior surface, which suggests that the bead was glazed using the efflorescence process. Most of the glaze has eroded from the exterior, but some patches of glassy matrix are found on the interior of the bead (Fig. 10c, d). The quartz particles are up to 221 µm in size but many are between 20 and 50 µm, which means that the quartz was ground quite fine before preparing the bead. The relatively low percentage of glass on the interior suggests that during this early period, a frit was not prepared first to make a more compact faience. The small bubbles scattered throughout the core of the bead could indicate the use of acacia gum as a binder. In the SEM EDS analysis of the various inclusions in the glass and interspersed with the quartz, there were grains of Ti and traces of Pd, Ni and Au. The presence of Ti, Ni, Pd and Au is not expected if the quartz is pure. However, the sands of the Indus River as well as the Soan and Jhelum Rivers are famous as a source of gold and other minerals mentioned above (Halfpenny & Mazzucchelli, 1999; Law, 2011b: 92–93). As part of my research, I have analysed sands collected from the current bed of the Ravi River which is some 6 km west of Harappa today as well as archaeological samples of sand found in different locations on the site that would have been brought from the Ravi River during the Harappa Phase occupation (c. 2200–1900 BCE). In both the

Glazed Steatite and Faience Technology at Harappa, Pakistan …

61

Fig. 10 Harappa, Period 1, Ravi Phase glazed faience bead, H96 7512-11, a End showing location of cotton fibre, b Detail of cotton fibre, c SEM image 100×, d SEM image 300×

prehistoric and the modern sand, samples of crystals that contain Zr, Ti, Fe and Kfeldspar have been identified. In the modern sand sample, monzanite crystals with Th and Ce and other trace elements were identified. The Ravi River sands near Harappa represent erosion from many different geological rock formations, most significantly from the Himalayan ranges to the north, but the specific origin of the Zr and Ti in the Ravi River sands has not been determined. Another possible source for good quality quartz that also has some traces of gold (Shah, 1973; Ahmed et al., 2018: 1) is in the Kirana Hills, which are only 120 km north of Harappa (Law, 2011b: 31). Studies of other raw materials from Period 1, Ravi Phase by Law, have shown that the site of Harappa was already connected to the Kirana Hills for quartzite grinding stones and to the northern Indus for steatite (Law, 2011b: 463, Fig. 13.2). In summary, it is possible that both the Ravi River quartz sands as well as quartz from the Kirana Hills may have been used in the glazed steatite and faience industry at Harappa during the Ravi Phase as well as in subsequent periods.

62

J. M. Kenoyer

3.3 Period 2, Kot Diji Phase Glazed Steatite and Faience During the Harappa Period 2, Kot Diji Phase, glazed steatite beads continued to be made, but the shapes and sizes of the beads changed and include short and long cylindrical beads, as well as long barrel and oval and lenticular long barrel beads (Fig. 5b) (Kenoyer, 2005a: 162). The production of very short cylindrical beads and microbeads continued in much the same way as from the previous Period 1, Ravi Phase, but the first examples of ‘very very short cylindrical’ beads (Kenoyer, 2017a: 158, Fig. 6) start to be produced in graduated sizes with larger diameters. These beads are more commonly referred to as ‘disc’ beads or ‘wafer’ beads in the earlier literature (Beck, 1928: 4) and have a very large central hole as compared to the smaller beads that have very small holes. They were still produced in the same manner as the smaller beads but would have been strung on larger cords for grinding and rounding prior to firing and for use in ornaments. The multiple sizes of these thin beads suggest that they were made in graduated sizes though no complete ornament from the Period 2, Kot Diji Phase, has been discovered. A group of beads found together in the excavations in 1999 do show some graduating sizes (Fig. 11). The external surface of these beads is always fired with a thin white glaze, which often peels off leaving a rough surface that shows traces of saw marks on both faces of the

Fig. 11 Harappa, Period 2, Kot Diji Phase, very very short cylindrical steatite beads in graduated sizes found together, H99

Glazed Steatite and Faience Technology at Harappa, Pakistan …

63

bead and grinding marks along the exterior edge. Other shapes of beads were also being produced, including oval barrel and long barrel or biconical beads. In terms of steatite bead sizes, the same basic range of bead lengths was being produced during the Period 2, Kot Diji Phase, though the sample from the Harappa 1996 excavations shows that the maximum length for long cylindrical beads is slightly shorter than those from the Period 1, Ravi Phase (Fig. 12a). In terms of very very short cylindrical beads, their maximum diameter reaches around 14 mm though the majority of the beads are between 2 and 3 mm in diameter (Fig. 12b), which is roughly the same range as the Period 1, Ravi Phase beads (Fig. 7b). The other major difference is the overall drill hole size. The bead drill hole for very very short cylindrical beads is up to 3.8 mm which indicates the use of fairly thick cords that would have been used for threading the beads during production as well as for use. The larger diameter cord may have been necessary to support larger weights of bead as a thin cord would have broken during the grinding process and also when the beads were worn as an ornament. The larger diameter holes are seen only on larger diameter beads, and the majority of small short and long cylindrical beads were still perforated with drills that range between 0.8 and 1.2 mm (Fig. 12c), which is much the same as is seen in the Period 1, Ravi Phase.

3.4 Period 2, Kot Diji Phase Glazed Seals During Period 2, Kot Diji Phase, we see the first evidence for the use of steatite to make glazed seals. Prior to this time period, button seals were made from bone, terracotta and ivory, but not from steatite (Durrani et al., 1995b; Jarrige et al., 1995; Kenoyer & Meadow, 2000: l). Two examples of unfired steatite seals were discovered at Harappa. One large seal has four intersecting lines that cross the centre of the seal and create eight triangular areas on the seal. The unfired grey steatite seal may have been used for stamping clay or some other material since the edges are smooth and rounded. The boss on the back was very small and appears to have been set off centre to accommodate a flaw in the rock (Fig. 13a). The second example is a broken, unfired steatite intaglio seal carved with an elephant motif that was made from light brown or tan steatite (Fig. 13b). The carving was relatively rough and not very precise. On the reverse side, there is evidence for the presence of a boss that was drilled at an angle as is seen in later Period 3 seals. Three other seals provide clear evidence for the firing of steatite at relatively high temperatures, and on two seals, there are traces of a thick blue-green glaze on the roughly carved surface. One small circular button seal or possibly an ornament has two holes drilled through the body of the ornament (Figs. 13c and 14a). The dark grey-coloured steatite can be seen below a whitened surface that indicates firing at over 900 °C, and traces of a thick blue-green glaze are seen on the interior circular area (Fig. 14a). This glaze has not yet been analysed, but on the basis of the analysis of Period 1, Ravi Phase glazes, it was probably made using copper as a colourant and a plant ash flux. After firing, the outer layer was

64

J. M. Kenoyer

Fig. 12 Harappa Period 2, Kot Diji Phase steatite beads, a Maximum length frequency, b Maximum diameter frequency, c Maximum drill hole diameter frequency

Glazed Steatite and Faience Technology at Harappa, Pakistan …

65

Fig. 13 Harappa, Period 2, Kot Diji Phase, glazed seals, a Square button seal, unfired steatite, H96-2743/7402-90, b Elephant seal, unfired steatite, H2000-4474/8994-01, c Square glazed button seal, H96-2740/7469-01, d Circular blue-green glazed steatite button seal, H96 7458-1, e Square blue-green glazed button seal, H2000 4495/9597-10

Fig. 14 a Harappa, Period 2, Kot Diji Phase, glazed seals, a Circular blue-green glazed steatite button seal, H96 7458-1, b Square glazed button seal, H96-2740/7469-01, c Square blue-green glazed button seal, H2000 4495/9597-10

66

J. M. Kenoyer

hardened, but through use and also possibly repeated stamping into clay, the exterior edges were worn and rounded. Another example of a seal with traces of glaze has a geometric motif that represents a form of house with a double pointed roof (Figs. 13d and 14b). Similar motifs are found on later Harappan seals as well as on moulded tablets and appear to represent a sign that is part of the Indus script (Kenoyer, 2006). The rough carving on this seal appears to have been intentional in order to allow the glaze to adhere to the surface of the steatite. Another very distinctive motif is a deeply carved design with five circle and dot motifs. One circle and dot motif are in the centre, and the other four are arranged around a four-pointed star (Fig. 13e). This seal was originally covered with a thick blue-green silica glaze, traces of which can still be seen along one edge and on the face of the seal (Fig. 14c). Similar seals with circle and dot motifs have been found at other Regionalization Era sites including Tarakai Qila, Pakistan (Allchin & Knox, 1981), and Kunal, Haryana, India (Khatri & Acharya, 1997). Other seals with similar carving styles and glaze have recently been discovered in new excavations at Rehman Dheri by Peshawar University and the Department of Archaeology, Khyber Pakhtunkhwa (Personal observation, courtesy of Dr Zakirullah Jan and Dr Abul Samad). The style of carving seen on these various seals suggests that they may have all been made in the same workshop or by craftspeople trained in a very specific workshop style or tradition. Further studies of the glazes and manufacturing methods are needed to determine the production area for these distinctive glazed button seals.

3.5 Period 2, Kot Diji Phase Faience Only a very small sample of faience beads have been recovered from chronologically secure Period 2, Kot Diji Phase contexts, at Harappa, but they do show some new developments in terms of bead shape and production technique. At present, no detailed analysis has been carried out on Period 2 glazed steatite or faience beads, but a brief discussion of the finds and new developments show the importance of these technologies in the emergence of Harappa as a major urban centre. One of the new developments in faience bead making is the production of segmented short or long barrel beads (Fig. 15a, b). These beads were probably made on a fine reed or wooden stick and scored to separate individual beads. They appear to have been fired together and then broken into individual beads or shorter segments of two or three beads after firing and glazing (Fig. 15b). The preparation of multiple beads at the same time suggests that glazed beads were in high demand and were being produced on a larger scale than in the earlier Ravi Phase. The other types of beads include long cylindrical faience beads (Fig. 15c), long barrel beads (Fig. 15d, e) and long lenticular barrel beads (Fig. 15f). Most of the beads have a blue-green colour and appear to have been made using the efflorescence technique.

Glazed Steatite and Faience Technology at Harappa, Pakistan …

67

Fig. 15 Harappa, Period 2, Kot Diji Phase, faience beads, a Short barrel segmented bead, bluegreen glaze, H98 8491-25, b Short barrel, yellow to blue-green segmented beads, c Long cylindrical blue-green glaze, H96 7464-1, d Long barrel blue-green faience bead, H96 7474-1, e Long barrel blue-green faience bead, H96 6509-5, f Lenticular, long barrel blue-green glaze, H99-4351/8492119

3.6 Period 2, Faience Bangles and Vessels Terracotta bangles were first produced in Period 1, and during Period 2, the variety of terracotta bangles that were produced increased dramatically. Red-fired terracotta bangles were made with different shapes, and black-fired terracotta bangles with incised designs were also produced. The diversity of bangle types probably relates to the needs of many different communities to differentiate themselves with ornaments that provide a non-verbal form of communication regarding ethnicity, status and wealth (Kenoyer, 1992b: 90–91, 96–97). It is not surprising that faience was also used to make bangles, but the production of faience bangles requires an important new technology of faience production since normal faience is not strong enough to hold the delicate shape of a bangle without breaking (Kenoyer, 1994b). Examination of faience bangles from Period 3, Harappa Phase, indicates that faience bangles were made with a special form of compact faience (McCarthy & Vandiver, 1990) that required the production of a glassy frit that was reground to form a finer paste that could then be re-fired to develop a stronger vitreous faience (Kenoyer, 1994b).

68

J. M. Kenoyer

Fig. 16 Harappa, Early Harappan faience objects; a Plain faience bangle Period 1, Ravi Phase (H98/8513-1), b and c Faience slag, Period 1, Ravi Phase (H98/85145,6), d, e & f Incised chevron design bangles with blue-green glaze, Period 2, Kot Diji Phase (H2000/8990-107, H99/8950-22, H90/1157-7), g incised faience vessel (H99/8492-120)

Only one example of a faience bangle has been recovered from a secure Period 1, Ravi Phase context, and it had a plain design (Fig. 16a). Lumps of glassy slag that might have been used to make a faience frit were also found from the later Ravi levels (Fig. 16b–c), and it is possible that the production of compact faience for making bangles was started during the Ravi Phase. More examples of faience bangles, including incised bangles, have been found from the Kot Diji Phase deposits, and all have remnants of a blue-green glaze (Fig. 16d–f). One of these bangles (H88/749) was available for SEM EDS analysis (Fig. 17), and the composition of the glass indicates that a plant ash flux was being used and copper with some zinc was used for a colourant to make a blue-green glaze. The presence of zinc with the copper could indicate that the copper came from a polymetallic ore source as is found in the southern Aravalli Range where copper, zinc and lead are found together (Patel & Ajithprasad, 2018). The quartz grain sizes range from 27.6 to 77.5 µm but most fall in the 30–40 µm range, which is very fine (Fig. 17b, d). The amount of glass in the core area of the bangle indicates that this is a compact faience possibly made using a reground frit that would result in a very strong structure of the bangle. Additional examples of plain and incised faience bangles have been recovered from levels that are at the interface between the Period 2 and Period 3 occupation. It appears that the production of faience bangles may have started towards the end of Period 1 and expanded during Period 2. The production of many different styles of faience

Glazed Steatite and Faience Technology at Harappa, Pakistan …

69

Fig. 17 Kot Diji Phase faience bangle SEM EDS, H88/749, a SEM 700×, b SEM 300× with measurements of external glaze and quartz grains, c SEM 100× showing compact glassy core and external glaze, d SEM 700× with quartz grain measurements

bangles, rings and other ornaments becomes more widespread at the very beginning of Period 3A and on through the Harappa Phase. Only one example of a faience vessel (H99 8492-120) has been recovered so far from Period 2, Kot Diji Phase deposits (Fig. 16g). Detailed studies of this fragment have not yet been carried out, but studies of Period 3, Harappa Phase faience vessels (see below), indicate that faience vessel production involved the use of compact faience. This is necessary in order to produce a thin strong vessel wall that is nonporous and will not break in the course of use. On the basis of faience beads, bangles and one vessel fragment, it is possible to suggest that faience production was becoming more important by the end of Period 2. The blue-green glazed faience beads and bangles would have been important high-status ornaments, and the faience containers were probably used to hold valuable aromatic oils used by elites. These developments appear to coincide with the expansion and formal division of the site into two distinct walled habitation areas, Mound AB and Mound E (Kenoyer, 1993; Meadow & Kenoyer, 2005). The elaboration of high value crafts like faience would have been closely linked to the demands of elites and expanding urban populations (Kenoyer, 2000).

70

J. M. Kenoyer

3.7 Period 3, Harappa Phase Glazed Steatite and Faience During the subsequent Harappa Phase, there is a dramatic increase in the use of glazed steatite for producing many different types of objects used by all levels of Indus society, as well as special items for elites. The range of objects includes beads of many different shapes and sizes, pendants, inlay, figurines and composite larger size sculptures as well as geometric seals and inscribed tablets. Faience objects included all of the same categories, but since faience is a plastic material it could be shaped, carved, moulded and glazed with different colours to produce an even greater variety of objects including new categories such as vessels. A selection of objects from each category of objects made from glazed steatite and faience will be presented below to illustrate the major features of production and glaze composition. It is not possible to discuss all aspects of production for each object, and the main focus will be on glazing technology and composition. The majority of the objects presented below come from the uppermost levels of Harappa, dating to Harappa Phase 3B and C, but a few examples from 3A are discussed when possible.

3.8 Period 3, Harappa Phase Glazed Steatite Beads and Pendants The production of glazed steatite beads increased dramatically during the Harappa Phase, and presumably, this was needed to meet the demand of the growing urban population as well as for trade to surrounding regions. In Period 3A deposits excavated on Mound AB (Trench 42), a copper saw and a collection of sawn steatite bead blanks were discovered that indicated the development of highly specialized production. The copper saw blade (Fig. 18a) measured 0.35–0.4 mm thick and was denticulated to allow for rapid production of thin slices of steatite from raw blocks (Fig. 18b– e) (Kenoyer, 1997: 268–269). The thin slices of steatite were sawn or chipped into squares or circular shapes (3–5 mm wide and 0.5–1.0 mm thick) (Fig. 18f, g) and then drilled using a fine copper drill with a bevelled tip. The surfaces of all fired steatite beads show evidence of whitening and traces of glaze that in some cases is still clearly visible (Fig. 19a), but in many instances, the glaze has flaked off revealing the underlying saw marks (Fig. 19b). The glaze clearly does not adhere well to the surface of the steatite beads, and the rough saw marks appear to have been left intentionally to help the glaze to stick better. Based on my own experimental replications, the glaze is quite well fused to the beads, but after long burial in the ground and more than 4500 years of weathering, the archaeological beads tend to lose the glaze. Most of the steatite beads found at Harappa are found in secondary contexts with only a few found in primary contexts that demonstrate how they were used. In the cemetery excavations, one of the female burials (127a) had anklets made of very very short cylindrical steatite beads (Dales & Kenoyer, 1991: 201, Fig. 13.12), and a male burial had a long necklace of around 308 of the same type of bead in graduated

Glazed Steatite and Faience Technology at Harappa, Pakistan …

71

Fig. 18 Harappa Phase, Period 3A, steatite bead production, a Copper saw, b Sawn steatite, H94 5108-9, c Sawn steatite, H94 3948-28, d Sawn steatite H94 3948-28, e Sawn steatite, H94 3936-11, f Steatite bead blanks, H97 Trench 42, Period 3a, g Harappa Period 3A, steatite bead blanks and roughouts, H97 7783

72

J. M. Kenoyer

Fig. 19 Harappa, Period 3, Harappa Phase glazed steatite beads, a H88 703-9, b H89 1011- 42_145

sizes (Fig. 20). These beads may have been made together as many of them seem to fit together precisely as a result of rubbing against each other in the manufacturing process. The surfaces of the beads are glazed, and it is surprising that they did not stick together when they were fired which is an issue that will be discussed below. The beads show that they were sliced from thin sheets that were not uniformly thick and then ground together on a string before firing. Bead size frequencies for Period 3 are not useful for comparison with the earlier periods since necklaces and anklets were made in graduated sizes. The most amazing aspect of Harappan steatite technology is the production of microbeads that are less than 1 mm in length and less than 1 mm in diameter. They were made from thin sheets of steatite that was sawn into tiny blanks that were then drilled and strung on a sturdy thread for grinding. They were probably fired in canisters using a flux to glaze the surface of the beads, and unlike the earlier Ravi Phase beads, none of these beads appear to have fused together during the firing. It is possible that they were buried in layers of ash mixed with steatite and bone powder to keep them from sticking together. The best example of how they were actually worn has been found at Harappa where a male burial (147a) was discovered with thousands of steatite microbeads arranged in a hair ornament that included three shell bangles or rings and a jasper bead (Dales & Kenoyer, 1989: 91, 1991: 200, Fig. 13.11) (Fig. 21a, b). Studies of similar micro beads found from Chanhudaro show traces of glaze on the surface and the underlying striae from grinding (Fig. 21c). Traces of charred fibre have also been found inside these microbeads from Chanhudaro, and my analysis indicates that these examples were strung on cotton thread and then fired and glazed. After removal from the firing canister, the charred cotton fibre was still preserved inside the beads, and the bead hole around the fibre was packed with steatite powder. This powder is either the remainder of the powder that accumulated inside the bead

Glazed Steatite and Faience Technology at Harappa, Pakistan …

73

Fig. 20 Glazed steatite necklace from burial (H194a) and other steatite beads, Harappa, Period 3, Harappa Phase

during grinding, or possibly it is the remainder of steatite powder used to separate the beads to keep them from sticking during firing. Further studies of this powder are in progress. It is possible that the beads were coated with a flux, and then, after the flux was dry, they were placed inside canisters while still strung on a cotton thread. To keep them from sticking, a mixture of powdered steatite and bone may have been used to coat the containers. It is also possible that the glaze was made through cementation by adding ash or flux to the powdered steatite and bone in much the same way is done when firing faience beads with the cementation process described by Wulff in Iran (Wulff et al., 1968). In his study of faience beads made with the cementation process, he notes that the optimal firing temperature is around 1000 °C and beads fired below this at 900 °C do not have a good glaze formed, while firing at 1100 °C results in beads fusing with the ash and flux used for the cementation process (Wulff et al., 1968: 101). Examples of strings of steatite beads that were

74

J. M. Kenoyer

Fig. 21 a Steatite microbeads from burial H88-147a, b Detail of H88-147a microbeads with human hair for scale, c SEM of microbeads from Chanhudaro showing charred cotton fibre inside the beads

fused with the firing canisters have been recovered from Lakhan-jo-daro (Fig. 22a) that could be evidence for the use of cementation. In addition, at Harappa, numerous examples of steatite beads fused with glassy slag and parts of firing containers or pieces of bone have been recovered (Fig. 22b). These examples could also indicate the use of cementation and not just accidental dropping of beads on molten slag. Further studies of the excavated examples mentioned above as well as experimental replications are needed to determine if cementation might have been used to create glazed surfaces on steatite beads in the Indus. One final note is that after the beads were fired, they had to be removed from the containers and restrung on a new fibre as the original fibre was charred in the firing process. Beads could have been restrung

Glazed Steatite and Faience Technology at Harappa, Pakistan …

75

Fig. 22 Lakhan-jo-daro steatite firing canisters, a Strands of long cylindrical steatite beads in ashy matrix (Photo by Qasid Mallah), b Very very short cylindrical steatite beads fused to a firing tile. (Photo by J.M. Kenoyer courtesy of Qasid Mallah)

on cotton thread, and this may have been the most common technique, but analysis of other samples of steatite microbeads from Chanhudaro by Irene Good identified the fibre inside the beads as being silk (Good et al., 2009). It is necessary to carry out additional studies to confirm this identification, but if it is confirmed, then for the microbeads at Chanhudaro, they were first strung on cotton, and then after firing, they were restrung on silk thread after they were removed from the firing container and prior to sale or use in ornaments. Examination of fired steatite beads indicates that all of the white-fired steatite was glazed using some form of flux with or without additional silica. Experimental replication of glazing using only plant ash flux (sajji or khar) results in a fine glassy surface on the fired steatite. The silica in this glaze must derive from the composition of the flux combined with small amounts of silica drawn from the steatite itself as no additional silica was added. Further studies of this glazing technique are being carried out in collaboration with Randall Law (Law, 2018), and it is possible that this is the same technique used to glaze steatite seals. Experiments using a glaze with added ground silica resulted in a thicker glassy surface than is normally seen on steatite beads, so this technique may have been used only for certain types of beads with thick glaze. Based on discoveries at Harappa and other sites, the firing of the steatite beads appears to have been done in straw and sand-tempered terracotta canisters or saggars combined with flat tiles covered with steatite powder and crushed bone (Kenoyer, 2005c). This white powdery layer comprised of fired steatite powder and/or crushed bone has been found on the interior of canisters as well as on firing tiles found at Harappa ( Miller et al., 1996; Miller, 2008a) and also at Lakhan-jo-daro (Mallah, 2008) and Chanhudaro (Didier, 2017; Vidale, 1987). Experimental replication of firing using both crushed bone and steatite demonstrates that both are very effective for keeping the glazed steatite or faience from

76

J. M. Kenoyer

sticking to the terracotta of the firing container or saggar (Kenoyer, 2005c). Sometimes the firing may have gone on too long or gotten too hot, and there are numerous examples from Harappa of steatite beads fused to the surface of canisters or tiles. One of the best examples of this type of firing accident come from the very large steatite bead manufacturing workshop excavated at the site of Lakhan-jo-daro in 2013–14 by Dr Qasid Mallah and his team from Shah Abdul Latif University Sindh. A vitrified terracotta firing canister was discovered that had a fragment of a white coated tile and strands of steatite beads fused to an ashy matrix that included large pieces of calcined bone (Fig. 22a). Another example had larger steatite beads and bone fused to a firing tile (Fig. 22b). In both cases, the kiln temperature must have reached around 1100 °C or higher for the steatite beads to fuse to the terracotta and surrounding ash as is noted by Wulff in his studies of cementation glazing in Iran (Wulff et al., 1968: 101). Analysis of the ashy powder on the steatite beads as well as the white ashy surfaces on other tile examples shows the presence of steatite powder and crushed bone, but some points have higher traces of Na, Mg, Ca, P and some S that may indicate the use of a flux mixed with the ash. These discoveries from Lakhan-jo-daro confirm that the steatite beads were fired on strings and many layers of beads would have been placed in a single canister or tile for firing. Further studies are needed to confirm if these vitrified firing containers were used in the cementation process or if they were used to fire steatite that had been coated using the application process.

3.9 Period 3 Glazed Steatite Button Seals and Intaglio Seals The production of glazed steatite button seals and glazed intaglio seals with script became highly developed and also diverse during Period 3 at Harappa as well as at other major Indus site. Most steatite seals appear to have been glazed with white glazes that often peel off after weathering to reveal the saw marks and carving lines of the original manufacture (Fig. 23). On a seal from Period 3A at Harappa, the traces of the glaze can be seen over the working marks on the surface of the seal (Fig. 23a), so it is evident that the technique used to glaze seals had begun at the very beginning of Period 3. On other seals, the glazed surface is cracked but has not spalled off yet (Fig. 23b, c). In other seals, the glazed surface has worn away or peeled off revealing the underlying unfired steatite (Fig. 23d), which indicates that in some cases the firing of the seals was done in manner that only heated and glazed the external surface and not the core of the seal. Attempts at replicating this process have been partially successful by heating seals in a container surrounded by burning charcoal and using a blowpipe or fan to heat the coals and bring the seal to a red-hot colour. This process however results in some carbon smudging of the surfaces, and no Indus seals show this feature. The precise technique used by Indus craftspeople to rapidly fire seals to high temperatures without carbon smudging is still a mystery. The large unicorn seal is from Period 3C (Fig. 23e), and it has a very fine glazed surface that shows very little cracking. The fine quality of this glaze on such a large

Glazed Steatite and Faience Technology at Harappa, Pakistan …

77

Fig. 23 Harappa, Period 3, Harappa Phase glazed steatite seals, a Bull seal, Harappa Period 3A (2600-2450 BCE), b Unicorn seal, steatite, Harappa, H87-262, Period 3C (2200-1900 BCE), c Unicorn seal, steatite, Harappa, H96-2736, Period 3C (2200-1900 BCE), d Chimera seal, steatite, Mohenjo-daro, Lahore Museum, P-1727, e Unicorn seal, steatite, Harappa, H99-4064, late Period 3C (2000-1900 BCE)

surface suggests that by the final phase of Period 3, seal makers had perfected the art of glazing steatite in a manner that resulted in a strongly fused glaze. Earlier scholars had identified the glaze on steatite seals as being a talcose glaze (Mackay, 1931: 379) or an alkaline treatment of the surface (Beck, 1934: 80–81; Vidale, 2000: 62) or a combination of the two as suggested by Law (2011b: 258, Appendix 7.15). Since talc cannot form a glaze, the glassy surface must be a form of silica that is derived either from the silica in the steatite or from an applied silica combined with a flux. In his detailed analysis of the white exterior glazed layers, Law found that the ‘The carved steatite body of the object is covered by an extremely thin (≈20 µm) layer composed of talc that has been heated to a temperature of 1200 °C or greater (and, thus, it is no longer talc but rather the minerals enstatite and cristobalite) as well as a calcium phosphate slip. …The calcium phosphate detected on the outer surface of the seal and preserved in a micro-crack within its body is perhaps a boneash slip/treatment of some kind that was added after the talcose layer was applied’. (Law, 2011b: 667). My own experiments at whitening and glazing the surface show that the external glazed surface can be achieved by first soaking the steatite in a solution containing plant ash flux or natron for several hours and then drying and firing the pieces in a muffle furnace to 1000 °C. The process by which this glaze forms is still being studied. Further experiments of weathering and analysis of the layers produced by weathering are also needed to see if this process can replicate what has been proposed by Law.

78

J. M. Kenoyer

Fig. 24 Harappa, Period 3, Harappa Phase, blue-green glazed steatite button seal/ornament

While most button seals and intaglio seals were fired and glazed with a white colour, some small ornamental button seals that may have been used as buttons continued to be glazed with blue-green silica glaze as was common in earlier periods (Fig. 24). Due to weathering, this glaze peels off, and this may explain why most of these types of objects appear to be white-fired steatite without any glaze at all. So far, no analyses have been conducted on the blue-green glazes found on steatite button seals, but we can assume that they are coloured with copper, which is the main colourant for glazed steatite in earlier periods (see above). The discovery of drips of blue-green and green glaze on pieces of bone at Harappa and lumps of frothy slag can be interpreted as evidence for the production of a frit that was then ground and used as an applied glaze. The frit may have been made by melting silica and a colourant, either copper (Cu) or in the case of green glaze, iron (Fe) combined with a plant ash flux that contains Na, K, Mg, P and Ca. These glaze-covered bone fragments and slag may also be related to the process of frit manufacture needed for the production of compact faience discussed above relating to faience bangles.

3.10 Period 3 Faience Production A workshop with steatite moulds and discarded faience tablets (Fig. 3) was discovered along the southwestern edge of Mound E in 2000 and 2001 that has provided important evidence for both the manufacture and the glazing technology of steatite beads, inscribed steatite tablets, faience beads and faience moulded tablets (Kenoyer, 2005c). The workshop is dated to Period 3B (2450–2200 BCE) on the basis of pottery found in association with the workshop debris, and it was located in a small, restricted

Glazed Steatite and Faience Technology at Harappa, Pakistan …

79

access area or courtyard between four walls. The main deposit of the workshop consisted of layers of manufacturing debris interspersed with layers of charcoal and ash. The absence of a formal kiln suggests that the area was used for firing faience and steatite using canisters fired in a bonfire type of furnace. It is possible that a temporary kiln was constructed but no traces were found in the debris. Firing experiments with an open bonfire have demonstrated that it is possible to reach temperatures around 935 °C which is sufficient for glazing both the experimental faience tablets and firing the steatite tablets (Kenoyer, 2005c). Kilns for firing faience have been reported from the site of Kuntasi (Dhavalikar, 1993; Dhavalikar et al., 1996) and also at Lakhan-jodaro (Personal Communication, 2013, Qasid Mallah), but there has been no analysis of the interiors to confirm that they were used specifically for faience firing. Faience bead making is also reported from the site of Bagasra (Gola Dhoro), but no specific kiln structure was found that could be linked to the production (Bhan et al., 2005). The earliest levels of the workshop on Mound E at Harappa had evidence of faience and steatite bead making, while in the later levels of the workshop, the discovery of five steatite tablets and eight moulded faience tablets demonstrates that tablet manufacture was also being carried out along with bead making. Over 732 over-fired, blackened and misfired beads were discarded in the workshop debris layers along with firing canisters (1009 fragments weighing over 37.39 kg), bone with faience slag (13.29 kg), bone without slag (16.66 kg) and faience slag without bone (7.44 kg). In addition, there were many pieces of unfired steatite as well as partly made steatite tablets and a steatite mould that matched a moulded faience tablet fragment (Fig. 3b3 and k). The canisters were vitrified inside with green and greenish black colour, and some had layers of bone and crushed steatite fused to the inner surface. Most of the beads that had traces of colour were blue-green glazed, and most of the tablets were either blue-green or yellowish white with some tablets made with two colours, black and yellowish white. In contrast to the limited colours of glazes used on steatite during the Harappa Phase, faience production expanded from primarily blue-green and white to a wide range of colours. The analysis of colourants found in the glazes indicates that iron was used to make deep red–orange, light red–orange, various shades of brown that almost look black. Deep azure blue was made using cobalt, which is the earliest evidence for the use of this mineral in the production of blue glaze in South Asia. Traces of cobalt have been reported from a purple faience bead found at Nausharo (dating to around 2600 BCE) (Barthélemy De Saizieu and Bouquillon, 2000: 101, Table 4), and at Harappa, there are numerous examples of dark blue beads that were probably coloured with cobalt and numerous examples of vessels with decorative lines that were made with deep blue cobalt, dating from around 2450–1900 BCE. Identification of the cobalt was confirmed using SEM–EDS as well as with LAICP-MS. The use of cobalt appears to have continued into the Late Harappan Phase (1900–1300/1000 BCE) on the basis of deep azure faience beads found in a small pot on Mound AB at Harappa, but so far, the beads have not been analysed. In addition to the development of new colours, faience craftspeople appear to have taken advantage of the plastic nature of the medium to fashion many new types of three-dimensional objects mentioned above. It is not possible to discuss all objects made of faience,

80

J. M. Kenoyer

so the main focus will be on beads, bangles and rings, seals, moulded tablets and vessels.

3.11 Period 3 Faience Beads Faience beads appear to have been made with hand modelling and sometimes carving the surface to incise lines or decorative surfaces. Most faience beads were made with only one colour, with blue-green from copper the most common, followed by yellowish white and some white beads (Fig. 25). The most common faience bead shapes include short spherical, short and long barrel, short and long bicone and long cylindrical. Microbeads made from faience have been found at Lothal (Dennys Frenez and Randall Law, personal communication) and possibly at other sites in Gujarat, but no examples have been recovered yet from Harappa or Mohenjo-daro. Several faience beads from Period 3 have been studied using SEM EDS. The sizes of the quartz particles indicate that they were made using compact faience, and the presence of copper colourant in both the core and the external glaze indicates the use of the efflorescence technique. One bead (Fig. 26a) has a distinct glassy layer on the exterior that is 90–117 µm thick, and the compact vitreous structure extends another 1.7 mm into the interior of the bead. The core of the bead has more bubbles and pores (Fig. 26b). The size of quartz particles in the main body of the bead is between 18 and 30 µm which suggests that the paste was made from a reground frit (McCarthy & Vandiver, 1990; Kenoyer, 1994b). The source of the quartz for this bead may derive from quartz sand as there are tiny fragments of zircon and titanium scattered throughout the body of the bead. In this bead, there was also a tiny piece of bone that may have been inadvertently included in the paste during the grinding process. Although most beads appear to have been made of compact faience, some very large beads that were incised with lines and cross hatching may have been made from less compact faience. These beads are usually found in the latest levels of the site dating to Period 3C, but some may come from the Late Harappan Period 4/5 as well. When studied with a hand lens or even under the microscope, the core of these beads is generally less glassy and without any colour, while the surface has a blue-green glaze. At first, it was thought that these beads may have been glazed with the application process or cementation, but traces of copper were found in the glassy part of the core area based on SEM–EDS, and it appears that these beads were made with the efflorescence technique. Beads made with two colours were usually white with red–orange, or dark brown lines that spiral around the bead or occasionally as a band in the middle of the bead (Fig. 25l). Some beads appear to have been made to simulate banded stone beads that were often produced from basalt that had veins of quartz or banded siltstone. The white glazes have higher amounts of Ca and P that may be the result of using bone as a colourant as well as a fluxing agent. The most vivid red–orange colour is found on beads that were made to imitate natural eye beads made from jasper or

Glazed Steatite and Faience Technology at Harappa, Pakistan …

81

Fig. 25 Harappa, Period 3, Harappa Phase selected faience bead types, a Segmented, b Long cylindrical, c Square long barrel, d Ring with lateral perforation, e Short barrel, f Long barrel, g Long bicone, h Long barrel, i Spherical, j Short cylindrical incised design, k Three-hole spacer, l Spiral banded, m Unperforated long cylindrical with dot motifs, n Ribbed circular, o Plano-convex long barrel and circular eye bead, p Long cylindrical made with joined pieces of faience

bleached carnelian (Fig. 25o). The red- and the white-coloured areas both have a glassy surface that is almost identical in terms of reflectivity, and yet they are not covered by an overlying glaze. This means that red paste and the white paste were made with appropriate amounts of flux so that they would both glaze at the same temperature. This is extremely difficult to accomplish, and so far, I have not been

82

J. M. Kenoyer

Fig. 26 Faience blue-green glazed bead SEM a Showing compact core and glassy surface glaze, b Sizes of quartz particles, H87 525-26

Glazed Steatite and Faience Technology at Harappa, Pakistan …

83

able to achieve either the red–orange colour or the glazing of two colours at the same time. The analysis of one of these beads by R.H. Brill determined that the colourant for the red faience was primarily iron (Fe2 O2 ) with traces of antimony (Sb2 O2 ). The flux was determined to be a combination of soda and potash (Na, K, Mg and Ca) (unpublished analysis by Robert R.H. Brill, Corning Museum of Glass). Analysis of a broken cylindrical ornament with white dots on a red background also revealed the use of iron for the red colouring (Fig. 25m), but experimental replication with various types of iron has not been able to achieve the brilliance of the Harappan red–orange glaze. One rare example of a bead with two colours was made from pieces of fired glassy faience that were cut and shaped to fit together around the ends of the bead (Fig. 25p). It is not known what type of adhesive was used to fuse these together, and it is possible that they were partly fused in the firing process. Many other types of ornaments were made with faience including inlay, buttons, ear plugs and gaming pieces.

3.12 Period 3 Faience Bangles and Rings Harappa phase faience bangles and rings were made in many different sizes and designs (Fig. 27). Most of the bangles were circular in shape and have a circular or oval cross section. The majority of the bangles were made with a plain blue-green glaze or occasionally with a white glaze. Wide plain bangles with a central ridge were made with flat sloping or concave sides. Another common form of bangle has a plano-convex or bevelled ridge with incised designs that include repeating chevron motifs, diagonal lines or variations of these motifs (Fig. 27a–k). Wide and narrow white glazed faience bangles were made with a single chevron incised design that imitates wide and narrow shell bangles. Wide bangles with a central ridge were also sometimes incised with diagonal or cross hatch designs. The most complex forms of bangles have deeply carved projections or wavy lines, and the shape of the bangle is not round but rather oval with an internal projection (Fig. 27d and n). This shape, which is similar to the shape of shell bangles, can be interpreted as a fertility symbol relating to the womb (Kenoyer, 1998: 109–110). All of the bangles were made using a single colour, but the incised lines resulted in some areas having darker or glassier highlights. So far, no examples of polychrome bangles have been recovered. Faience bangles from Period 3 appear to have been made exclusively from compact faience as this is the only way to make a strong circlet that could withstand casual impact without breaking. The glassy extends throughout the bangle, and the exterior has a thin layer of glass that gives it a smooth and shiny appearance (Fig. 28). It is interesting to note that unlike terracotta bangles which are often found in complete circlets and unbroken, faience bangles are rarely found as complete circlets. It is possible that faience bangles were intentionally broken when they were no longer being worn rather than simply being discarded. This breaking of the bangles might have had some cultural significance since it is highly likely that faience bangles may have been much more valuable than other materials. The glassy matrix extends

84

J. M. Kenoyer

Fig. 27 Harappa Phase, Period 3, a to k Faience bangles, l and m Faience rings, n Complete ridged bangle, Harappa 13,041, National Museum Karachi, NMP 54.3447

throughout the bangle, and based on the presence of copper colourant in the core as well as in the glaze, these ornaments were made using the efflorescence technique. The flux used for melting the silica was a plant ash (sajji or khar), and the source of the quartz appears to have been from Ravi River sand, though some bangles may have been made from the fine powder of white rock quartz pebbles. The quartz particles range in size from 20 to 70 µm but most are smaller than 50 µm. Small rings made of faience are also relatively common during Period 3 and could have been worn on the fingers or toes or tied into the hair. One additional possible

Glazed Steatite and Faience Technology at Harappa, Pakistan …

85

Fig. 28 Faience bangle with blue-green glaze SEM- EDS, CAL 28, H89 1060

use is for placing around the feet of pet birds as a form of ornamentation and also identification. This practice is common throughout the Punjab today and might have been a practice in ancient Harappa as well. The rings were often made as single circlets with a blue-green glaze, but many rings were made of two or more circlets joined together, or as wide bands with various types of incised lines or motifs (Fig. 27 l, m). The analysis of the faience ring composition indicates that they were made in the same way as the larger bangles, with compact faience using the efflorescence technique and that copper oxide was the main colourant for making blue-green glazes.

3.13 Period 3 Faience Glazed Button Seals and Moulded Inscribed Tablets The production of glazed seals with geometric motifs and moulded faience tablets represent new types of faience objects that were not found in the earlier periods at the site (Fig. 29). The motifs on square seals were probably made using a mould, but the boss on the back and the perforated hole would have been made by carving and drilling the faience when it was partly dry. It is not possible to analyse these objects as they are all accessioned to the museum and cannot be taken out of the country for analysis. The main source of information is from surface visual inspection and comparison with other objects that have been analysed such as beads and bangles. Many of the button seals have a core that is yellowish white with little trace of colourant. The surfaces have faint traces of blue-green colour and rarely traces of glassy external glaze (Fig. 29a–e). These indications suggest that the process of glazing may have been through the application process and not efflorescence as was used in beads and bangles. The same interpretation applies to some moulded faience tablets that have traces of blue-green glaze on the exterior but have very little evidence for colour in the core (Fig. 29f). There are some exceptions to this pattern however, and it is possible that some of the tablets were made with compact faience and efflorescence technique. Although most faience seals and tablets were made with only one colour of faience, there are some examples of tablets made with two colours, primarily grey-black with yellowish white (Fig. 29g, h).

86

J. M. Kenoyer

Fig. 29 Faience glazed seals and tablets, Harappa Phase, Period 3 (a Faience button seal, double stepped cross motif, Harappa, H97-3373; b Faience button seal, single stepped cross motif, Harappa, NMP 52.3015; c Faience button seal, swastika motif, Harappa, H99-3814; d Faience button seal, circular with cross motif; e Faience moulded tablet with blue-green glaze; f Faience moulded tablet with blue-green glaze, H98-3443; g Faience moulded tablet with two colours, H2001-5090; h Faience moulded tablet with two colours, H2001-5082)

Fig. 30 Period 3, Faience vessels, a H90/3041-3, b CAL73, c CAL33, d CAL13, e CAL58, f H95/5152-1, g Faience surma container, H98/8156-28

Glazed Steatite and Faience Technology at Harappa, Pakistan …

87

3.14 Period 3 Faience Vessels Faience vessels manufacture required craftspeople to develop innovative new techniques to build shapes and join them together to create small to miniature vessels (Fig. 30). The vessel shapes include small globular bottles, long narrow bottles and some very large vessels for which only the lids have been recovered (Fig. 30b). Some faience vessels were made in one colour, primarily blue-green, but others have applied designs made from other colours such as white, reddish orange and deep cobalt blue. The composition of the faience for vessels that have been analysed reveals the use of compact glassy faience, and the process of manufacture involves building the vessel around a core made of some soft material such as cotton wrapped in a piece of textile. The impression of a woven textile is often found on the interior surface of many vessel fragments. Vessels were also often made in two segments that were then joined together using a strip of faience to reinforce the join on the inside, while the outside was smoothed to remove all traces of the join. The decoration of the exterior surface was done primarily by a process of application of lines of colour using thin coils or strips of plastic faience (Fig. 30a, e–f). These designs appear to have been added, while the faience was still plastic as the colour is deeply embedded into the body of the vessel. In one rare example, three different colours of faience, white, reddish-brown and grey, were moulded together to simulate a rock with many large crystals like porphyry. After drying, the rims and bases as well as some exterior surfaces were trimmed and carved to perfect the shape before firing and glazing.

3.15 Period 3 Faience Figurines The production of hand-modelled and moulded faience figurines is also a new development that becomes quite elaborate during Period 3 at Harappa (Fig. 31a, b). Many miniature figurines, such as seated rams, monkeys and squirrels, were produced at the sites of Harappa and Mohenjo-daro. Some of these objects may have been made with moulds while others appear to have been carved to accentuate specific features. Based on the fine compact structure of the cores and the presence of some blue-green colour throughout the core and exterior, some faience figurines appear to have been made using the efflorescence technique. It is possible that application was also used, but this needs to be confirmed with more detailed analysis. The types of moulded figurines most commonly produced in faience are resting ram figurines, some of which have a hole that would be used to wear them as a pendant or bead (Kenoyer, 1998, Fig. 6.34). No moulds have been found yet, and it is possible that the moulds were made from wood. More complex forms include seated monkeys that may have been moulded first and then the details were carved to accentuate the eyes, mouth and various body parts (Kenoyer, 1998, Figs. 7.13, 8.23). The most complex form of faience figurine was made with composite pieces formed separately and then joined to produce a full image. The horn of what would have been a bull figurine has

88

J. M. Kenoyer

Fig. 31 Period 3, faience figurines, a Faience horn, CAL, b Faience leg and hoof., H20004440/2121-90

been analysed and indicates the use of compact faience and efflorescence technique (Fig. 31a). Another unique example is the leg of a bull figurine discovered near the faience workshop described above (Fig. 31b). This leg has a hoof made with white faience and the rest of the leg made from a darker grey brown faience. The shape may have been moulded but the details of the cloven hoof were carved when the faience was partly dry. Samples of all the different types of faience objects discussed above have been analysed to determine the overall proportions of unmelted quartz, glass and voids to better understand the range of variation found in the production of faience objects. Cross sections of objects were made and polished to produce flat sections that could be studied under the SEM. For standardized comparison, the core of each object was documented to determine the three main variables. Ternary plots of the different groups of objects shows that there is a wide range of compactness for bangles (Fig. 32a–c). For one example, CAL88 (Fig. 32a), the exterior glazed area (CAL88a) was compared with the glassy area near the edge of the bangle (CAL88b) and then with two different areas of the core (CAL88c and CAL88d). This one example shows

Glazed Steatite and Faience Technology at Harappa, Pakistan … Fig. 32 Faience Ternary plots, a Kot Diji and Harappa Phase faience bangles, b Ravi and Harappa Phase faience beads, c Harappa Phase vessels and figurines

89

90

J. M. Kenoyer

Fig. 33 Late Harappa geometric seal, glazed steatite, H98 3493/8314-01

that the exterior has the highest glassy portion but even different parts of the core are relatively similar (within 10%) in terms of percentage of quartz, glass and voids. Overall, most bangles have more than 60% glassy matrix, between 30 and 50% quartz, and between 10 and 40% voids. The blue dot CAL55 represents the Kot Diji Phase bangle and it falls within the overall range of the later Harappa Phase bangles. On the plot of faience beads, the grouping of the beads is a bit tighter (Fig. 32b), with less variation in terms of glass and quartz percentages, but the voids range from 10 to 50%. The Ravi Phase faience bead (blue dot 200) fits right in the middle of the cluster and is clearly part of a pattern that continues in the later period. Faience vessels and figurines also are very close to the centre and show very little variation (Fig. 32c).

3.16 Period 4/5 Late Harappa Steatite and Faience Although it is beyond the scope of this paper, it is important to note that the Late Harappa Phase sees a strong continuity in both steatite and faience production with some new features that indicate a robust craft tradition. While in the past scholars have argued that the Late Harappan period is one of decline, evidence from Harappa shows exactly the opposite. The Late Harappan period is one of transformation and new innovation in pottery making, faience production and other aspects of urban life (Kenoyer, 2005b). In steatite seal making, fine white glazed steatite seals continued to be produced but they no longer had unicorn motifs or Indus script (Fig. 33). New styles of glazed steatite beads and glazed faience beads were also produced (Fig. 34). Faience beads with carved designs and with two or more colours were produced to simulate stone beads as well as to create new styles that had never been seen before. Beads from the site of Sanauli (Sharma et al., 2004) provide one of

Glazed Steatite and Faience Technology at Harappa, Pakistan …

91

Fig. 34 Late Harappa faience beads, H96 7330

the most important examples of the new styles of faience during the Late Harappan period, and the future analysis of these beads will provide an important window on the elaboration of faience technology during this important transitional time period. So far there is no evidence for the production of glass during the Harappan or Late Harappan periods. A red glass like bead found in a Late Harappan context was at first thought to be a form of early glass (Kenoyer, 2005a: 167), but detailed scientific analysis by Law later demonstrated that this bead was in fact a natural form of red kaolin, a rock and not glass (Law, 2011b: 556, Appendix 4.6). Nevertheless, it is still important to consider that compact glassy faience and the elaboration of faience technology during the Late Harappan period may have been an important transitional technology that eventually led to the development of an indigenous glass technology in the Painted Grey Ware Period in South Asia (Singh, 1989; Basa, 1992; Kanungo, 2010).

4 Conclusion In this brief overview, it has been shown that the development of glazing technology involving steatite and faience has a long history in South Asia beginning around 5500BCE for glazed steatite at Mehrgarh and for faience by around 4000–3700 BCE at Mehrgarh, Nausharo and Harappa. The styles of objects and developments in technology reflect an autochthonous process that does not seem to have been

92

J. M. Kenoyer

influenced or in any way derived from parallel traditions that were developing in West Asia (Mesopotamia and the Levant) and North Africa (Egypt) at approximately the same time. Among the unique aspects of Indus glazed steatite and faience production is the manufacture of microbeads as well as beads of different colours that imitated stone beads. The Indus craftspeople may have used different techniques for glazing depending on the most optimal methods of production and the final form of an object. The efflorescence technique appears to have been used for most objects, but there is some indication that glaze was applied to the exterior of frit bodies, and in the case of vessel decoration, lines of different coloured faience were applied to the exterior of the vessels. There is also evidence that cementation may have been used in the production of large quantities of glazed steatite beads, but so far, no evidence for cementation in the production of faience objects. The Indus Tradition craftspeople developed a technique of making coloured frit that was reground to produce compact glassy faience specifically to make strong vitreous bangles and rings, non-porous vessels, delicate figurines and different types of beads. Further excavations and research on faience ornaments and manufacturing waste from sites in the Indo-Gangetic region will undoubtedly turn up new evidence that shows how this technology continued to develop during the Late Harappan and subsequent Early Historic Periods. One of the most important aspects of glazed steatite and faience research is the analysis of archaeological objects as well as experimental reconstructions of different aspects of faience production. Scholars throughout South Asia are encouraged to submit their broken steatite and faience objects for proper scientific analysis that requires preparing thin sections that can be studied under the SEM–EDS and by glaze analysis using LA-ICP-MS and other techniques. The use of non-destructive techniques can be applied to complete objects that cannot be cut and polished, but these techniques do not provide the detailed information needed to test specific hypotheses regarding faience composition or production technology. Regional surveys of steatite as well as quartz sands and sources of good quality quartz rock must also be undertaken to develop a better understanding of where the craftspeople acquired the basic raw material needed for making steatite and faience objects. The other area that requires creative experimental approaches is the manufacture and firing of steatite and faience objects. It is unlikely that there was only one technique for making faience objects so experiments are needed to see how different techniques can be used to create similar objects. In my ongoing research, I am engaged in experiments to make frit, and experiments for regrinding using tools that were available to the Harappans are also being undertaken. I am experimenting with different recipes needed to make self-glazing pastes to determine which combinations produce results that are similar to those seen on ancient objects. Finally, the firing of faience objects is being studied through use of modern muffle furnaces as well as wood-fired furnaces and kilns. The more people who can be involved in this research the better. At the workshop in IIT Gandhinagar, Dr. Massimo Vidale and I were able to demonstrate how to undertake an experimental project using local sands from the Sabarmati River. In our experiment, it was clear that these sands do not produce the type of faience that is seen in the Indus sites, but more research is needed to see if

Glazed Steatite and Faience Technology at Harappa, Pakistan …

93

these sands could be processed in a manner that would allow them to be used to make faience. We also used pure rock crystal brought specially from Madhya Pradesh to try and produce faience. This too was not very successful. Even the firing of the faience was not completely successful, but through the whole process, we learned a great deal about firing technology and processes of manufacture that we can now combine with other knowledge to build a better understanding of these ancient technologies. Carrying out a workshop for everyone to participate in was an important way to share knowledge and experience, and hopefully, the conference participants have been able to continue with this type of work at their own institutions. Finally, it is important to develop ways to show how the study of ancient technologies has potential applications to present day crafts. No one in South Asia presently is involved in the glazing of steatite, and yet steatite carving is widely carried out throughout both India and Pakistan. Faience technology is also something that has died out in both regions, though some frit-based glazed pottery traditions are still carried out in some areas. The production of fired steatite and faience in Egypt is an important craft industry that provides income to many communities who produce replicas for sale to tourists at museums and other public markets. Future publications will include more detailed discussions of the recipes and techniques used in my experiments that could be used by anyone interested in developing these crafts. Acknowledgements First of all, I would like to thank Dr. Alok K. Kanungo for organizing the workshop at IIT Gandhinagar and Director, Prof. Sudhir K. Jain, for his support and encouragement over the past many years. My research on Indus Faience began under the guidance of my advisor, the late Dr. G.F. Dales, and I wish to acknowledge his support and encouragement in all of my various studies of Indus crafts. Most of the data presented in this paper derive from the Harappa Archaeological Research Project (HARP) excavations at Harappa, and I would especially like to thank the Government of Pakistan, Department of Archaeology, for facilitating our continued work at Harappa, and my co-directors Dr. Richard H. Meadow and Dr. Rita P. Wright as well as my many students for their ongoing support. Over the years I have worked with many people to better understand glazed ceramics and specifically want to thank Dr. Pamela Vandiver for her guidance and training and the Smithsonian Institution for supporting my study with her in 1993. I also want to specially thank my colleagues at Mohenjo-daro, Dr. Massimo Vidale, and at Harappa, Dr. Richard Meadow, Dr. Rita P. Wright, Donna Strahan, Harriet Beaubien, Dr. Heather M-L Miller, Dr. Randall Law, and all the HARP staff (Tazeem ul Hassan and Ghulam Husain) and workers who helped with the excavations and conservation of materials. I also want to thank Dr. Qasid Mallah for allowing me to study the faience and steatite from his recent excavations at the site of Lakhan-jo-daro, Sindh. In the analysis of the faience, I want to specially thank Dr. Laure Dussubieux for the LA-ICP-MS analysis at the Field Museum and Dr. John Fournelle and Bil Schneider for their help with the SEMEDS, Department of Geosciences, UW-Madison. I also want to thank many colleagues and students who have also been working on faience from other sites and with whom I have had the opportunity to exchange ideas and discuss various theories. Special thanks to Dr. Kuldeep K. Bhan, Dr. K. Krishnan, Dr. P. Ajithprasad and others at Maharaja Sayajirao University of Baroda. Also thanks to Dr. V.N. Prabhakar for allowing me to look at some of the materials from sites that he had surveyed, such as Mitathal. Additional thanks to Dr. Akinori Uesugi and Dr. Ayumu Konosukawa for sharing their information on sites that they have been working at. I also want to thank all of the other many scholars who have allowed me to study beads from their excavations in India, Pakistan, Oman, China and other countries and look forward to presenting these results in future collaborative papers. My ongoing research at Harappa and the Indus Valley Civilization in general has been supported by numerous organizations, namely the National Science Foundation, the National Endowment for the Humanities, the National Geographic Society, the Smithsonian Institution, the American School of

94

J. M. Kenoyer

Prehistoric Research (Peabody Museum of Archaeology and Ethnology, Harvard University), the University of Wisconsin, www.HARAPPA.com, Global Heritage Fund, educational grants from the US State Department, US Embassy Islamabad.

References Ahmed, Z., Faisal, M., Abbasi, T., Malik, H., Rasool, W., Shah, S. A., & Rafique, F. (2018). Geological Field Report on Kirana Hills, Sargodha, Pakistan. Department of Earth and Environment Sciences, Bahria University. Allchin, F. R., Allchin, B., Durrani, F. A., & Khan, M. F. (Eds.). (1986). Lewan and the Bannu Basin. British Archaeological Reports. Allchin, F. R., & Knox, J. R. (1981). Preliminary report on the excavation at Tarakai Qila. In H. Härtel (Ed.), South Asian Archaeology, 1979 (pp. 245–250). Dietrich Reimer. Barthélemy De Saizieu, B. (2003). Les Parures de Mehrgarh: Perles et Pendentifs du Néolithique Préceramique à La Période Pré-Indus: Fouilles 1974–1985. Editions Recherche sur les Civilisations. Barthélemy De Saizieu, B., & Bouquillon, A. (1994). Steatite working at Mehrgarh during the neolithic and chalcolithic periods: quantitative distribution, characterization of material and manufacturing process. In A. Parpola & P. Koskikallio (Eds.), South Asian Archaeology, 1993 (pp. 47–60). Suomalainen Tiedeakatemia. Barthélemy De Saizieu, B., & Bouquillon, A. (2000). Émergence et Évolution des Matériaux Vitrifiés dans la Région de l’Indus du 5e au 3e Millénaire (Merhgarh-Nausharo). Paléorient, 26, 93–111. Basa, K. K. (1992). Early glass beads in India. South Asian Studies, 8, 91–104. Beck, H. (1940). Chapter XV. Part 1. Report on Selected Beads from Harappa. In M. S. Vats (Ed.) Excavations at Harappa: Being an account of archaeological excavations at harappa carried out between the years of 1920–21 and 1933–34 (pp. 392–431). Government of India Press. Beck, H. C. (1928). Classification and nomenclature of beads and pendants. Archaeologia, 77. Beck, H. C. (1934). The use of the microscope in the study of ancient beads. Journal of the Royal Microscopical Society LIV, 186–194. Belcher, W. R. (1998). Fish exploitation of the Baluchistan and Indus valley tradition: An ethnoarchaeological approach to the study of fish remains. PhD Thesis. University of Wisconsin-Madison. Belcher, W. R. (2000). Marine subsistence of the Indus valley tradition: Fish remains from Balakot and Allahdino (Pakistan). In M. Taddei & G. De Marco (Eds.), South Asian Archaeology 1997 (pp. 3–16). Is. I. A. O. Bhan, K. K., Sonawane, V. H., Ajithprasad, P., & Prathapchandran, S. (2005). A Harappan trading and craft production centre at Gola Dhoro (Bagasra). Antiquity, 79, 1–7. Bouquillon, A., & Barthélémy De Saizieu, B. (1994). Steatite working at Mehrgarh during the Neolithic and chalcolithic periods: quantitative distribution, characterization of material and manufacturing processes. In A. Parpola & P. Koskikallio (Eds.), South Asian Archaeology 1993 (pp. 47–59). Suomalainen Tiedeakatemia. Bouquillon, A., Barthélémy De Saizieu, B. (1995). Découverte d’un Nouveau Matériau dans les Parures de la Période Pré-Indus de Mehrgarh (Balochistan): La “Faïence” de Stéatite, Technè 2, Autoportrait d’un Laboratoire, Le Laboratoire de Recherche des Musées de France, 50–55. Bouquillon, A., Barthélémy De Saizieu, B., & Duval, A. (1995). Glazed steatite beads from Mehrgarh and Nausharo (Pakistani Balochistan). Materials Research Society Symposium Proceedings, 352, 527–538. Chase, B., Ajithprasad, P., Rajesh, S. V., Patel, A., & Sharma, B. (2014). Materializing Harappan identities: Unity and diversity in the borderlands of the Indus civilization. Journal of Anthropological Archaeology, 35, 63–78.

Glazed Steatite and Faience Technology at Harappa, Pakistan …

95

Chase, B., Ajithprasad, P., & Sasidharan Vasantha, R. (2016). The Identification of diversity: Material cultures and social practice in Harappan Gujarat. In V. Widorn, U. Franke, & P. Latschenberger (Eds.), South Asian archaeology and art: Contextualizing material culture in south and central Asia in pre-modern times (pp. 111–124). Brepols Publisher. Clark, S. R. (2007). The social lives of figurines: Recontextualizing the third millennium BC terracotta figurines from Harappa (Pakistan). PhD Thesis. Harvard University. Clark, S. R., & Kenoyer, J. M. (2017). South Asia – Indus Civilization. In T. Insoll (Ed.), The Oxford handbook of prehistoric Figurines (pp. 493–519). Oxford University Press. Dales, G. F. (1991). Some Specialized Ceramic Studies at Harappa. In R. H. Meadow (Ed.), Harappa excavations 1986–1990 (pp. 61–70). Prehistory Press. Dales, G. F., & Kenoyer, J. M. (1989). Excavation at Harappa—1988. Pakistan Archaeology, 24, 68–176. Dales, G. F., & Kenoyer, J. M. (1991). Summaries of five seasons of research at Harappa (District Sahiwal, Punjab, Pakistan) 1986–1990. In R. H. Meadow (Ed.), Harappa excavations 1986–1990 (pp. 185–262). Prehistory Press. Dales, G. F., & Kenoyer, J. M. (1993). The Harappa Project 1986–1989: New investigation at an Ancient Indus City. In G. L. Possehl (Ed.), Harappan Civilization: A Recent Perspective (pp. 469–520). Oxford & IBH Co., Pvt. Ltd. Dani, A. H. (1970–71). Excavations in the Gomal Valley, Ancient Pakistan 5, 1–177. Davis, M. (2016). Stone tools as indicators of socio-economic variability at Harappa, Pakistan. In V. Widorn, U. Franke, & P. Latschenberger (Eds.), South Asian Archaeology and Art: contextualizing material culture in South and Central Asia in pre-modern times (pp. 51–66). Brepols Publisher. Dhavalikar, M. K. (1993). Harappans in Saurashtra: The mercantile enterprise as seen from recent excavation of Kuntasi. In G. L. Possehl (Ed.), Harappan civilization: A recent perspective (pp. 555–568). Oxford & IBH Co., Pvt. Ltd. Dhavalikar, M. K., Raval, M. R., & Chitalwala, Y. M. (1996). Kuntasi: A Harappan emporium on west coast. Deccan College Post-Graduate & Research Institute. Didier, A. (2017). Nouvelles Recherches sur les Débuts de la Civilisation De l’Indus au Pakistan. Les Fouilles De Chanhu-Daro (Sindh). Par Mme Aurore Didier, Académie Des Inscriptions & Belles-Lettres, 2, 969–1002. Durrani, F. A., Ali, I., & Erdosy, G. (1995a). The Beads of Rehman Dheri. Ancient Pakistan, 10, 15–81. Durrani, F. A., Ali, I., & Erdosy, G. (1995b). Seals and Inscribed Sherds of Rehman Dheri. Ancient Pakistan, 10, 198–233. Dussubieux, L., Robertshaw, P., & Glascock, M. D. (2009). LA-ICP-MS analysis of African glass beads: Laboratory inter-comparison with an emphasis on the impact of corrosion on data interpretation. International Journal of Mass Spectrometry, 284, 152–161. Good, I., Kenoyer, J. M., & Meadow, R. H. (2009). New evidence for early silk in the Indus civilization. Archaeometry, 51, 457–466. Good, I., Kenoyer, J. M., & Meadow, R. H. (2011). A Reply to Ji-Huan He. Archaeometry, 53, 1257–1258. Halfpenny, R., & Mazzucchelli, R. H. (1999). Regional multi-element drainage geochemistry in the Himalayan Mountains, Northern Pakistan. Journal of Geochemical Exploration, 67, 223–233. Hegde, K. T. M., Karanth, R. V., & Sychanthavong, S. P. (1982). On the composition and technology of Harappan microbeads. In G. L. Possehl (Ed.), Harappan civilization (pp. 239–244). Oxford and IBH Publishing Co. Hoffman, B. C. (2019). Indus tradition copper and bronze metallurgy: A model from Harappa. PhD Thesis. University of Wisconsin-Madison. Hoffman, B. C., & Miller, H.M.-L. (2009). Production and consumption of copper-base metals in the Indus civilization. Journal of World Prehistory, 22, 237–264. Jamison, G. M. (2017). The organization of Indus unicorn seal production: A diachronic comparative study of style, skill, and sociopolitical organization. PhD Thesis. University of Wisconsin-Madison.

96

J. M. Kenoyer

Jarrige, C. (2008). The figurines of the first farmers at Mehrgarh and Their Offshoots. Pragdhara, 18, 155–166. Jarrige, C., Jarrige, J.-F., Meadow, R. H., Quivron, G. (Eds.), (1995). Mehrgarh field reports 1975 to 1985—from the Neolithic to the Indus Civilization. Dept. of Culture and Tourism, Govt. of Sindh and the French Foreign Ministry. Joshi, J. P., & Parpola, A. (1987). Corpus of Indus seals and inscriptions. Collections in India. Suomalainen Tiedeakatemia. Kanungo, A. K., & Trivedi, M. (2019). Meeting report: Ancient Indian glass. Current Science, 117(3), 355–357. Kanungo, A. K. (2010). Antiquity of Glass in India: Excavations at Kopia. In V. Tripathi & P. Upadhyay (Eds.), Archaeology of the Ganga Basin: Paradigm Shift (pp. 451–476). Sharada Publishing House. Kenoyer, J. M. (1989). Socio-economic structures of the Indus civilization as reflected in specialized crafts and the question of ritual segregation, In J. M. Kenoyer (Ed.),Old problems and new perspectives in the archaeology of South Asia (pp. 183–192). Dept. of Anthropology, U.W.Madison. Kenoyer, J. M. (1992a). Harappan craft specialization and the question of Urban segregation and stratification. Eastern Anthropologist, 45, 39–54. Kenoyer, J. M. (1992b). Ornament styles of the Indus Tradition: Evidence from recent excavations at Harappa. Pakistan, Paléorient, 17, 79–98. Kenoyer, J. M. (1993). Excavations on Mound E, Harappa: a systematic approach to the study of Indus Urbanism. In A. J. Gail & G. J. R. Mevissen (Eds.), South Asian Archaeology, 1991 (pp. 165–194). F.S. Verlag. Kenoyer, J. M. (1994a). Experimental studies of Indus valley technology at Harappa. In A. Parpola & P. Koskikallio (Eds.), South Asian Archaeology, 1993 (pp. 245–262). Suomalainen Tiedeakatemia. Kenoyer, J. M. (1994). Faience ornaments of Harappa and the Indus civilization. Ornament, 17(35– 39), 95. Kenoyer, J. M. (1997). Trade and technology of the Indus Valley: New insights from Harappa. Pakistan, World Archaeology, 29, 262–280. Kenoyer, J. M. (1998). Ancient cities of the Indus Valley civilization. Oxford University Press. Kenoyer, J. M. (2000). Wealth and socio-economic hierarchies of the Indus Valley civilization. In J. Richards & M. Van Buren (Eds.), Order, legitimacy and Wealth in early states (pp. 90–112). Cambridge: Cambridge University Press. Kenoyer, J. M. (2005a). Bead Technologies at Harappa, 3300–1900 BC: A Comparison of Tools, Techniques and Finished Beads from the Ravi to the Late Harappan Period. In C. Jarrige & V. Lefèvre (Eds.), South Asian archaeology, 2001 (pp. 157–170). CNRS. Kenoyer, J. M. (2005b). Culture change during the late harappan period at Harappa: New insights on Vedic Aryan issues. In E. F. Bryant & L. L. Patton (Eds.), The Aryan Invasion: Evidence, Politics, History (pp. 21–49). Routledge. Kenoyer, J. M. (2005). Steatite and Faience manufacturing at Harappa: new evidence from mound E excavations 2000–2001. Museum Journal: National Museum of Pakistan, Karachi, 3–4, 43–56. Kenoyer, J. M. (2006). The origin, context and function of the Indus script: Recent insights from Harappa. In T. Osada & N. Hase (Eds.), Proceedings of the pre-symposium and the 7th Esca Harvard-Kyoto roundtable (pp. 9–27). Research Institute for Humanity and Nature, RIHN. Kenoyer, J. M. (2012). Households and neighborhoods of the Indus tradition: An overview. In B. J. Parker & C. P. Foster (Eds.), New perspectives in Household archaeology (pp. 373–406). Eisenbrauns. Kenoyer, J. M. (2017a). Stone beads of the Indus Tradition: New perspectives on Harappan bead typology, technology and documentation. In A. K. Kanungo (Ed.), Stone beads of South and Southeast Asia: Archaeology, ethnography and global connections (pp. 149–164). Indian Institute of Technology Gandhinagar / Aryan Books International.

Glazed Steatite and Faience Technology at Harappa, Pakistan …

97

Kenoyer, J. M. (2017b). Textiles and trade in south asia during the proto-historic and early historic period. In B. Hildebrandt (Ed.), Silk, trade and exchange along the silk roads between Rome and China in Antiquity (pp. 9–33). Oxbow Books. Kenoyer, J. M. (2017c). Using SEM to study stone bead technology. In A. K. Kanungo (Ed.), Stone beads of south and Southeast Asia: Archaeology, ethnography and global connections (pp. 405–433). Indian Institute of Technology Gandhinagar/Aryan Books International. Kenoyer, J. M., & Meadow, R. H. (2000). The ravi phase: A new cultural manifestation at Harappa. In M. Taddei & G. De Marco (Eds.), South Asian Archaeology, 1997 (pp. 55–76). Istituto Italiano per l’Africa e l’Oriente/Istituto Universitario Orientale. Kenoyer, J. M., & Meadow, R. H. (2016). Excavations at Harappa: 1986–2010: New insights on the Indus civilization and Harappan burial traditions. In G. R. Schug & S. R. Walimbe (Eds.), The companion to the South Asian Past (pp. 145–168). Wiley. Kenoyer, J. M., & Miller, H.M.-L. (1999). Metal technologies of the Indus Valley tradition in Pakistan and Western India. In V. C. Pigott (Ed.), The emergence and development of metallurgy (pp. 107–151). University Museum. Kenoyer, J. M., & Miller, H.M.-L. (2007). Multiple crafts and socio-economic associations in the Indus Civilization: perspectives on continuity and change from Harappa, Pakistan. In I. Shimada (Ed.), Rethinking craft production: The nature of producers and multi-craft organization (pp. 152– 183). University of Utah Press. Khatri, J. S., & M. Acharya (1997). Kunal: The Earliest Pre-Harappan Settlement. In J. P. Joshi (Ed.), Facets of Indian civilization recent perspectives-essays in Honor of B.B. Lal (pp. 87–91). Aryan Books International. Khatri, J. S., & Acharya, M. (2005). Kunal excavations: new light on the origin of the Harappan Civilization. In A. Agrawal (Ed.), In Search of Vedic Harappan Relationship (pp. 104–117). Aryan Books International. Konosukawa, A. (2013). Diversity of Harappan civilization: A case study of the Ghaggar Basin (with Special Reference to Seals). PhD Thesis. Deccan College Post-Graduate & Research Institute. Law, R. W. (2002). Potential steatite sources for the Indus civilization. In M. A. Halim (Ed.), Indus valley civilization: Collection of papers presented in the international colloquium on Indus Valley civilization at Islamabad (6th–8th April 2001) (pp. 158–69). Ministry of Minorities, Culture, Sports, Tourism, & and Youth Affairs, Government of Pakistan. Law, R. W. (2006). Moving mountains: The trade and transport of rocks and minerals within the greater Indus Valley Region. In E. C. Robertson, J. D. Seibert, D. C. Fernandez, & M. U. Zender (Eds.), Spatial and spatial analysis in archaeology (pp. 301–313). University of Calgary Press. Law, R. W. (2011a). Inter-regional Interaction and Urbanism in the Ancient Indus Valley: A geologic provenience study of Harappa’s rock and mineral assemblage. Manohar. Law, R. W. (2011b). Inter-regional interaction and Urbanism in the Ancient Indus Valley: A geologic provenience study of Harappa’s rock and mineral assemblage, Occasional Paper 11. Kyoto: Research Institute for Humanity and Nature. Law, R. W. (2013a). The important stone and metal resources of Gujarat during the Harappan period. Heritage: Journal of Multidisciplinary Studies in Archaeology, 1, 319–343. Law, R. W. (2013b). Initial geologic provenience studies of stone and metal artefacts from Rakhigarhi, Heritage: Journal of Multidisciplinary Studies in Archaeology, 2, 74–100. Law, R. W. (2014). Evaluating potential lapis lazuli sources for ancient South Asia using Sulfur Isotope analysis. In C. C. Lamberg-Karlovsky, B. Genito, & B. Cerasetti (Eds.), My Life is Like the Summer Rose: Maurizio Tosi E L’archeologia Come Modo Di Vita (pp. 419–429). Arcaeopress. Law, R. W. (2018). The art of the Harappan microbead—revisited. In D. Frenez, G. Jamison, R. Law, M. Vidale, & R. H. Meadow (Eds.), Walking with the Unicorn: Social organization and material culture in Ancient South Asia. Jonathan Mark Kenoyer Felicitation Volume (pp. 327– 342). Archaeopress. Law, R. W., Carter, A., Bhan, K., Malik, A., & Glascock, M. D. (2013). INAA of agate sources and artifacts from the Indus, Helmand, and Thailand regions. In D. Frenez & M. Tosi (Eds.), South Asian archaeology, 2007 (pp. 177–184). Archaeopress.

98

J. M. Kenoyer

Ludvik, G. E. (2018). Hard stone beads and socio-political interaction in the intermediate bronze age southern Levant, ca. 2500–2000 BCE. PhD Thesis. University of Wisconsin-Madison. Mackay, E. J. H. (1931). Ch. XXVIII. Ivory, shell, faience, and other objects of technical interest. In J. H. Marshall (Ed.), Mohenjo-Daro and the Indus civilization: Being an official account of archaeological excavations at Mohenjo-Daro carried out by the government of India between the Years 1922 and 1927 (pp. 562–588) A. Probsthain. Mackay, E. J. H. (1937). Bead making in ancient Sind. Journal of the American Oriental Society, 57, 1–15. Mackay, E. J. H. (1938). Excavations at Chanhudaro by the American school of Indic and Iranian studies and the Museum of fine arts. Boston, Smithsonian Report , 1937, 469–478. Mackay, E. J. H. (1938b). Further excavations at Mohenjo-daro: Being an official account of archaeological excavations at Mohenjo-Daro carried out by the Government of India between the years 1927 and 1931. Government of India. Mackay, E. J. H. (1943). Chanhudaro excavations 1935–36. American Oriental Society. Mallah, Q., & Rajput Shafiq, T. (2016). Exceptional objects from Lakhan-Jo-Daro. Frontier Archaeology, 10–14, 196–229. Mallah, Q. H. (2008). Recent archaeological discoveries in Sindh, Pakistan. In T. Osada & A. Uesugi (Eds.), Linguistics, archaeology and the human past, occasional paper 3 (pp. 27–76). RIHN. Marshall, J. H. (1931). Mohenjo-Daro and the Indus civilization: Being an official account of archaeological excavations at Mohenjo-Daro Carried out by the Government of India between the Years 1922 and 1927. A. Probsthain. McCarthy, B., & Vandiver, P. B. (1990). Ancient high-strength ceramics: Fritted faience bangle manufacture at Harappa (Pakistan), ca. 2300–1800 B.C. In P. B. Vandiver, J. Druzik & G. S. Wheeler (Eds.), Materials issues in art and archaeology (Vol. 2., pp. 495–510). Materials Research Society. Meadow, R. H., & Kenoyer, J. M. (2001). recent discoveries and highlights from excavations at Harappa: 1998–2000. Indo Koko Kenkyu: Indian Archaeological Studies, 22, 19–36. Meadow, R. H., & Kenoyer, J. M. (2005). Excavations at Harappa 2000–2001: New Insights on Chronology and City Organization. In C. Jarrige & V. Lefèvre (Eds.), South Asian Archaeology, 2001 (pp. 207–225). Editions Recherche sur les Civilisations - ADPF. Meadow, R. H., & Kenoyer, J. M. (2008). Harappa excavations 1998–1999: New evidence for the development and manifestation of the Harappan Phenomenon. In E. M. Raven (Ed.), South Asian archaeology, 1999 (pp. 85–109). International Institute of Asian Studies. Meadow, R. H., & Patel, A. K. (2003). Prehistoric Pastoralism in Northwestern South Asia from the Neolithic through the Harappan Period. In S. Weber & W. R. Belcher (Eds.), Ethnobiology and the Indus Civilization (pp. 65–94). Lexington Books. Miller, H. M.-L. (1994). Metal processing at harappa and mohenjo-daro: Information from nonmetal remains. In A. Parpola & P. Koskikallio (Eds.), South Asian Archaeology, 1993 (pp. 497– 510). Suomalainen Tiedeakatemia. Miller, H. M.-L. (1996). The distribution of high-temperature manufacturing at Harappa, Pakistan (3rd Mill. BCE). Society for American Archaeology. Miller, H. M.-L. (1997a). Locating ancient manufacturing areas: High-temperature manufacturing debris from surface surveys at Harappa, Pakistan. In B. Allchin (Ed.), South Asian archaeology, 1995 (pp. 939–946). Oxford & IBH. Miller, H. M.-L. (1997b). Pottery firing structures (Kilns) of the Indus civilization during the third millennium B.C. In P. Rice & W.D. Kingery (Eds.) Prehistory and history of ceramic kilns (pp. 41–71). American Ceramic Society. Miller, H. M.-L. (1999). Pyrotechnology and society in the cities of the Indus Valley. PhD Thesis. University of Wisconsin-Madison Microfilms. Miller, H. M.-L. (2008a). The Indus talc-faience complex: types of materials, clues to production. In E. M. Raven (Ed.), South Asian archaeology, 1999 (pp. 111–122). International Institute of Asian Studies.

Glazed Steatite and Faience Technology at Harappa, Pakistan …

99

Miller, H. M.-L. (2008b). Issues in the determination of ancient value systems: The role of talc (Steatite) and Faience in the Indus civilization. In E. Olojdam & R. H. Spoor (Eds.), Intercultural relations between South and Southwest Asia: Studies in commemoration of E.C.L. during caspers (1934–1996) (pp. 145–157) BAR International Series 1826. Miller, H. M.-L., Bouquillon, A. & Barthélémy B. (1996). White-Coated ‘Dishes” from Harappa, Pakistan: A refractory puzzle from the mid-third millennium BC. In International symposium on archaeometry, 1996. University of Illinois at Urbana-Champaign. Miller H. M.-L., & Kenoyer, J. M. (2018). Invisible value or tactile value? Steatite in the faience complexes of the Indus valley tradition. In: D. Frenez, G. Jamison, R. Law, M. Vidale, & R. H. Meadow (Eds.), Walking with the Unicorn: Social organization and material culture in Ancient South Asia. Jonathan Mark Kenoyer Felicitation Volume (pp. 389–394). Archaeopress. Miller, L. J. (2003). Secondary products and urbanism in South Asia: The evidence for traction at Harappa. In S. Weber & W. R. Belcher (Eds.), Ethnobiology and the Indus civilization (pp. 251– 326). Lexington Books. Mughal, M. R. (1968). Harappa-1966 (Cemetery R 37). Pakistan Archaeology, 5, 63–68. Nardi, I. (2006). The theory of citrasutras in Indian painting: A critical re-evaluation of their uses and interpretations. Routledge. Patel, A., & Ajithprasad, P. (2018). Analytical study of Harappan copper artifacts from gujarat with special reference to Bagasra. In D. Frenez, G. Jamison, R. Law, M. Vidale, & R. H. Meadow (Eds.), Walking with the Unicorn: Social organization and material culture in Ancient South Asia. Jonathan Mark Kenoyer felicitation volume (pp. 443–452). Archaeopress. Possehl, G. L. (1991). A short history of archaeological discovery at Harappa. In R. H. Meadow (Ed.), Harappa excavations 1986–1990 (pp. 5–12). Prehistory Press. Ringdalen, E., & Tangstad, M. (2016). Softening and melting of SiO2 , an important parameter for reactions with quartz in Si production. In R. G. Reddy, P. Chaubal, P. C. Pistorius, & U. Pal (Eds.), Advances in molten slags, fluxes, and salts: Proceedings of the 10th international conference on molten slags, fluxes and salts (pp. 43–51). Springer. Sant, U., Baidya, T. J., Nikoshey, N. G., Sinha, N. K., Nayan, S., Tiwari, J. K., & Arif, A. (2005). Baror: A new Harappan site in Ghaggar Valley—a Preliminary report. Puratattva, 35, 50–59. Shah, S. G. M., & Parpola, A. (1991). Corpus of Indus seals and inscriptions. 2. collections in Pakistan. Suomalainen Tiedeakatemia. Shah, S. M. I. (1973). Occurrence of Gold in the Kirana group. Sargodha, Punjab, Pakistan, Geological Survey of Pakistan: Information Release, 68, 1–14. Sharma, D. V., Nauriyal, K. C., Prabhakar, V. N., & Vishnukant. (2004). Sanauli: A late Harappan burial site in the Yamuna-Hindon Doab. Puratattva, 34, 35–44. Singh, R. N. (1989). Ancient Indian glass: Archaeology and technology. Parimal Publications. Stoltman, J. (1991). Ceramic petrography as a technique for documenting cultural interaction: An example from the upper Mississippi Valley. American Antiquity, 56, 103–120. Thomas, K. D. (1983). Tarakai Qila: Site, economy and environment. In B. Proudfoot (Ed.), Site, Environment and economy (pp. 127–144). British Archaeological Reports. Tite, M. S., & Bimson, M. (1986). Faience: An investigation of the microstructures associated with the different methods of glazing. Archaeometry, 28, 69–78. Tite, M. S., Shortlan, A., Maniatis, Y., Kavoussanaki, D., & Harris, S. A. (2006). The composition of the soda-rich and mixed alkali plant ashes used in the production of glass. Journal of Archaeological Science, 33(9), 1284–1292. Tite, M. S., Shortland, A. J., & Vandiver, P. B. (2008). Raw materials and fabrication methods used in the production of faience. In M. S. Tite & A. J. Shortland (Eds.), Production technology of faience and related early vitreous materials, Monograph 72 (pp. 37–56). Oxford University School of Archaeology. Trivedi, P. K. (2009). Excavations at Tarkhanewala-Dera & Chak 86 (2003–2004). Archaeological Survey of India.

100

J. M. Kenoyer

Ullah, K. B. M. S. (1931). Appendix I. Notes and analyses. In J. H. Marshall (Ed.), Mohenjo-Daro and the Indus civilization: being an official account of archaeological excavations at mohenjodaro carried out by the government of India between the years 1922 and 1927. (pp. 686–690). A. Probsthain. Vandiver, P. (1982). Technological change in egyptian faience. In J. S. Olin & A. D. Franklin (Eds.), Archaeological Ceramics (pp. 167–179). Smithsonian Institution Press. Vandiver, P. (1983). The manufacture of faience, appendix A. In A. Kaczmarczyk & R. E. M. Hedges (Eds.), Ancient Egyptian faience (pp. A1–144). Aris and Phillips. Vats, M. S. (1940). Excavations at Harappa: Being an account of archaeological excavations at Harappa carried out between the years of 1920–21 and 1933–34. Government of India Press. Vidale, M. (1987). The paste plaques and cylinders of chanhudaro: A descriptive report. Annali, 47, 57–66. Vidale, M. (1989a). Early Harappan steatite, faience and paste beads in a Necklace from MehrgarhNausharo (Pakistan). East and West, 39, 291–300. Vidale, M. (1989b). A steatite-cutting atelier on the surface of moenjodaro. Annali, 49, 29–51. Vidale, M. (1995). Early beadmakers of the Indus tradition: The manufacturing sequence of talc beads at Mehrgarh in the 5th Millenium B.C. East and West, 45, 45–80. Vidale, M. (2000). The archaeology of Indus crafts: Indus craftspeople azd why we study them. Istituto Italiano per l"Africa e l’Oriente. Vidale, M., & Miller, H. (2000). On the development of Indus technical virtuosity and its relation to social structure. In M. Taddei & G. De Marco (Eds.), South Asian Archaeology, 1997 (pp. 115– 132). Istituto Italiano per l’Africa e l’Oriente. Weber, S. A., & Belcher, W. R. (Eds.). (2003). Ethnobiology and the Indus civilization. Lexington Books. Wheeler, R. E. M. (1947). Harappa 1946: The defenses and cemetery R-37. Ancient India, 3, 58–130. Wright, R. P. (1991). Patterns of technology and the organization of production at Harappa. In R. H. Meadow (Ed.), Harappa excavations 1986–1990 (pp. 71–88). Prehistory Press. Wright, R. P. (1993). The Harappa museum pottery collections: Some typological and technical evidence. Pakistan Archaeology, 28, 231–241. Wright, R. P., Lentz, D. L., Beaubien, H. F., & Kimbrough, C. K. (2012). New Evidence for jute (Corchorus Capsularis L.) in the Indus civilization. Archaeological and Anthropological Science, 4, 137–143. Wulff, H. E., Wulff, H. S., & Koch, L. (1968). Egyptian faience: A possible survival in Iran. Archaeology, 21, 98–107.

Traditional Bead and Bangle Crafts in India Alok Kumar Kanungo

Abstract Production of glass was one of the most advanced technical developments of the primeval world. It required knowledge of furnace building, glass recipes and pyrotechnology for maintaining the furnace temperature for weeks. Likewise, working with glass also involves maintaining equilibrium of temperature in the furnace/kiln to keep the glass in semi-molten stage for the entire period of work. The antiquity of glassmaking and glass working in India is 3500 years old; the evidences are mainly in the form of beads and bangles. The furnace-wound glass bead–banglemakers of Western Uttar Pradesh and drawn glass bead-makers of Chittoor district of Andhra Pradesh are few of the last living craftsmen who have inherited a predominant part of technological know-how from their ancestors. There have been very few alterations to their production method since they began the craft production. Their furnaces are either indigenously developed or made based on the knowledge craftsmen acquired from where they migrated. These cottage industries thus hold the key to many archaeological puzzles about glass beads-bangles and furnaces.

1 Introduction The antiquity of glass beads and bangles in India goes back to 3500 years old (Kanungo, 2021). For more than 2500 years, India has been one of the leading glass bead–bangle providers to the world. And the traditional glass bead–bangle making crafts have survived in certain corners of India without major alterations for centuries. For nearly two decades, I have visited the famed Purdilnagar, Jalesar, Akrabad, Hasayan (PJAH) region in western Uttar Pradesh and Papanaidupet in Andhra Pradesh to investigate the traditional glass bead–bangle making. Beadsbangles from these areas continue to dominate the world glass jewellery market at an unprecedented scale.

A. K. Kanungo (B) IIT Gandhinagar, Gandhinagar, India Flinders University, Adelaide 5001, SA, Australia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_4

101

102

A. K. Kanungo

2 Glass Beads Beads are used either as decorative items for humans and animals or in craft items. Glass has been regarded as the most favoured raw material for bead making. Glass bead production is a unique process as is glass production. Technically, glass beads are produced by heating raw glass above its melting point and allowing it to cool without crystallization. Glass beads are usually small with varying shapes and versatility in designs. Due to these eye catching features, there has been temporal and spatial mobility and frequently glass beads are used by different generations of the same family as heirloom. Thus, it is an important marker of trade and an object to be studied in its spatio-temporal context. There are varied ways to shape a bit of glass into a bead. Over the centuries, glassworkers have attempted every feasible technique to make beads such as winding, dipping, drawing, mould pressing, marvering, cutting, grinding and drilling, and at times a combination of these methods are used for making glass beads. Many a time, those were indigenous innovations. Different modes of manufacturing beads have been discussed by Alister Lamb (1965), van der Sleen (1973), Lugay (1974), Francis Jr. (1983, 1993), Kucukerman (1987), Robert Liu (1989), Ross and Pflanz (1989), Bronson (1990), Karklins and Jordan (1990), Kishore Basa (1993), Karlis Karklins (1993), Kock and Sode (1995), Kanungo (2001a, 2001b, 2004a, 2004b, 2006, 2014, 2016, 2019), Carroll and Allen (2004) and Holland and Holland (2006). Most of the beads found at archaeological sites in the Indian subcontinent seem to be produced using either drawing or winding techniques. These two most ancient and important techniques of bead manufacturing continued to be in use in India even today: drawing technique at Papanaidupet and furnace winding at PJAH. For both the raw material and the fuel have been the same.

3 Bead Productions Cycle 3.1 Raw Material Nowadays, all of the raw glass in cake or cane form comes from the glass furnaces of Firozabad, the glass city, which produces glass for all Indian bead- and banglemakers. The raw glass travels 72 km north to Purdilnagar and PJAH and 1300 km south to Papanaidupet. The glasses also travel 500 km west to Banaras (Varanasi, Uttar Pradesh), where a booming business produces fine lamp beads. In both PJAH and Papanaidupet raw glass used to be produced till the end of twentieth century (Kock & Sode, 1995; Sode & Kock, 2001; Kanungo, 2016; Gill, 2017). While carrying out ethnographic fieldwork in the glass furnaces of Firozabad, an interesting fact is noticed that the bead–bangle-makers who come to the glass furnaces to buy raw glass, always get some beads-bangles along with them to the

Traditional Bead and Bangle Crafts in India

103

factory for colour matching and very often leave behind these beads-bangles at the factory.

3.2 Fuel Fuels are usually of three types: (1) wooden logs, (2) brush wood and (3) cow-dung cakes. Wooden logs are mostly bought from the market; brush woods are collected from the nearby forest/bushes, and the cow-dung cakes are prepared by the villagers. The furnace is first heated with wooden logs and to maintain the high-temperature cow-dung cakes are added. For increasing and decreasing the temperature in the furnace, brush wood is used. Although the raw material and the fuel remain same, the method of production and the end product differ temporally and spatially.

4 Ethnographic Survey of Bead–Bangle Production Cycle in PJAH Cluster In the field of bead studies, furnace-wound glass beads have become synonymous to the villages of Purdilnagar, Jalesar, Akrabad and Hasayan (PJAH), primarily because of the mass-scale export-quality bead production taking place in this cluster for centuries. For the manufacture of wound beads, glass is heated in a crucible inside the furnace. An iron rod (mandrel) is dipped into the glass, and while being taken out, with a dexterous twist, the glass bead is produced. While hot, it may be shaped with paddles and other tools or other glass variant may be added for decorative purpose. The furnace winding and moulding technique is regarded as the oldest, simplest and the most common method of bead making (Basa, 1993: 93; Francis, 1992: 15; Sleen, 1973: 23). One characteristic that is common in all wound beads is that the fabric of the glass, including additives/residues, is oriented around the perforation. In other words, striation marks can be seen around the perforation in such beads (Fig. 1). The antiquity of this process in India can be traced to 1000 BCE, during Painted Grey Ware (PGW) culture. Jalesar and Purdilnagar are situated on the road between the glass city of Firozabad and the Indian capital, New Delhi. From Firozabad, at about 40 km, comes the town of Jalesar (Etah district) and at about a distance of 65 km from Firozabad is Purdilnagar (Hathras district). Further to the north-west, on the south from Purdilnagar to Aligarh, is the small settlement of Akrabad (Aligarh district), at a distance of about 42 km from Purdilnagar. Hasayan is located in the district of Hathras and is located at 12.5 km south-west of Purdilnagar (Fig. 2). All these four villages make the cluster which

104

A. K. Kanungo

Fig. 1 Striation marks around the perforation in wound beads

is famous for the production of traditional furnace-wound glass beads and joint-less bangles.

4.1 Firing the Furnace Generally, the furnace owner, the bead–bangle merchant and the craftsmen work independently but are dependent on each other. Work in the glass working furnace begins early in the morning, at around 4 O clock, with the stoker lightning the furnace with wooden planks and cow-dung cakes below ground, with a floor in the furnace (open at the centre), holding the glass away from direct flames. It takes about an hour for the furnace to reach the right temperature (about 600 °C) for the day’s work to begin. An iron pipe instead of any fixed tuyere is used to blow air into the fire. The crucibles are filled with glass pieces. It takes another hour of firing for the glass to melt sufficiently to be eligible for the manufacture of bead and arm-rings (bangles and bracelets). Often, new solid glass (at times uniform coloured broken bangles) is added at the back of the crucible while work goes on from the front. When the added glass melts, the bead-maker mixes those well and pulls molten glass to the front of the crucible with a stirring iron. Furnace for bead and bangle making are same although the ports, the craftsmen and their expertise varies. The following describes the methods of production of various designs based on the observation made in different furnaces, in general, and

Traditional Bead and Bangle Crafts in India

105

Fig. 2 Locations of native glass furnaces: Purdilnagar, Jalesar, Akrabad, Susamayee and Hasayan

in Sattar Bhai Bead Furnace (27° 39 32.76”N; 78° 22 21”E) at Purdilnagar, in particular.

4.2 Bead Making The craftsmen squatting in front of individual ports of a circular shaped furnace make wound beads with ductile glass wound upon a mandrel. The bead-maker picks a small amount of molten glass from the crucible with a thin and short mandrel (a pointed slender 1 m long iron rod) and then winds the pick on a second mandrel (a comparatively thicker but pointed slender 1.5 m long iron rod). The bead-maker goes on winding several picks on the same mandrel (Fig. 3). The wound material is given shape by rolling forwards and backwards on a smooth metallic surface (usually a piece of old well-worn railway girder) with the mandrel as an axis. The bead is shaped with frequent use of a form-iron and/or by rolling in a single or multi-channelled

106

A. K. Kanungo

Fig. 3 Winding of glass around the mandrel

trough (Fig. 4). While the bead is hot, it may be designed variously with paddles, form-irons and other tools. Form-iron is an oblong piece of iron plate which is used to forge the bead into its final shape. A wound bead can be re-inserted in the furnace when it is still hot and stripped or coated with another coloured glass with the help of the shorter mandrel. Two colours can be maintained in one crucible, because in these crucibles, glass is not allowed to become liquid. The craftsman puts back the mandrel with beads in the furnace for five to ten minutes for the first phase of annealing. After removing the mandrel from the furnace, the beads are separated from each other with the help of mala (an arrow shaped iron tool). Upon completion of making the beads, the mandrel is stroked hard with the mala so that the perforation becomes a little bigger and the bead is slipped off during the brief period when the iron cools and contracts faster than the glass. The bead is allowed to slide down into a small clay pot or tin with annealing cotton in it, placed immediately below the working port facing the annealing port. When the pot is full with beads, it is covered with hot ash from the furnace and placed under sunlight (if available) and allowed to cool slowly (Fig. 5). This is necessary to prevent stresses and cracks that might appear in the bead due to sudden cooling. Still one does find plenty of such cracked beads and bangles around the furnace area and throughout the village (Fig. 6). However, in most of the cases, glassworkers put the cracked beads and bangles back into the furnace for re-melting. The ones having oxidization marks or polychrome ones are generally not recycled, unless a certain combination is desired. This leads to accumulation of such debitage in large quantity, which has the potential of altering the statistics of debitage of all varieties of bead production.

Traditional Bead and Bangle Crafts in India

107

Fig. 4 Shaping the bead in a trough

Fig. 5 Annealing

4.3 Pendant Making The port in which pendants are made is comparatively bigger in size. The craftsman picks and pulls a lump of glass out of the crucible. Using a scissor, he cuts the required amount of glass and puts that on a mould. Then he uses a mala and presses the molten glass in the mould. Within five minutes, the pendant is removed from the mould to a plate or a tile made of rajmahal clay. Rajmahal clay works like kaolin. After six pendants are made and placed on the plate, certain amount of fine rajmahal clay powder is sprinkled on top of the pendants. The tile with pendants is then transferred

108

A. K. Kanungo

Fig. 6 Production debitage near the Furnace

to a U-shaped clay ledge built on the hanging crucible inside the working port but near the opening. After annealing for about ten to fifteen minutes, the craftsman puts the pendants in the tin container insulated with annealing cotton on the annealing port for the second phase of annealing. By the time the tin is filled with pendants, the craftsman requires a break from work, and on return he replaces the annealing tin for another set of pendants (Fig. 7). A decade and half ago, when I was carrying out fieldwork in the same area, none of the craftsmen were using annealing cotton. Only a few used tin, but most of them were using ceramic pots placed under the outer periphery of the working port as an annealing chamber. As a result, cracked by-products were more abundant outside the furnace area.

4.4 Eye Bead Making Small eyes are first made by putting a small black or dark blue glass disc at the centre of a white disc while white glass is still in a semi-molten stage. Then navy blue colour (preferably) glass is taken on a mandrel, and the required amount is cut from the mandrel using a scissor and placed on a tile. The shape to the disc is given by a spring tong. When the disc is still in the semi-molten stage, the prepared eye is put at the centre and pushed to the required position. At one periphery, a perforation is made with the help of a punch/nail. Using the spring tong repeatedly, the shape of the

Traditional Bead and Bangle Crafts in India

109

Fig. 7 Making of pendants

disc is maintained (Fig. 8). Annealing phases are followed, as is done for pendants described above. There is a form of design that is also known as eye beads in the Indian context, i.e. to decorate the initial monochrome opaque beads with applied threads or eye motifs; such beads are found from a number of early historic sites. Another Indian decorative technique, which has been in use since ancient times, is the so-called powder-glass beads. These are usually black in colour, with a pattern of powdered glass melted into the surface. Such beads are seen in a number of museums and are recorded in various publications, especially collected from early historic archaeological sites such as Kausambi (Gupta, 1997), Ahichchhatra (Dikshit, 1952), Ter (Chapekar, 1969), Nevasa (Sankalia et al., 1960) and Kopia (Kanungo, 2013). To produce such beads, the desired pieces of minute or crushed glass are placed for heating on an iron plate positioned just in front of the working port. A wound bead, while still hot, is rolled over the plate and then heated in the furnace again, so that the attached glass melts to give the form of a bead. The bead is given its

110

A. K. Kanungo

Fig. 8 Making of eye beads

final shape with a form-iron. Then, the hot wound glass is coated with a thin (final) layer of transparent glass. At times, a specific patterned design is made by rolling the hot wound on a patterned engraved iron plate filled with fine powdered glass, so that the glass powder sticks to it (Fig. 9). The glass is then heated, and the bead is finally moulded in a form-iron.

Traditional Bead and Bangle Crafts in India

111

Fig. 9 Rolling of wounds over an engraved pattern and pattern imprinted beads

4.5 Foiled Bead Making To make gold/silver foiled beads, a master craftsman picks the required amount of glass from the crucible as for any other bead. The glass is then rolled into a cylinder shape on an iron rail. While it is still hot, silver foil gets instantly wrapped on it, which is kept on the iron rail (Fig. 10). The silver foil wraps itself around the wound. A layer of transparent glass is then wound on the silver foil surface. The final product can be given varied shapes. Melon shape is one of the widely used ones. This is made using a tin separator, having a U-shaped groove at the designing or the moulding end. After the transparent glass is given a ball shape, the separator is used to make deep ridges on the surface along the diameter (Fig. 11). For shine, the bead is given a twist in a small bowl having fine sand and corundum powder before sliding it into the annealing container.

Fig. 10 Winding of silver foils

112

A. K. Kanungo

Fig. 11 Groove is given to wound

4.6 Mould Bead Making After winding the hot glass on the rod, a press tong mould is used for giving various shapes. A very thin seam is usually seen in moulded glass beads (Fig. 12). On joining the mould, the flow of excess glass creates a thin circumference around the piece or the seam, which is caused by the differences created by differential cooling of the glass in the upper and the lower faces of the mould. In the recent past, with technological development, such marks are removed successfully by polishing the beads in a rotating drum with corundum powder (Fig. 13). Likewise, workshops engaging in giving metallic and matt finish to beads using various chemical and synthetic colour have been mushrooming in Purdilnagar and Jalesar.

5 Bangle Making Joint-less glass bangles are in use right from PGW culture in India and are produced in large quantity at PJAH in western Uttar Pradesh. The process of bangle production observed: • The making of joint-less bangles begins with the picking of required amount of molten glass on the mandrel from the crucible.

Traditional Bead and Bangle Crafts in India

113

Fig. 12 Moulded beads with seams and before cleaning

Fig. 13 Use of corundum powder in wooden drums

• A wire (unaar) is inserted between the molten glass loop and the mandrel. • The mandrel is continuously rotated with the unnar kept constant to give the molten material a circular shape with a considerably larger perforation. • While it is still in red hot condition (malleable form), the loop is mounted on a khalbut. • The wounded glassy material is placed along the diameter of the cone and continuously rotated with the unnar being constant. • The khalbut is rotated on a rail grid using the palm.

114

A. K. Kanungo

• Alternatively, an iron stand with a U-shaped depression on top is used to support the khalbut. The portion where the khalbut is joined with the mandrel is kept on this depression, rather than on the rail grid, and is continuously rotated. • In both methods, the end of the iron rod is placed on a stone slab as the rotation creates friction between the two surfaces and helps a smooth rotation. The cone is made to work facing the port so that the bangles on khalbut get the required temperature. This permits gradual enlargement of the wound glass from the tip of the cone. • When the required size is attained, the bangle-maker flips the bangle on the cone so that both the sides of the bangle get internal polish and near symmetrical concaveness. It also helps to maintain uniform thickness of the bangle. In half-aminute, the ring attains the proper size and becomes cool enough to slip off the cone (Fig. 14). During this process of shaping the bangle, many a time, the bangle is rolled over a powder of golden sparkle, also known as kirkira, to achieve the sparkling effect on its surface. Coating the surface of bangle is also done by using a glittering strip called zari, which gives a symmetric pattern. In most cases, the bangle-maker picks zari pieces of different colour on the glass loop from an iron plate kept on the outer periphery of the working port. These decorative pieces are inlayed at regular interval, and then the wound is put in the furnace for a minute or two and taken out to give the shape in a channelled turf so as to ensure proper fusing and shape (Fig. 14). At times,

Fig. 14 Glass bangle making using Khalbut

Traditional Bead and Bangle Crafts in India

115

the mala (a multi-purpose tool, shaped as an arrow) is put to use to press the designer pieces on the glass wound, so that they do not disorient or fall. As per the requirement, a layer of transparent molten glass is coiled over the coated surface, and then the bangle is transferred to the khalbut for subsequent resizing and cooling. The look can be enhanced endlessly depending on the craftsmen’s imagination by twisting, pressing, stamping, pinching, crimping, embossing, appliqueing and polychrome designs. Once a bracelet or bangle is made, the craftsman puts it on a plate kept near the working port for annealing till the second bracelet/bangle is made. After this, he slides the bracelet/bangle to the annealing plate/pot for further annealing. The method which precedes the khalbut was perhaps the two-mandrel method. Gaborieau (1977) provided a detailed account of the two-mandrel technique practised by the Churihars, bangle-maker communities living in hills of western Nepal. Using a much smaller and readily devised two-chamber kiln, the bangle-maker winds/loops glass on a mandrel. Once satisfied, the mandrel is struck repeatedly with a mala to detach the loop and begin the enlargement. The slightly enlarged loop is transferred to a second mandrel, which is smeared with rajmahal clay on a small portion of its length. Through a dexterous series of rotations of these mandrels in different directions, while holding the bangle at the kiln mouth to keep it at the desired temperature and plasticity, the bangle is enlarged to circular shape (Fig. 15). While the ring still on the mandrel, the bangle-maker could reintroduce it to heat at the kiln’s mouth and, if desired, make it polychrome with application of different coloured glasses. It is

Fig. 15 Glass bangle making using two mandrels (Courtesy Marc Gaborieau)

116

A. K. Kanungo

important to mark that the kiln used by this community is much small in size though of same shape and style as that of the PJAH (for details see Trivedi, 2021). In between these two ends, one finds reference of large-scale use of mid-sized proto-industrial bangle workshops in Punjab (Kanungo 2021: Fig. 14), where kilns have six working ports and two air flow and firing windows, and use of khalbut towards the end of nineteenth century (Hallifax, 1892).

6 Furnace At PJAH, beads and bangles are produced at same furnaces, and one finds a huge concentration of furnaces immediately after entering the villages. In contrast to general opinion that glass-working areas are placed on the outskirts of habitational areas as they generate enormous heat, they are often located within the habitation area. The walls of the furnaces are however not shared with any residential house (Fig. 16). The furnace is entirely made of clay. A number of rectangular, sun-dried clay plates are placed standing upright around the pit that is used to serve as the base of the fire-chamber. These plates separate individual workstations. The number of plates shows how many craftsmen can work around the finished furnace. The number of working windows in the furnace vary between 13, 15, 17 and 19, and it is never an even number (Fig. 17). The circumference of the furnace is bounded with a

Fig. 16 A Bead furnace house at Purdilnagar

Traditional Bead and Bangle Crafts in India

117

Fig. 17 Bead and bangle-makers at work

line of clay bricks at the ground level. The fuel is placed on an iron mesh. Brick nodules are placed on the mesh, and they act as heat retainers. The 3.5–4 feet deep upturned triangular shaped pit performs two functions: one as an ash receptacle and two providing vent/passage for constant oxygen flow for the fuel to continuously burn (Fig. 18). A clay dome is made over the standing clay plates. In any case, the top is left open and is covered with a clay plate only when the furnace is in use. This opening facilitates a regular change of crucibles. The furnace is made of coarsely tempered clay mixed with chopped straw. When the furnace is fired, the straw burns off, leaving a highly porous clay furnace that can cope with the great range of temperatures between the day’s high temperature and the night’s cooling effect. The opening from which glass is picked is called the window, and inside is a small open crucible for the glass to be worked on, which measures about 25 × 40 cm and is made of clay mixed with chopped straw. The windows are uniquely designed for various products like beads, bangles and other objects (Fig. 19). Underneath, on the outer periphery, a small individual pit facing yet another port holds the annealing pot. It is also necessary to change the crucibles at certain intervals, because it will normally be burnt after about two weeks. When crucibles are replaced or the structure is repaired, the outside of the old furnace is built again with new, straw-tempered clay. With this type of furnace, one can produce temperature that is high enough for the making of glass beads and arm-rings with the least possible consumption of fuel. In normal circumstances, the furnace can be used for about six months, depending entirely on how deftly it is built and maintained. Thus, one sees continuous

118

A. K. Kanungo

Fig. 18 Schematic section of the furnace

reconstruction of furnaces in the village, and they are generally constructed at the same place after demolishing the earlier one completely. Many a time when a furnace develops cracks, it is abandoned right there and then and a new furnace is created in the adjacent land.

Traditional Bead and Bangle Crafts in India

119

Fig. 19 Different windows for bangles and beads, pits for annealing pot is seen

7 Traders Associated with Glass Trading In these and surrounding villages, there is a section of people who generally do not produce beads themselves but prefer to play the role of middlemen, facilitating trade with bead traders from big cities like Varanasi, Delhi, Mumbai, Hyderabad and Bangalore. There are a number of such local traders in Purdilnagar village; however, none of them send beads to the international market directly. Banaras Beads Limited (BBL), the biggest trader of glass beads in India, has an office at Varanasi, dealing with international trading of beads, which are ordered through middlemen. These middlemen also send beads to other cities, and the details of their modus operandi and BBL’s network are presented below as a case study. Orders for beads of specific style come to BBL → the bead is replicated or designed by the master craftsmen at BBL → on satisfaction the customer gives the order and leaves BBL → middlemen are called to BBL and production method for the specific shape is explained → at times moulds and glass are also given to the middlemen → villagers working in varied glass furnaces come to the middlemen and go away with the models and orders for specific quantity, made with specific glass type using the specific mould shape. In lieu sometime they are given advance payment → on completion of the job craftsmen get the beads back to the middlemen → quality assurance and weighing of products takes place at the place of middlemen before the balance payment is paid and new orders are placed → bead traders (middlemen)

120

A. K. Kanungo

Fig. 20 Sorting by Middlemen and the sorted debitage

re-sort the beads. In the process of sorting the debitage encompassing half-moulded, closed perforation, cracked beads, beads with air bubble and transportation debitage, all of which are generally discarded nearby, leading to piling of such debitage near a place where beads are not actually made (Fig. 20).

8 Mould-Makers There are a few smithy masters in the village who are engaged only in mould-/dyemaking, and one finds plenty of model beads with them. These models may have travelled from any part of the world and might have had different glass composition than the local ones. The same applies to the model which goes to the local furnaces to be copied as per the consumer’s requirement. Interesting enough, the dye-maker generally breaks the model beads into two parts to make the mould/dye more accurate (Fig. 21). Thus, this also becomes a location for bead debitage. In case the mould/dye-maker’s place is away from the production place, the corresponding debitage could accumulate even at a distant area.

Traditional Bead and Bangle Crafts in India

121

Fig. 21 Mould-makers

9 Stringing and Cleaning of Beads Once the beads are produced, either the middlemen or BBL gets the hole cleaned of the oxidization marks. The cleaning is done by pulling the bead forward and backward on an aluminium wire fixed to the ground or to a wooden plate (Fig. 22).

122

A. K. Kanungo

Fig. 22 Women cleaning the oxide residues from the perforations

This takes place in many villages surrounding the region, and it is mainly carried out by the women folks. In this process of cleaning the perforation, a few of the beads break apart. After clearing the perforation, the beads are cleaned thoroughly and polished on a wooden grinder and rubbed on a gunny bag or coir mat in BBL. After sorting, the beads are either packed for export or strung by ladies. For cheap labour, most of the beads are sent to different neighbouring villages for stringing.

10 Lamp Beads In India, today lamp winding technique is practised and mastered by the BBL, Varanasi. The industry was established in 1940 by the Late Shri Kanhaiya Lal Gupta after getting training under the able guidance of a Czech couple, Mrs. & Mr. Heinrich, who had visited the Ceramic Engineering Dept. of Banaras Hindu University in 1938 and taught lampwork beads. The training program continued till 1962 in the University, and a number of craftsmen from Purdilnagar got trained. Lamp winding table is fitted with a blowing burner and a pedal bellow. A worker sits behind an array of six to seven kerosene nozzles, each with a blowpipe that intensifies and directs the flame to a spot in front of a stone crucible. Flame is topped with a tin cover for protecting vision and providing clarity of observation. The blowpipes are run by bellows pumped continuously by the worker’s foot that makes it possible to achieve a temperature high enough for the glass to be melted and worked on the nozzles. A stone stand with a channel is kept below the flames. Then the craftsman, putting a finger cap on his index finger, picks up a copper wire

Traditional Bead and Bangle Crafts in India

123

Fig. 23 Lamp Bead Making

pre-coated with rajmahal clay (a substitute of kaolin), in one hand, and one or two glass canes, in the other, and starts melting it in the flame and winding it around the wire (Fig. 23). While the beads are hot, they are shaped to near perfection in iron half moulds. These are dies made of small metal cubes with depressions on one face corresponding to half the ultimate section of the bead. Grooves running from the centre of the depression to the edge of the die allow the wire to rest in them. The bead and wire are laid on the half-mould, and the wire is twirled, shaping the bead by this rolling action in the depression. Both chipna (a flat iron strip with raised edge) and kahari (a chisel-shaped spatula) are used for pressing and moulding glass, and dohri chimti (spring mould) is used for both shaping and separating the beads while they are still in a semi-molten state on the mandrel. While the glass is still hot, the bead may be shaped or given decoration using glass canes of other colours. These beads can be annealed in the lamp flame. After seven or eight beads are made up on/around the wire, it is kept upright in a pot/tin cane full of sand. When cooled, the clay allows the beads to easily skip off the wire, leaving a white coloured coating in the hole. Such beads are produced in hundreds of villages in the vicinity of Varanasi and around Purdilnagar (for details see Kanungo, 2001b).

124

A. K. Kanungo

11 Chevron/Millefiori For more than half a century, a few artisans at Purdilnagar have practised the art of making chevron canes. Either they produce chevron beads out of these canes or after slicing the discs from these canes they make designed millefiori beads, bangles and other varied objects such as bowls and fountain pot (for details see Kanungo, 2019). At present, chevron glass canes and tubes are being produced only at two kilns, i.e. wood fuelled kiln owned by Heider Ali (27° 39´ 28.38 N; 78° 22´ 13.30 E) (Fig. 24) and gas fuelled kiln owned by Mahammad Hasain (27° 39 23.76 N; 78° 22 23.88 E) (Fig. 25). The art of chevron and then millefiori was mastered by Heider Ali in 1977 by the art of trial and error. His interest was caught by a millefiori bead brought to him by the Managing Director of BBL, Shri Ashok Gupta, who had bought it in Venice. He duplicated similar chevron canes as used in the Venetian millefiori bead at a bead furnace in village Hasayan (his wife’s village) and then on has been producing such beads in his village, Purdilnagar in a customized wood fuelled kiln. Hasain Bhai learned the art from Heider Ali in 1990 and later innovated a gas fuelled kiln to produce chevron. Their products are sold as chevron beads and millefiori products in South and South-East Asia and Africa. The kilns are comparatively smaller than the bead–bangle furnace, having six to nine working ports, each having independent hanging crucible. All crucibles carry independent colour glasses. The entire act of producing chevron canes or tubes is carried out by only one master craftsman, with the help of a junior artisan.

Fig. 24 Gas fuelled Chevron Kiln

Traditional Bead and Bangle Crafts in India

125

Fig. 25 Wood fuelled Chevron Kiln with Ash Pit

11.1 Chevron Making Applying layers of required coloured glass stripes (from different crucibles) all around a base colour glass cylinder on a steel pipe, required pattern is made in the section of the mass (Fig. 26). If the aim is to draw tube for pre-perforated chevron,

126

Fig. 26 Applying layers of different colour glass for making Chevron Pattern

Fig. 27 Blowing air into glass

A. K. Kanungo

Traditional Bead and Bangle Crafts in India

127

then it is necessary to frequently blow air through the same steel pipe so as to ensure it holds its shape (Fig. 27). Joining a mandrel (tip of which is dipped in the molten glass) with the other end of the mass while making the entire designed glass malleable in the kiln, the mass is taken to a drawing platform (next to furnace room) and pulled on both side of a

Fig. 28 Drawing of Chevron

128

A. K. Kanungo

prefixed hook (Fig. 28). The pulled canes are then annealed and sliced to the required sizes. The flow of work is as follows: Breaking the glass → Putting glass in crucibles → Melting of glass → Gob formation → Parison preparation → Reheating → Applying different colour glass till the required chevron pattern is achieved by reheating with each application of glass → Blowing → Reheating → Attaching a rod at other end → Drawing the glass tube → Cutting → Annealing → Sorting → Packing → Dispatching.

11.2 Millefiori Beads There are two ways of making millefiori beads. One way is to place the chevron sticks on an iron plate. The sticks are placed on one end so that the pattern at the other end faces upwards. The iron plate is put just outside the working port so that the sticks are heated up. Then a small quantity of glass is caught on the pontil from inside the furnace and is shaped into a narrow oblong bead that is then rolled across the sticks so that the end-pieces stick firmly on to the bead. The bead and the mosaic sticks are then melted together in the furnace with the bead being given its final shape with the form-iron. This technique leaves the finished millefiori patterns sitting on the outside of a core of monochrome glass. The second method of making millefiori beads is by firstly placing discs of the chevron/murine canes in a pre-determined pattern on a kaolin (rajmahal clay) tile. The discs are brought to a required malleable state by placing the tile near the working port. Secondly, semi-molten glass is taken on a pontil and rolled over the chevron discs so that they get fixed in the required pattern. Thirdly, the designed core is again placed inside the furnace for the chevron discs to get embedded and give it a millefiori like look. Then the finishing is done by rolling the millefiori over the rail grid with simultaneous use of form-iron.

12 Zigzag-Patterned Bangle Nowadays, zigzag-patterned bangles have also come into vogue as a result of which the technique of bracelet-making has also undergone innovations. To produce the same, first the bracelet is moulded on a khalbut, following the same procedure as that for above-described bangle. While it is still in the mouldable stage, it is transferred to an octagonal ridged iron khalbut of conical shape with a zigzag-patterned upper base. When the diameter of the bracelet is widened to reach the size of the zigzagpatterned base, a negative zigzag-patterned cap is pressed. The length of the cap is slightly smaller than the metal khalbut, and the in-between space creates the required zigzag-patterned bracelet (Fig. 29).

Traditional Bead and Bangle Crafts in India

129

Fig. 29 Zigzag-patterned glass bangle making

13 Ethnographic Survey of the IP Bead Production Cycle at Papanaidupet The most common drawn beads are known as the Indo-Pacific beads. These are generally three to five mms in diameter. They are usually round in cross section, but their profiles are varied. They can be oblate, discoid, tubular or of other shapes. One characteristic that is common in all drawn tubes and beads is that the fabric of the glass, including any additives/residues, is oriented along the perforation. In other words, striation marks can be seen along the perforation in such tubes/beads (Fig. 30). They are traditionally known as mutisalah bead in south-east Asia and seed bead in south Asia. On the basis of their distribution from southern Africa to Korea, Francis (1986) coined the term ‘Indo-Pacific Beads’ (IP). Their distribution is even wider than the name suggests. It is important to note here is that the finds of such beads in in-land areas are multiplying by years. Recently, the site of Kopia in north India has yielded these beads from mid fifth century BCE (Kanungo, 2013, 2016). Distribution and antiquity of this bead are described in detail by Kanungo (2016). Francis (2002, 2004) and Stern (1987a,1991) have hypothesized that the technique followed at Arikamedu in ancient times is similar to that practiced in present-day Papanaidupet. Recently, Kanungo (2010, 2013) and, Rajan et al. (2013) and Rajan & Kumar (2014) have respectively provided evidence to suggest that IP beads were also produced at

130

A. K. Kanungo

Fig. 30 Indo-Pacific Bead Tube’s Waste from Arikamedu, striation marks along the perforation are seen

Kopia in Uttar Pradesh and Porunthal in Tamil Nadu around the same time as that of Arikamedu and probably by using the same technology. There have been number of claims about the first IP bead production centre to be in south India (Francis was more particular on Arikamedu to be the site), but lately both archaeological and compositional evidences have piled up in support of existence of several independent IP bead-making centres at different points across south and southeast Asia in precommon era time period (Brill, 1999; Dussubieux, 2001; Dussubieux & Gratuze, 2003; Robertshaw et al., 2003; Kanungo, 2004a, 2013, 2016; Popelka et al., 2005; Lankton & Dussubieux, 2006; Dussubieux et al., 2008). Today, Papanaidupet is the only surviving traditional Indo-Pacific bead producing centre in the world (Stern, 1987b; Francis, 1990; Kanungo, 2000–01, 2004a, 2016). Papanaidupet is located in Yerpedu Mandal in Chittoor district of Andhra Pradesh State. The drawn/IP bead production cycle has been practised in this village for more than two hundred years.

Traditional Bead and Bangle Crafts in India

131

13.1 Drawing Furnace Furnaces are located at the southern outer periphery of the village at a safe distance from habitation area. A dome-shaped furnace is constructed using clay, mud-bricks and iron, on a five feet deep pit. In between the upturned triangular-shaped pit and dome-shaped furnace, there is an iron rod mesh for placing the fire atop with ashes to fall down. It has four openings for different purposes, a high port with iron base for the lada (the lada is the most important tool and is typical of IP bead production. It is a heavy iron pipe having one flaring end to hold the glass), a low one for fuel, a trough on one side having high platforms made of mud-bricks for putting raw glass and colouring agents on it, and a five-feet-long annealing tunnel on the fourth side through which the tubes are drawn. The glass-chunk platform is constructed at a 20° sliding fashion towards the furnace, and this is at a higher level than all the other ports. The lada port and the drawing tunnel are made at the same height, whereas the fuel port is at the lowermost level (Fig. 31). The ports are placed and designed in such a manner that the fire gets directed to the glass-chunk port, and when this port is closed, it remains in the centre instead of getting out of the lada port. This furnace requires regular maintenance (on every alternate day), through the application of a thin layer of mud to its surface. After the use of the furnace for a year or less, it develops too many cracks to be used further. Then the entire furnace is broken, and the workers shift to a new furnace. A replacement is built at the old site. The furnace is housed in a structure that is granite stone pillared, wooden beamed and tile roofed. The high roof made of tiles helps for the outflow of the heat and smoke. Towards the end at a level lower by one foot, a platform is made for the master to sit and pull the tubes. The right side and the drawing end of the entire structure have mud walls, and the other sides are left open. This arrangement protects and maintains the fire at the required temperature. At the same time, it lets in enough light and air for the workers. There are two more structural requirements, one is a raised stone platform, which is marginally tilted towards the working side for shaping the glass cone. In between this platform and the platform for glass chunks, there is a structure for placing molten glass and replacing the gedda paru (metal-sheathed pole with hafting of a long wooden handle) by the lada.

13.2 Melting of Glass As the sun sets, the workers get into the furnace enclosures one by one. Before the process starts, a slurry of cow-dung is smeared all over the glass-chunk platform. Logs are placed inside the furnace and fired. Between 50 and 60 kg of glass chunks are placed on the designated platform. They are covered with customized (semi-circular shaped with a hole on the body) ash-smeared pot pieces. This is done apparently to avoid the sticking of glass to the platform and the potsherd coverings.

132

Fig. 31 Schematic plan of drawing furnace

A. K. Kanungo

Traditional Bead and Bangle Crafts in India

133

Fig. 32 Indo-pacific bead-making furnace at work

The fire in the furnace is increased, and both the upper and lower portions of the lada port are covered with an inverted half-pot and an iron plate with a hole, respectively. These holes help in picking up the hot coverings when required. The fire goes out of the glass platform port and going through the potsherds start melting the chunks in 40–50 min (Fig. 32).

13.3 Making the Glass Cone Once the glass melts, the potsherd coverings are taken out with iron hooks. The glass is lifted around a pre-heated gedda paru and put in the furnace through the lada port, for melting and kneading. In the meantime, another gedda paru is inserted into the whole glass for twisting it to get a better mixture (Fig. 33). When the glass attains required malleability and is well mixed, the whole mass is taken out and placed on the platform in between the raised platforms for glass chunk and cone making. This area keeps the glass in semi-molten stage for long enough to replace the gedda parus with a lada. The molten glass is then put around the lada (Fig. 34). Once glass is put on the lada, both the gedda parus are allowed to cool off under natural conditions so that the glass sticking on it falls off. The gedda parus are then smashed on a nearby stone, which causes thin, slightly curved glass flakes, having rough oxidized markers on the inside surface and shiny, smooth exteriors, (often with a tail) to come off (Fig. 35).

134

Fig. 33 Mixing of glass

Fig. 34 Glass are mounted on Lada

A. K. Kanungo

Traditional Bead and Bangle Crafts in India

Fig. 35 Gedda Paru and the Flakes

Fig. 36 Pierced cone

135

136

A. K. Kanungo

The lada with glass is inserted into the furnace (again) and continues to be heated till it becomes malleable enough to be given the required shape. It is then taken out of the furnace and is run atop the raised stone platform and manoeuvred to a conical shape. A long iron rod with pointed end called chetak (smothered butted ended long iron rod) is put into the lada and thrust at the base of the cone for perforating the cone from base to apex (Fig. 36).

13.4 Drawing the Tubes Temperature is lowered in the furnace so as to maintain the malleable state of the glass cone. The glass platform port is covered with an inverted half-pot supported by a few more. The glass cone with the lada is placed inside the furnace. The upper portion of the port is covered with a comparatively bigger inverted half-pot, and the

Fig. 37 Rolling and unrolling of Lada

Traditional Bead and Bangle Crafts in India

Fig. 38 Grabbing the tip of the cone with a U-tipped rod

Fig. 39 Horn-shaped glass

137

138

Fig. 40 Drawing the tubes

Fig. 41 Old master checking the perforation

A. K. Kanungo

Traditional Bead and Bangle Crafts in India

139

Fig. 42 Swollen glass tubes (Knots)

lada is placed between two driven poles to prevent the cone from rolling out in any other direction than the tunnel. The poles are put at three and half feet distance away from the lada port, and the distance between the two poles is equal to that of the lada port (Fig. 37). Two workers place themselves in front of the furnace to maintain the continuous movement of glass mass moving back and forth on the base of lada port (made of iron). This keeps the glass around the lada and does not allow it to fall off or melt. On the opposite side, a master takes a metal hook, with a U-shaped tip, peeps through the tunnel and grabs the tip of the cone (Fig. 38). The lump of glass, which comes with the hook due to the pull, is broken off, and tube is pulled holding a small piece of cloth/sack in hand to a distance of five meters. The broken off glass lump is detached in a U shape and invariably breaks into two horn-shaped parts (Fig. 39). Simultaneously, the chetak is again re-inserted in the lada and kept at place throughout the process. The master drawer places one leg against the comparatively high tunnel platform to stabilize himself and draw the tube continuously (Fig. 40). The thickness and thinness of the tube depend on the master who pulls the tube: the more the speed, the thinner the tube. A slight mistake in pulling the tube makes it to touch the floor, and the moment it touches the floor impurity gets in which blocks the hole leading to mid-joint balloon like swellings. Some times on such occasions, with air pressure the hole opens up.

140

A. K. Kanungo

Fig. 43 Lada and its flakes

At regular intervals, an old master with the help of a kerosene lamp keeps on checking whether the perforation is through or not (Fig. 41). He breaks a small piece from a drawn tube and checks whether the light is passing through. If not, he instructs the workers on the lada to push the chetak a little more for making a complete perforation. One can find such glass rods piled around the furnace area. When broken at a swollen point, this rod gives a knot-like appearance (Fig. 42). Once the glass on lada becomes too small in quantity for the tubes to be drawn, it is taken out and put under the morning sun. Patches of glass left on the lada eventually fall off in broken concentric flake shapes (Fig. 43), whose inner sides have oxidization marks from the lada.

14 Cutting the Tubes With the rising sun, the work of converting the tubes into beads starts. Six to eight workers are seated in a row each holding a dozen tubes in a symmetrical manner and the other hand is held flat in between two small polished bamboo (hand-held) strips. The symmetrical ends of the tubes are laid on a blade driven into the ground, and with another blade tiny pieces are sliced off onto the cloth fixed against it (Fig. 44). The ends of the tubes, about 2–3 cm in length are discarded (as it becomes impossible

Traditional Bead and Bangle Crafts in India

141

Fig. 44 Slicing of tubes

Fig. 45 Re-heating of the segments

to hold them to cut). These pieces are expected in plenty near the cutting area, as they are not passed on further for any other processing. A few people sometimes collect these and sort them into different symmetrical shapes to export to north-east

142

A. K. Kanungo

India where they are profusely used in necklaces and waist griddles. Many a times, moderate knots through which the air has passed are seen at this place.

15 Rounding Operation After the segmentation, the segments are taken to a kiln for reheating. Reheating is done for rounding the sharp edges and giving a glaze to the beads. This kiln is

Fig. 46 Pounding for Rounding and Glazing

Traditional Bead and Bangle Crafts in India

143

generally located close to the main drawing furnace. The kiln which is situated above a two and half feet deep pit, squarish in shape with rounded edges and tapering ends, and has two ports. One is a multi-purpose port for putting the fuel, removing the ash as well as for placing the tray with bead segments and the other is a small port for ventilation. The kiln is heated up first with brush wood, along with a few logs placed at the bottom. During this time, the firing port is covered with inverted half-pots. The top platform of the kiln has a small raised mud boundary for drying the cow-dung powder and to get rid of moisture. The workers pack the cut glass tube segments in the dried cow-dung powder. Pouring a considerable amount of these segments mixed with cow-dung powder on a ceramic tray the master puts the tray inside the kiln with the help of a double-forked wooden pole. Sitting at a distance of three meters he stirs the segments with the help of a bent iron bar for about 20 min (Fig. 45). After allowing the segments to cool, a worker then sieves off all the cow-dung ash. Sometimes segments stick together in this operation and they are discarded. A few examples of beads sticking to the tray are also found thrown at the periphery. Then the segments are placed with rice husk and pounded lightly with a wooden pounder for further rounding and glazing (Fig. 46). This process converts the sharp-edged segments to smooth edged beads. Once again, the beads are sieved to get rid of the husks. Then beads are collected in metal buckets and poured several times from a height to get rid of all husks. The third sieving is done using sieves of different sizes to sort the finished beads in descending sizes (Fig. 47).

Fig. 47 Sieving the beads

144

A. K. Kanungo

16 Stringing the Beads At Papanaidupet, women from almost every household are involved in stringing the beads. This stringing has a very typical Papanaidupet style. Each woman holding more than a dozens of long needles (15 cm in length) in the hand, repeatedly picks up the minute glass beads from a winnowing basket on their laps (Fig. 48). Many a time, they encounter beads whose perforations are too small to pass through the needles. The stringer places an oval-shaped polished stones as a base and break the bead by striking it with the another polished stone (Fig. 49). Many beads are found to have collapsed/ blocked holes and quite a few with holes too small for perforation with the needle, all these along with a few knots which come in the lot are discarded at the stringer’s place. Due to non-specialized and labour-consuming nature of the work, cutting of tubes to segments and stringing of beads takes place in many nearby villages of Papanaidupet too. Thus cutting, sieving and stringing debitage piles up at these Fig. 48 Woman stringing glass beads

Traditional Bead and Bangle Crafts in India

145

Fig. 49 Breaking the stuck bead off the needle

places barring any evidences of production of beads. Once sold/exported unstrung, it is not unlikely that bead debitage is found at the end user’s or the stringer’s place (Kanungo, 2002).

17 Signature of Furnace/Kiln When a furnace or kiln is abandoned, it does not take more than six months for the upper hollow body to collapse and start filling up the pits. In a year’s time, the entire upper structure collapses and the windblown soil and other material fills the pits. In two years’ time, only the pit line of the furnace is visible (Fig. 50). The abandoned kiln vanishes from the scene faster when the furnace material are reused. The stone platforms and basin are lifted to be used either in other furnaces or for any other purposes be it commercial or domestic. The clay is taken for making floors at other places. Due to the vegetation overgrowth at the abandoned site, we do get evidences of only the buried pit lines.

146

A. K. Kanungo

Fig. 50 Furnace after two years of abandonment

18 Observations In South Asia, starting around the mid-2nd millennium BCE till the medieval period, glass was mostly used for the manufacture of ornaments and specifically for beads and bangles. Complex technologies were developed to produce these beads, either from long drawn tubes of glass or by winding the molten glass on a pontil and then altering the shape and design using various moulds or sandwiching techniques or applying glass of different colour over one another, giving the final output a polychrome or a mosaic finish. These beads and bangles were traded locally, and they were traded all over the World through various intermediatories. This is reflected in archaeological records as well as in ethnographic practices. The antiquity of furnace winding beads and bangles in India is 3000 years old, and drawn bead is 2500 years old. The glass crafts in its traditional forms are still practised in western Uttar Pradesh and Chittoor district of Andhra Pradesh. Though the origin of glass working at villages of Purdilnagar, Jalesar, Akrabad and Hasayan (PJAH) is obscure, however, historically it can be traced back to the Sultanate period, c. 1,450 CE. Likewise, the drawn bead making at Papanaidupet can be traced to seventeenth century CE. Archaeological investigations in these villages are much needed. There are many centres of furnace-wound bead production around the world. However, nowhere there are furnaces having as many as nineteen working ports and

Traditional Bead and Bangle Crafts in India

147

yet generate the required temperature to melt the glass, as seen in PJAH. Also, there are no references of multiple-coloured glasses being worked independently at the same time from the same crucible. The inherited furnace-making technology and its plan in PJAH and Papanaidupet appear to be more indigenously developed than many of the other known places. The design and depth of furnace pit plays a crucial role in generating and maintaining the required temperature, which is the prime requirement to work with glass. This observation was possible through ethnographic survey in PJAH and Papanaidupet, a feature, which would have otherwise been difficult to visualize at any archaeological site. Due to its functional properties, the furnaces are broken and made repeatedly in adjacent place, which leads to finding of a number of fired pits in close proximity. The base size/circumference of the pit on which the furnace is made is no way bigger than any tandoor or altar. Thus, care should be taken while interpreting firing places or pits in close proximity to each other in archaeological context. The characteristic glass bead wastes from the archaeological contexts needs to be interpreted scientifically, and conclusive evidence can thus be made about the place where wastes are found, for instance, whether the place was a production centre or a far off production-associated activity area like, place of middlemen, perforation cleaning, cutting, sieving, annealing, glazing, rounding, stringing, glass factory, mould/dye-maker, or an end user. Even if a site has brought to light convincing evidence of bead making, there should be an attempt to suggest which methods (drawing/winding) were employed. The drawing and winding furnaces are different from each other; the required skills, expertise, tools and debitage are specific to these respective furnaces and methods. Bangles are used for ornamentation, symbolic purposes, as well as for offerings. The style and colour composition of bangles are not only controlled by the likings but also by the prevalent traditions. The usage pattern and its relation to socio-cultural milieu, continuously create more broken pieces than intact bangles in cultural deposit. Though use of bangles is must for societal tradition in larger part of south Asia, yet the use is both variable and dynamic in nature. The visibility factor must have played its role in creating this interesting cultural association with bangles. Contextual study of debitage of glass bangle production of different regions is required to understand the origin and spatio-temporal technological evolution of both bangle and furnace along with the understanding of the functional use of the end products.

References Basa, K. (1993). Manufacturing methods of monochrome glass beads in South Asia. Man and Environment, 18(1), 93–100. Brill, R. H. (1999). Chemical analyses of early glasses (Vol. 2). The Corning Museum of Glass. Bronson, B. (1990). Glass and beads at Khuan Lukpad, southern Thailand. In I. C. Glover & E. Glover (Eds.), Southeast Asian archaeology 1986 (pp. 213–229). BAR International Series. Carroll, B. H., & Allen, J. D. (2004). Bead making at Murano and Venice. Beads, 16, 17–37. Chapekar, B. N. (1969). Report on Excavation at Ter 1958. Poona.

148

A. K. Kanungo

Dikshit, M. G. (1952). Beads from Ahichchhatra. Ancient India, 8, 33–63. Dussubieux, L. (2001). L’Apport de l’ablation laser couplée à l’ICP–MS à l’étude du verre archéologique de l’Océan Indien, Ph.D. Dissertation. France: Université d’Orléans. Dussubieux, L., & Gratuze, B. (2003). Non-destructive characterization of glass beads: Application to the study of glass trade between India and Southeast Asia. In 9th International Conference of the European Association of Southeast Asian Archaeologists, May 27th–June 2nd 2002, pp. 135–148. Sigtuna. Dussubieux, L., Kusimba, C. M., Gogte, V. D., Kusimba, S. B., Gratuze, B., & Oka, R. (2008). The trading of ancient glass beads: New analytical data from South Asian and East African soda-alumina glass beads. Archaeometry, 50(5), 797–821. Francis, P. Jr. (1983). Some thoughts on glass bead making. In Proceeding of the 1982 Glass Trade Bead Seminar (pp. 193–202). Rochester Museum and Science Center, Rochester, Research Notes16. Francis, P. Jr. (1986). Bead report, XVIII: the Asian bead study tour, part IV: a little tube of glass. Ornament, 10(1): 54–57, 74–78. Francis, P. Jr. (1990). Glass beads in Asia, part II, Indo-Pacific beads. Asian Perspectives, 29(1), 1–23. Francis, P. Jr. (1992). 20 Easy steps to identifying most beads in most collections. Lapis Route Books. Francis, P. Jr. (1993). West African powder glass beads. Ornament, 16(4), 96–97, 100–101. Francis, P. Jr. (2002). Asia’s maritime bead trade, 300 BC to the present. University of Hawaii. Francis, P. Jr. (2004). Beads & selected small finds. In V. Begley, P. Francis Jr., I. Mahadevan, K. V. Raman, S. E. Sidebotham, K. W. Slane, & E. L. Will (Eds.), The ancient port of Arikamedu: New excavations and researches 1989–1992 (Vol. 2). Ecole Française d D’Extreme-Orient. Gaborieau, M. (1977). Bracelets et Grosses Perles de Verre Fabrication Et Vente En Inde Et Au Népal. Objets Et Mondes Paris, 17(3), 111–130. Gill, M. S. (2017). A single ingredient for primary glass production: Reassessing traditional glass manufacture in Northern India. Journal of Glass Studies, 59, 249–259. Gupta, M. C. (1997). A study of the beads from Kausambi. Swabh Prakashan. Hallifax, C. J. (1892). Monograph on pottery and glass industries of the Punjab, 1890–91. Holland, S., & Holland, T. (2006). Folded glass beads: An Islamic innovation. Ornament, 29(4), 34–37. Kanungo, A. K. (2001a). Glass beads in Indian archaeology: An ethnoarchaeological approach. Deccan College Post Graduate & Research Institute Bulletin, 60–61, 337–353. Kanungo, A. K. (2001b). Glass beads in India: Lamp winding & moulding techniques. Man and Environment, 26(2), 99–108. Kanungo, A. K. (2002). Bondo beads: An ethnoarchaeological approach. South Asian Studies, 18, 121–128. Kanungo, A. K. (2004a). Glass beads in ancient India: An ethnoarchaeological approach. British Archaeological Reports International Series 1242. Kanungo, A. K. (2004b). Glass beads in ancient India and furnace-wound beads at Purdalpur: An ethnoarchaeological approach. Asian Perspective, 43(1), 123–150. Kanungo, A. K. (2006). Glass bead production centres: An ethnoarchaeological assessment. In G. Sengupta, S. Roychoudhury, & S. Som (Eds.), Past and present ethnoarchaeology in India (pp. 411–27). Pragati Publishers & Kolkota: CASTEI, India. Kanungo, A. K. (2010). Antiquity of glass in India: excavations at Kopia. In V. Tripathi & P. Upadhyay (Eds.), Archaeology of the Ganga Basin: Paradigm shift (pp. 451–476). Sharada Publishing House. Kanungo, A. K. (2013). Glass in Ancient India: Excavations at Kopia. Kerala Council for Historical Research. Kanungo, A. K. (2014). Indian glass beads: Archaeology to ethnography. Research India Press. Kanungo, A. K. (2016). Mapping Indo-Pacific Beads vis-à-vis Papanaidupet. Aryan Books International/International Commission on Glass.

Traditional Bead and Bangle Crafts in India

149

Kanungo, A. K. (2019). Chevron and Millefiori in India. Journal of the Borneo International Beads Conference 2019, pp. 69–88. Sarawak. Kanungo, A. K. (2021). Glass in Indian Archaeology, ancient literature, historical records and colonial accounts. In A. K. Kanungo and L. Dussubieux (Eds.), Ancient glass of South Asia— Archaeology, ethnography and global connection. Springer Nature/IIT Gandhinagar. Karklins, K. (1993). The A Speo method of heat rounding drawn glass beads and its archaeological manifestations. Beads, 5, 27–36. Karklins, K., & Jordan, D. (1990). An Early 19th-Century Account of Beadmaking in Murano and Venice. The Bead Forum, 17, 5–8. Kock, J., & Sode, T. (1995). Glass, glassbeads and glassmakers in Northern India. Rosendahl. Kucukerman, O. (1987). Glass beads: Anatolian glass bead making: The final traces of three millennia of glass making in the mediterranean region. Turkish Touring and Automobile Association. Lamb, A. (1965). Some observations on stone and glass beads in early Southeast Asia. Journal of the Malaysian Branch of the Royal Asiatic Society, 38(2), 87–124. Lankton, J. W., & Dussubieux, L. (2006). Early glass in Asian maritime trade: A review and an interpretation of compositional analyses. Journal of Glass Studies, 48, 121–144. Liu, R. K. (1989). Mosaic face beads. Ornament, 12(3), 22–23. Lugay, J. B. (1974). Determination of the Methods of Manufacture of Glass Beads, in Proceedings of the First Regional Seminar on Southeast Asian History and Archaeology, Manila, 1972 (pp. 148– 181). National Museum. Popelka, R. S., Glascock, M. D., Robertshaw, P., & Wood, M. (2005). Laser ablation ICP–MS of African glass trade beads. In R. J. Speakman & H. Neff (Eds.), Laser ablation ICP–MS in archaeological research (pp. 84–93). University of New Mexico Press. Rajan, K., & Kumar Yathees, V. P. (2014). Archaeology of Amaravathi River Valley: Porunthal Excavations. Indira Gandhi Rashtriya Manav Sangrahalaya/Sharada Publishing House. Rajan, K., Yathees Kumar, V. P., Selvakumar, S., Ramesh, R., & Balamurugan, P. (2013). Archaeological Excavations at Porunthal District Dindugul, Tamil Nadu. Man and Environment, 38(2), 62–85. Robertshaw, P., Glascock, M. D., Wood, M., & Popelka, R. S. (2003). Chemical analysis of ancient African glass beads: A very preliminary report. Journal of African Archaeology, 1, 139–146. Ross, L. A., & Pflanz, B. (1989). Bohemian Glass Beadmaking: Translation and Discussion of a 1913 German Technical Article. Beads, 1, 81–94. Sankalia, H. D., Deo, S. B., Ansari, Z. D., & Ehrhardt, S. (1960). From history to prehistory at Nevasa (1954–56). Deccan College Post-Graduate and Research Institute. Sleen, W. G. N. van der. (1973). A handbook on beads. Librarie Halbert. Sode, T., & Kock, J. (2001). Traditional raw glass production in Northern India: The final stage of an ancient technology. Journal of Glass Studies, 43, 155–169. Stern, E. M. (1987a). On the glass industry of Arikamedu (Ancient Podouke). In H. C. Bhardwaj (Ed.), Archaeometry of Glass (pp. 26–36). Indian Ceramic Society. Stern, E. M. (1987b). The secret of Papanaidupet. Glastechnische Berichte, 60, 346–351. Stern, E. M. (1991). Early Roman export glass in India. In V. Begley & R. D. De Puma (Eds.), Rome and India: The ancient sea trade (pp. 113–124). The University of Wisconsin Press. Trivedi, M. (2021). Glass bangles in South Asia: Production, Variability and Historicity. In A. K. Kanungo, & L. Dussubieux (Eds.), Ancient Glass of South Asia—Archaeology, ethnography and global connection. Springer Nature/IIT Gandhinagar.

Scientific Study and Care of Glass

Elemental Compositions and Glass Recipes Laure Dussubieux

Abstract Glass composition has become a key tool for the study of ancient glass. This approach consists of measuring the concentrations of the different elements present in the glass. Information are then used to identify glass types and glass recipes to infer glass technology, provenance, circulation and sometimes, relative dating. Decades ago, only the major and the minor elements (the most abundant elements) were determined, while nowadays, trace elements (elements present in very small concentrations) are also accessible. With the full glass composition, it is possible not only to distinguish different recipes but also to identify the use of different raw material sources among workshops sharing the same recipes. From 1923 to 1933, the Field Museum and Oxford University excavated the site of Kish, located in modern Iraq, 80 km south of Baghdad. Kish is an ancient city occupied as early as 3200 BCE through the seventh century CE. A little more than 400 glass beads were identified in the anthropology collections at the Field Museum with dating, when available, ranging from the Early Dynastic to the post-Sasanian period. The scarcity of the contextual information for most of the beads as well as an often-poor preservation state created challenges for their study. Using laser ablation—inductively coupled plasma—mass spectrometry, a group of beads from Kish with an uncertain chronology and provenance was analysed. Based on the elemental composition of the glass, a second to first century BCE dating was confirmed and a provenance from South or Southeast Asia was proposed.

1 Introduction In archaeology, the study of ancient glass is a field of its own which contributes to a better understanding of the past by addressing a certain number of questions related to three sub-topics: glassmaking, glass trading and the use of glass. As far as glassmaking is concerned, questions deal with the reconstruction of ancient recipes, the ingredient selection and the manufacturing technologies for the production of glass L. Dussubieux (B) Field Museum, Chicago, USA e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_5

153

154

L. Dussubieux

and glass objects, giving an insight into the transfer of ideas and/or craftsmen. Also related to glassmaking is the organization of the glass industries. Their importance and their evolution can be studied in the wider context of the political situation of a given region. Trading is at the centre of many research projects as it informs us directly about interactions and exchange networks with interrogations related to the intensity of the trading and its reach. Finished objects, raw glass, as well as raw materials (e.g. sand or other ingredients), can be tracked to trace the circulation of goods in ancient time. Finally, the use of the finished objects poses questions related to their status and the status of the people using them. With glass being used for the manufacture of personal ornaments, a number of questions are related to the sociocultural significance of wearing certain types of glass objects based on their nature, colour and styles. To address these questions, the comprehensive study of ancient glass has to be truly interdisciplinary and combines a variety of approaches conducted by scholars with very different backgrounds. Archaeologists are at the forefront of this research as they are the ones recovering glass artefacts and recording their contextual information. Their ability to identify indicators of activities related to glass is key to correctly characterizing glass-related workshops but is also a challenging task as those indicators might be difficult to recognize (Fisher, 2008). Ethnography is providing information on glass from communities that kept alive ancient traditions providing clues useful in interpreting archaeological evidence (e.g. Kanungo, 2004, 2016). A great attention is given to typology, which looks at the physical attributes of the glass objects in order to draw meaningful comparison (e.g. Wood, 2011). Studying ancient texts is also a way to recover information about ancient glass even if sometimes the view of the author might distort the reality. Some of those ancient accounts can be tested, as well as hypotheses dealing with glass technology, by recreating ancient processes of manufacturing (e.g. Freestone, 2008; Oppenheim et al., 1970). The use of modern analytical methods to understand the property of the glass and its chemistry has developed widely in the last decades. In recent years, access to the elemental compositions of glass using analytical techniques that are less destructive, more comprehensive and quicker has expanded the possibilities of provenancing and tracking glass along intra- and inter-regional trade routes. It is this last approach that will be most prominently featured in this paper, the study of glass through its elemental composition, as it has become quite an unavoidable tool for the glass researcher, more especially in areas where typology is uninformative, texts are rare and contextual information is scarce.

2 Ancient Glass Recipes Elemental composition is inherently connected to the recipe used by the glassmakers but also to the source of the selected ingredients (e.g. Henderson, 2016; Morretti & Hreglich, 2013; Velde, 2013). In ancient time, glass was mostly a silica-based material obtained from the melting of a sand or crushed silica-rich rocks. Pure silica melts

Elemental Compositions and Glass Recipes

155

around 1700 °C, which was a temperature difficult to reach by ancient craftsmen. To lower this melting point, a flux was added. A flux is an alkali- or alkali-earthrich ingredient that chemically reacts with silica during the fusion of the glass and facilitates its melting. For ancient glass, two main types of flux were available: some have a vegetable nature and were obtained by burning plants or wood containing a high content of soda, potash and/or lime. Other fluxes are taken from mineral deposits such as the natron deposits of Egypt or obtained from saltpetre. Lead has also been used as a flux. Glass must also contain a stabilizer, generally calcium, essential to produce a durable glass that is resistant to water. Calcium can be added as a third ingredient but can also be incorporated indirectly to the glass with the sand or the flux. Although the manufacture of a base glass requires the mixing of two or three ingredients, Indian glassmakers have the particularity of having developed a single ingredient recipe. The raw material used in Indian glass is called reh. Reh is a silica-rich soil containing a natural mix of immature sand with high alumina concentrations and a sodic efflorescence able to produce a vitreous material when heated in a glass furnace (Brill, 2003; Gill, 2017; Kock & Sode, 1995). A last ingredient to consider is added to modify the colour and/or the transparency of the glass. Colouring ingredients, but also decolourizers, ingredients that remove an undesirable natural colour in the glass, as well as opacifiers, can be used. Those ingredients are diverse in nature, but many of them involve transition elements (e.g. iron, copper, cobalt, etc.) and a number of oxides or salts containing lead, antimony, tin…etc. While nowadays the modern glass industry selects fairly pure ingredients, the composition of raw materials entering the glass batch in ancient times is complex. It included not only the major elements of interest for the glass recipe, but also a myriad of other elements, added by accident but present for some of them in significant quantities (Degryse & Shortland, 2019). Also, similar ingredients taken from different sources could feature different trace element patterns reflecting slightly different geological environments. Other factors need to be taken into account, such as recycling, that can impact the composition of glass. An increase of the trace element level is observed as impurities of the crucible enter the glass when it is melted. Glass recycling can also induce low levels of colouring elements in the glass, as pieces of different colours could have been melted together. Thanks to previous research, it is possible to connect a certain number of compositions to glass recipes. This has been a slow process that has depended largely on technological advances to be able to reduce the quantity of sample necessary, increase the number of elements and also lower the concentrations that could be measured.

3 A Few Interesting Facts About the History of Glass Study Martin Heinrich Klaproth, a German chemist, is credited with the first elemental analysis of glass at the end of the eighteenth century CE. He identified and quantified the colouring elements and a few other constituents in three Roman glass mosaic

156

L. Dussubieux

samples from the ruins of the villa of Tiberius in Capri. For his experiment, it is important to notice that 13 g of glass were necessary (Caley, 1949). If during the 19th and the beginning of the twentieth century, a number of elemental analysis of glass are reported, they are isolated, usually focused on the colorants and the material often lacked context (Caley, 1962). Around the middle of the twentieth century, as more data are accumulating, very significant publications start appearing. William Ernest Stephen Turner, a British chemist, published different papers summarizing the state of ancient glass knowledge at that time and laid down the foundation of ancient glass research. Among others, his papers deal with ingredients and their chronology, glass compositions and glass technology (Turner, 1956a, 1956b and 1956c). Sayre and Smith’s article in 1961 (Sayre & Smith, 1961) is one of the first publications that based its conclusion on a large quantity of data to define glass groups. Based on the analysis of 200 glass samples from Europe, the Middle East and Africa dating from the 2nd millennium BCE to the tenth century CE, the authors provided clear parameters to identify five glass groups that are still being relevant today to classify ancient glass (second millennium BCE, Antimony-rich, Roman, Early Islamic and Islamic lead glasses). A turning point in glass study was reached when R.H Brill was hired as a research scientist by the Corning Museum of Glass where he spent 39 years until he retired. Not only, in the course of his carrier, he built a huge database of glass compositions from all over the world that still constitutes the ultimate reference in the field (Brill, 1999—volume 1 & 2; Brill & Stapleton, 2012), but he also provided glass researchers with four glass standards with compositions mimicking that of ancient glasses. Those four glasses (Glasses A, B, C and D) were manufactured by Corning and distributed to a number of laboratories around the world in the middle of the 1960s in the framework of the first round-robin related to the analysis of ancient glass (Brill, 1972). They are still widely used as external standards or controls (Adlington, 2017; Vicenzi et al., 2002; Wagner et al., 2012). Brill also contributed significantly to the study of ancient Indian glass (Brill, 1987, 1999) with compositions from Arikamedu, Kausambi, Hastinapur Brahmagiri and Rupar and more recently Kopia (Kanungo & Brill, 2009; Brill & Fullagar, 2013). He noticed the domination of mineral soda – high alumina glass in this region (Brill, 1987). In India, the earliest analysis of ancient glass seems to have been carried out by Mr Muhammad Sana Ullah. He was an archaeological chemist at the Archaeological Survey in India. Starting in the 1920s, he conducted analysis on glass samples from Taxila (Pakistan), Udaigiri, Kurkshetra and Assam (Varshney, 1950). Other significant publications about Indian glass include Subramanian (1950) reporting data about a blue bead from Arikamedu, Roy and Varshney (1953) with an early study about glass from Kopia, and Varshneya et al. (1988) dealing with Nevasa glass. In 1987, Bhardwaj who already contributed greatly to the advancement of the knowledge in ancient Indian glass with a first major synthesis published in 1979 (Bhardwaj, 1979) offered a more recent overview on this topic in the edited volume of the Proceedings of the Archaeometry session of the XVI International Congress on Glass that took place in India in 1986 (Bhardwaj, 1987). Singh (1989) is another useful reference including scientific data of ancient Indian glass.

Elemental Compositions and Glass Recipes

157

4 Analytical Techniques in General and LA-ICP-MS More Specifically Since the inception of glass studies, a range of analytical approaches have been employed from the most destructive ones to totally non-destructive methods. The number of elements that can be measured varies widely. Ideally, researchers should choose the less destructive method able to analyse the largest range of elements with the lowest detection limits, but options offered to them are often limited. Below is provided a quick description of a number of analytical techniques widely used nowadays including X-ray fluorescence, nuclear activation and optical atomic spectrometry before concluding this section with a description of LA-ICP-MS. Countless studies were conducted using X-ray based methods to obtain the elemental composition of glass (e.g. Chandraratne et al., 2012; Tantrakarn et al., 2012; Varshneya et al., 1988; Zuliskandar et al., 2011). With this method, the atoms at the surface of a glass sample are probed with a beam of X-rays causing electrons to vacate the inner orbital they initially occupied. These electrons are replaced by outer orbital electrons, producing X-ray emissions in the process. Each specific atom emits X-rays with a characteristic energy while the intensity of the emission is related to the concentrations of the atoms in the material. This technique is quick and cost effective. It can be used in a non-destructive way on non-prepared samples, but caution is advised when objects are corroded as it only probes layers of materials close to the surface. More details about this analytical technique applied to the study of glass can be found in Janssens (2013) and Kaiser and Shugar (2012), and this second publication being more specifically focused on hand-held devices. Scanning electron microscopy (SEM) with energy-dispersive spectrometry (EDS) or with wavelength dispersive spectrometry (WDS) is also an X-ray fluorescencebased technique. An SEM directs a focused electron beam to the surface of the sample. The interaction of the electrons with the sample creates various signals (including X-ray fluorescence) used to obtain information about the surface topography of the glass and its composition. When attached to an EDS or WDS detection system, Xrays are collected and used to identify and quantify the atoms in the sample. EDS detection records X-rays of all energies simultaneously and produces a spectrum with X-ray photon energy (x-axis) versus intensity (y-axis). The WDS detection uses a crystal that depending on its position will be able to diffract only photons with a given wavelength at a time. Spectral resolution, sensitivity and reproducibility is better for WDS detection, but EDS is faster and more convenient to use (Verità et al., 1994). In both cases, the imaging capability of the instrument makes it possible to look at the composition of inclusions and other heterogeneities in the glass or to map the distribution of elements over the surface of the sample. SEM-WDS and electron probe micro-analysers or EPMA are fairly similar technologies. EPMA is equipped with a range of crystals in conjunction with WDS detection for a larger range of elements and better sensitivity. It can probe tiny surfaces with diameter smaller than 1 micron. Best results (for EPMA and SEM) are obtained on flat and polished samples usually coated with a conductive material such as carbon. With WDS, the electron beam may

158

L. Dussubieux

generate at the surface of the sample elevated temperatures that can have adverse effects on the composition of the samples. Strategies such as de-focalization of the electron beam or measurements using very short acquisition times are implemented with glass to avoid the migration of an element such as sodium (Henderson, 1988). Neutron activation analysis (NAA) is a bulk technique able to quantify a large range of elements with a sensitivity that can reach the ppb level. It can be totally non-destructive. It has been used quite extensively for the study of ancient glass as exemplified by the work of Hancock in North America (Hancock, 2013; Hancock et al., 1994, 2000) and the work of others in Africa (Davison, 1972), Southeast Asia (Rahman et al., 2008) and elsewhere (Gratuze et al., 1992). With this technique, glass samples are bombarded by neutrons inducing nuclear reactions that are followed by the emission of other neutrons, charged particles or photons. The detection of these particles and the measurement of their characteristic energies is a way to identify the specific nuclear reaction that is taking place and, hence, identify the elements present in the sample. By measuring the intensity of the emitted radiation, the amounts of different elements in the sample can be determined with a high degree of accuracy and precision (Glascock, 2013). Other bulk analytical techniques used with ancient glass are atomic absorption spectroscopy (AAS) and atomic emission spectrometry (AES) also called optical emission spectrometry (OES) that is able to determine the concentrations of trace elements (e.g. Zhang et al., 2005; Ramadan Abd-Allah, 2010). With AES, the sample is atomized, and light is emitted at a wavelength that will be specific of the element. The greater the intensity of the light, the higher the concentration of the element. A monochromator isolates a particular wavelength of light, and a photomultiplier converts the light into an electric current. Increasingly, inductively coupled plasma (ICP) is used as an excitation source. With AAS, a solution is vaporized in a flame. Light of a particular wavelength is passed through the atomic vapour where the atoms absorb some of its energy. The light then passes through the monochromators to select the light of the chosen wavelength. Its intensity is measured by a detector. The amount of light absorbed indicates the quantity of the element present in the original sample (for details, see Lajunen & Perämäki, 2004). AAS is much more versatile than AES and can detect over 70 elements. More and more glass data are now produced using laser ablation—inductively coupled plasma—mass spectrometry (LA-ICP-MS). LA-ICP-MS combines microsampling capabilities with the possibility to measure up to 60 different elements with limits of detection that can reach the ppb level. Its use for glass analysis was developed in the mid-1990s (Gratuze et al., 1993) and since then applications kept multiplying. The minimally destructive approach of this technique allows access to a larger range of artefacts, minimizing bias. As it is a fairly quick technique (30 or more samples can be analysed in a day), larger corpus can be studied insuring better representativeness of the samples. Laser ablation (LA) is used for the direct introduction of solid samples in the ICP-MS. No sample preparation is required. A CCD camera connected to a computer allows the visualization of the surface of the sample. The laser beam interacts with the surface of the objects, vaporizing a small quantity of material that is flushed by a

Elemental Compositions and Glass Recipes

159

gas carrier, which is generally argon or helium. The parameters of ablation depend on the material and on the instrumentation. The diameter of the laser beam ranges from about 50–100 µm. The trace left at the surface of the object by the laser is generally invisible to the naked eye. A pre-ablation is set up to clean the surface of the object in order to remove contamination and to avoid eventual corrosion, which affects the results of the analysis. An ICP-MS is comprised of two main parts: (1) the inductively coupled plasma, which decomposes, atomizes and ionizes the material to be analysed in a very hot environment (temperatures can reach 10,000 K) that is sustained with a radiofrequency electric current produced by an induction coil wrapped around a series of concentric quartz tubes called the plasma torch; (2) the mass spectrometer, which filters the ions according to their charge-on-mass-ratio. To avoid ions colliding with other molecules, the mass spectrometer is kept in a vacuum. An interface ensures the transfer of the ions from the plasma at atmospheric pressure to the mass spectrometer maintained in a low-pressure environment. The difference of pressures between the plasma and the mass spectrometer creates a supersonic jet, and the ions are accelerated as they go through the interface. They are then focused by a series of lenses and mirrors before entering the mass spectrometer. More details are available in Fricker and Günther (2016). At the Field Museum in Chicago, USA, between 2005 and 2015, we used a Varian Inductively Coupled Plasma—Mass Spectrometer (ICP-MS) connected to a New Wave UP213 laser for direct introduction of solid samples. With this instrument, to determine the composition of glass samples with concentrations in the ppm range while leaving no visible damage at the surface of the object, the single-point analysis mode was selected, and the parameters of the ablation were set up as follow: laser beam diameter was 80 µm, pulse frequency was 15 Hz, and 70% of the laser energy was used (0.2 mJ). The gas carrier in the laser is helium for enhanced sensitivity. No sample preparation is necessary, but a pre-ablation time of 20 s is essential to avoid possible surface contamination or corrosion. Concentrations are calculated from four measurements after subtraction of the blank. The reproducibility of the measurements is improved when using an internal standard in order to correct possible instrumental drifts or changes in the ablation efficiency. The isotope Si29 is selected as it is present in high concentrations in all glasses and therefore is accurately measured. The concentration of the internal standard has to be known in order to obtain absolute concentrations. The concentrations for major and minor elements, including silica, are calculated assuming that the sum of their concentrations in weight percent in glass is equal to 100% (Gratuze, 1999). The acquisition of fully quantitative data is achieved by using external standards. They need to have a composition as close as possible to that of the samples. Three different standards with compositions covering the range of compositions found in archaeological glass are analysed several times throughout the day. The first external standard is a standard reference material (SRM) manufactured by NIST (National Institute of Standards and Technology): SRM 610. It is a soda-lime-silica glass doped with trace elements in the range of 500 ppm. Certified values are provided by NIST for a very limited number of elements. Values from Pearce et al. (1997) are used for

160

L. Dussubieux

the other elements. Two additional standards are used. They are Corning B and D and were manufactured by Corning to mimick the composition of ancient glasses (Brill, 1999, vol. 2: 544). With this protocol, the detection limits range from 10 ppb to 1 ppm for most of the elements. Accuracy ranges from 5 to 10% depending on the elements and their concentrations. A more detailed account of the performances of this technique can be found in Dussubieux et al. (2009).

5 Case Study: Glass Beads from Kish, Iraq With this case, the goal is to show how composition including major, minor, and trace elements can help understand better a glass assemblage in terms of chronology and circulation. This section deals with a string of glass beads from the Field Museum collection with the catalogue number 228806 that was excavated at the site of Kish in Iraq (Fig. 1). The site of Kish is located in what is now South-Central Iraq, about 80 km south of modern Bagdad and 12 km east of Babylon. The site, occupied from the Ubaid to the Abbasid period, extends 5 miles in length and 2 miles in width and was divided in the past by the ancient course of the Euphrates that since then has moved westward in what is now its current location. Kish is an important archaeological site as it was one of the earliest cities in the world.

Fig. 1 String of glass beads mixed with beads made from other materials. © The Field Museum, Image No. CL0020_228806_OverallView_XMP2, Cat. No. 228806

Elemental Compositions and Glass Recipes

161

For ten seasons, from 1923 to 1933, the site was excavated by a joint Oxford–Field Museum expedition directed by Stephen Langdon of Oxford University. Langdon spent very little time at Kish after he nearly died during one of his visits but was assisted by two successive field directors. Ernest Mackay was in charge of supervising the field work from 1923 to 1926. He left prematurely to excavate the great Harappan city of Mohenjo-daro in the Indus Valley. Charles Watelin, a French archaeologist, replaced Mackay but unfortunately died shortly after the end of the 10th season of excavation. Artefacts found at Kish during those ten excavations seasons were divided between the Iraq National Museum, the Ashmolean Museum in Oxford and the Field Museum. No final report of the excavations has ever been published. Among the 32,000 objects received by the Field Museum, a little more than 400 individual glass beads were found. Catalogue number 228806 is a string of beads made of stone beads and fourteen beads made of glass (KIB128 to 141). The glass beads are all very similar: they are oblate in shape and manufactured with the drawn method as evidenced by visible lines parallel to the perforation of the beads. The rounded shape of the bead was certainly obtained by heating the beads after cutting small sections from a tube. The beads display a limited range of colours as only dark blue (7 beads) and green (7 beads) are represented. All the beads look opaque. The size of the beads is fairly uniform with diameters ranging from 5 to 8.5 mm. All the beads but one (KIB129) are in fairly good condition with a few beads with no visible trace of weathering at all while for some other beads a whitish or brownish thin layer of corrosion is visible at the surface of the glass. The Field Museum database indicates that this string of beads is dated from the Early Dynastic III period or 2600–2350 BCE. A little tag attached to the string show a similar dating but followed with a question mark. In Langdon (1924: Plate XXV), a picture of the string appears (along with another string of beads) with the legend ‘Strings of Beads from the Sumerian Period’. The Sumerian period ranges from 4500 to 2004 BCE. Immediately, this dating seems questionable as glass production started around the middle of the 2nd millennium BCE (e.g. Rehren & Pusch, 2005; Smirniou & Rehren, 2011; Smirniou et al., 2018; Tite & Shortland, 2003). Before that, glass findings are fairly rare. The beads were found in an area where was discovered a ‘Sumerian Palace’; however, this area included also later graves, dating from the Greek Period that Mackay (1929) described a few years later. They consist of an enclosure measuring 8.40 × 5 m with mud-coated walls 30 to 50 cm-high. The floor was covered with bricks taken from buildings from earlier periods. Bones belonging to six bodies stacked on top of each other were found in the northern part of the chamber (Mackay, 1929: 118). Objects found in the grave include a bronze ring and various items made from clay sometimes glazed (lamps, a vase, dishes). A ceramic figurine, 26 cm in height, was also found in the grave. It is described as ‘a nude female with an elaborate headdress’. It was manufactured using a mould made of two pieces and was certainly originally covered in painted stucco. A second figurine was placed in the grave. It was made of alabaster and represents a woman in a recumbent posture ‘common to the period’ but was missing its head. Last but not least are mentioned ‘two string of beads (that) include ivory, glaze, glass, limestone and quartz (Reg. No 796; Field. 797; Bagdad)’

162

L. Dussubieux

(Mackay, 1929: 203–204). The Field Museum records confirm that the objects with the registration number 796 and catalogue number 228806 are the same. In the microfiches provided with Moorey’s publication (Moorey, 1978), objects with Reg. No. 796 appear under ‘Grave O: Seleucid/Parthian (no registered grave number). This burial vault of the Seleucid/Parthian period, probably of the second or first century BC,…’ giving us a dating concurring with Mackay’s context and a priori more likely than an Early Dynastic one.

6 Results The composition of only 13 beads will be discussed as the analysis carried out on KIB129, with high silica concentration (71%) and low alkali content (Na2 O + K2 O < 4%) revealed that corroded material was sampled. Results reported in weight percent of oxide (%) for the major and minor elements and parts-per-million (ppm) for the trace elements are available in Table 1. The thirteen reminding beads have very similar compositions. To discuss the general recipe used for these glass beads, the reduced composition of the beads, including SiO2 , MgO, Al2 O3 , K2 O, CaO and Fe2 O3 was calculated (Table 2) to only take into account the elements brought to the glass by the silica source and the flux according to Brill (1999). The glass used to manufacture the beads found at the Kish site has a composition rich in soda (11.7 ± 1.6%) and lime (5.4 ± 0.8%). They are the most abundant constituents in the glass after silica. Magnesia and potash, with, respectively, 0.5 ± 0.1% and 1.3 ± 0.2%, have fairly low concentrations, usually found in glasses manufactured with soda taken from a mineral source. To find comparative material with a composition close to that of the Kish glass beads, it is necessary to look in different directions. Iraq was part of Mesopotamia and later, of the Sasanian Empire, where flourished centuries after centuries glassmaking centres using soda plant ash as a flux. Iraq is adjacent on its western part to the Syro-Palestinian region well known for its natron-based glass production during the eighth century BCE–eighth century CE period. Access to glass further east was possible through the terrestrial Silk Road that was running through Baghdad. Access to the Persian Gulf and the Indian Ocean was also facilitated by the proximity of the Euphrates that at one point run through Kish before being displaced westward away from the city. From a general point of view, the soda plant ash glass composition produced in the Middle East is a soda-lime glass that can be characterized by potash and magnesia concentrations higher than 1.5% (Sayre and Smith 1961; Henderson et al., 2004). The glass beads from Kish have lower concentrations for both these constituents indicating rather the use of a soda-rich flux with a mineral origin such as natron. Natron glass from the Syro-Palestinian region has also rather high lime (>5%) and rather low-alumina (3700–1900 BCE): Technological and experimental studies of production and variation. In A. K. Kanungo & L. Dussubieux (Eds.). Ancient Glass of South Asia—Archaeology, Ethnography and Global Connection. Singapore/Gandhinagar: Springer Nature/IIT Gandhinagar. Kock, J., & Sode, T. (1995). Glass, glass beads and glassmakers in Northern India. Vanlose: THOT Print. Lad, G. (1983). Mahabharata and archaeological evidence. Poona: Deccan College Post-Graduate and Research Institute. Lal, B. B. (1954–55). Excavation at Hastinapur and other explorations in the Upper Ganga and Sultej Basins 1950–52: New lights on the dark age between the end of the Harappa Culture and the Early Historic Period. Ancient India, 10–11, 5–151. Lal, B. B. (1987). Glass technology in Early India. In H. C. Bharadwaj (Ed.), Archaeometry of glass (pp. 44–56). Central Glass and Ceramic Research Institute.

256

A. K. Kanungo

Lucas, A. (1926). Ancient Egyptian materials and industries. Edward Arnold and Co. Mackay, E. J. H. (1931). Personal ornaments. In J. H. Marshall (Ed.). Mohenjodaro and the Indus Civilization (Vol. II, pp. 509–548). London: Arthur Probsthain. Margabandhu, C. (1971). Material culture of Central and Western India and the Deccan from circa the third century B.C. to the third century A.D. based on the evidence of excavated remains. Ph.D. Dissertation. Nagpur: Nagpur University. Margabandhu, C. (1975). Aspects of metal technology during the early historic period. Indica, 12(2), 71–76. Margabandhu, C. (1985). Archaeology of the Satavahana Kshatrapa Times. Delhi: Sundeep Prakashan. Marshall, J. H. (1911–12). Excavations at Bhita. Archaeological Survey of India annual reports (pp. 29–94). Marshall, J. H. (1951). Taxila, 3 Vols. Cambridge University Press. McCarthy, B., & Vandiver, P. (1991). Ancient high strength ceramics: At Harappa (Pakistan). c. 2300–1800 B.C. In Materials Research Society Symposium Proceedings (Vol. 185, pp. 496–509). Pittsburgh. Mehta, R. N., & Shah, D. R. (1968). Excavation at Shamalaji. Maharaja Siyajiroa University of Baroda. Mill, J. (1826). The history of British India (Vol. II). Baldwin, Cradock & Joy. Mishra, A., Singh, D., & Sharma, A. (2010). A study of glass bangles from Abhaipur, District Pilibhit Uttar Pradesh. Man and Environment, 35(1), 103–108. Mukharji, T. N. (1895). A monograph on the pottery and glassware in Bengal. Narain, A. K., & Roy, T. N. (1976). Excavations at Rajghat (Vol. III). Varanasi. Rajan, K. (2009). Note on Porunthal Excavations. Avanam: Journal of Tamil Nadu Archaeological Society, 20, 109–115. Rajan, K., & Yathees Kumar, V. P. (2014). Archaeology of Amaravathi River Valley: Porunthal excavations. Indira Gandhi Rashtriya Manav Sangrahalaya/Sharada Publishing House. Rajan, K., Yathees Kumar, V. P., Selvakumar, S., Ramesh, R., & Balamurugan, P. (2013). Archaeological excavations at Porunthal District Dindugul, Tamil Nadu. Man and Environment, 38(2), 62–85. Ramachandran, K. S. (1980). Archaeology of South India—Tamil Nadu. New Delhi: Sandeep Prakashan. Raman, K. V. (1991). Further evidence of Roman trade from coastal sites in Tamil Nadu. In V. Begley & R. Daniel (Eds.), Rome and India: The Ancient Sea Trade (pp. 125–133). University of Wisconsin Press. Rogers, A. (1900). Note on the presentation by J. A. Baines on ‘The Industrial Development of India’ (pp. 569–588). Journal of the Asiatic Society for Arts, 48, 584–585. Roy, P., & Varshney, Y. P. (1953). Ancient Kopia glass. Glass Industry, 34, 366–392. Russell, F. (1793). A short history of the East India Company. London: Sewell and Debrett. Sankalia, H. D. (1966). Tripuri excavations–1966–Short Report & Exhibition. Singhai Press. Sankalia, H. D., & Dikshit, M. G. (1952). Excavations at Brahmapuri (Kolhapur) 1945–46. Poona: Deccan College Post-Graduate and Research Institute. Sankalia, H. D., Subbarao, B., & Deo, S. B. (1958). The excavations at Maheshwar and Navdatoli: 1952–53. Deccan College Post-Graduate and Research Institute/Maharaja Sayajirao University of Baroda. Sankalia, H. D., Deo, S. B., & Ansari, Z. D. (1969). Excavations at Ahar (Timbavati). Pune: Deccan College Post-Graduate & Research Institute. Sankalia, H. D., Deo, S. B., Ansari, Z. D., & Ehrhardt, S. (1960). From history to prehistory at Nevasa (1954–56). Poona: Deccan College Post-Graduate and Research Institute. Sastri, I. C. (1917). Yukti-Kalpatary. Siddheswar Machine Press. Selvakumar, V. (2021). History of glass ornaments in Tamil Nadu, South India: Cultural perspectives. In Kanungo, A. K., & Dussubieux, L. (eds.) Ancient glass of South Asia - archaeology, ethnography and global connection. Springer Nature/IIT Gandhinagar.

Glass in Indian Archaeology, Ancient Literature, Historical …

257

Sen, S. N., & Chaudhuri, M. (1985). Ancient glass and India. New Delhi: Indian National Science Academy. Sharma, D. V., Nauriyal, K. C., & Prabhakar, V. N. (2006). Excavations at Sanauli 2006: A Harappan Necropolis in the Upper Ganga-Yamuna Doab. Puratattva, 36, 166–179. Sharma, G. R. (1969). Excavations at Kausambi (1949–50). Memoirs of the Archaeological Survey of India 74. New Delhi: Archaeological Survey of India. Singh, A. K. (1983). Pottery and trade—A study of Roman pottery from Indian sites. Journal of the Bihar Research Society, 53–54, 140–174. Singh, P. L., Upadhyay, P., Kumar, M., Kumar, A., Shukla, D. K., Gupt, C. B., Singh, U., & Afroj, M. (2018). Preliminary report on excavations at Sarethi, District Firozabad, Uttar Pradesh. Man and Environment, 43(1), 66–78. Singh, R. N. (1989). Ancient Indian Glass: Archaeology and technology. Parimal Publication. Singh, R. N., & Abdurazakov, A. A. (1990). Glasses from Khairadih: Some chemical consideration. International Symposium on Archaeometry Abstracts, 2–6 April 1990. Heidelberg. Sofianopoulas, A. J. (1952). Primordial glasses. Journal of Chemical Education, 29, 503–504. Srinivasan, K. R., & Banerjee, N. R. (1953). Survey of South Indian Megaliths. Ancient India, 9, 103–115. Stern, E. M. (1987). On the glass industry of Arikamedu (Ancient Podouke). In H. C. Bhardwaj (Ed.), Archaeometry of glass (pp. 26–36). Indian Ceramic Society. Stern, E. M. (1991). Early Roman export glass in India. In V. Begley & R. D. De Puma (Eds.), Rome and India: The ancient Sea trade (pp. 113–124). The University of Wisconsin Press. Stone, J. F., & Thomas, L. C. (1956). The use and distribution of faience in ancient east and prehistoric Europe. In Proceeding of Prehistoric Society (Vol. 22). Tewari, R., Hemraj, & Srivastava, R. K. (1995–96). Excavations at Hulaskhera, Lucknow: 1978–79 to 1986–87 (A Preliminary Report), Pragdhara, 6, 95–113. Thapar, B. K. (1957). Maski 1954: A Chalcolithic site of the Southern Deccan, Ancient India, 13. Tite, M. S., Freestone, I. C., & Bimson, M. (1983). Egyptian faience: An investigation of the methods of production. Archaeometry, 25(1), 17–27. Trivedi, R. K. (1967). Selected crafts of Gujarat. Census of India, 1961: V, part VII-A. Delhi. Upadhyay, P. (2017). Recent archaeological investigation in Karamnasa valley with special reference to excavations at Latif Shah: 2016. Bharati, 40, 53–75. Varshney, A. K., Stephen Tong, S. C., & Gogte, V. D. (1988). Analysis of early glass objects from Nevasa excavations in India. Transactions of the Indian Ceramic Society, 47(5), 149–155. Vikrama, B., & Pradhan, A. (2017). Implications of a recent hoard of copper objects from Harinagar District Bijnor, Uttar Pradesh. Puratattva, 47, 73–82. Watt, G. (1889). A dictionary of the economic products of India (Vol. I). Superintendent of Government Printing.

Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts Bhuvan Vikrama

Abstract The early glass in India has been generally dated around 1000 BCE. In the world context, the date of glass production is believed to be a Bronze Age speciality between 2500 and 1500 BCE. In the South Asian context, there has not yet been any convincing evidence of glass crafts in either urban Harappan or rural Chalcolithic cultures; however, fine quality of faience has been a regular find during that period. Muhammad Sana Ullah who first studied the faience from Harappan sites while explaining its fine quality expressed ‘that it is of the nature of glass’. Contrary to the present knowledge, this chapter argues that there existed a possible knowledge of glass crafts in India from at least 2000 BCE on the basis of the evidence revealed at recent excavations at Harinagar. Although the evidence is scanty, the chapter suggests that the context and chronology established firmly through AMS dating support the claim.

1 Introduction Among the materials used by humans and encountered by archaeologists, glass is one of the man-made materials which requires a high degree of skill and elaborate technology in its manufacture. The earliest known glass is believed to have been created in the Middle East several millennia ago (Kurkjian & Prindle, 1998: 797), and by the Late Bronze Age, the glass industry was established in Egypt and Mesopotamia (Rehren, 2021). The culture contact between the civilizations of Indus Valley and the Mesopotamia is well established. Item of luxury and ornaments were integral part of trade contact; however, Indus Valley Civilization has not yet reported any evidence of glass. Faience and glazes have been reported in both time and space of Harappan horizons. The prelude to the second urbanization in India started with introduction of Iron, and that is the time when glass was found across Northern India.

B. Vikrama (B) Archaeological Survey of India, Guwahati, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_10

259

260

B. Vikrama

2 Pre-Iron Age Though the iron making and iron working were prevalent in the Central Ganga Plain and the Eastern Vindhyas from the early second millennium BCE in association with pottery comparable to those generally considered as the characteristics of the Chalcolithic period (Tiwari, 2003: 543), culturally the onset of Iron Age is generally taken to begin from Painted Grey Ware (PGW) culture. However, to translate it into chronological terms may pose difficulties as the origin and spread or adoption of iron technology has not been uniform. This goes even with the Chalcolithic cultures too. If the entire period of human evolution is classified into material or technology-based ages, the Stone Age with all its variants lasted till 5000 BCE when the Chalcolithic Age was ushered in and continued till the beginning of the Iron Age about 1200–1000 BCE. Therefore, Pre-Iron Age is the Chalcolithic Age, a period before 1000 BCE. In India, the antiquity of glass has been traced to cultural period dominated by the PGW, which is considered both as copper using culture and as an Iron Age culture and tentatively dated 1200–800 BCE. Only a few Pre-Iron Age/Chalcolithic sites have yielded glass.

2.1 Bhagwanpura Bhagwanpura a Late Harappan-PGW overlap site has brought forward many specimens of faience and glass objects from the PGW levels. It is noteworthy, here, that PGW levels have not yielded any iron objects. From the overlap phase, i.e. Pd. IB, it yielded two beads of glass—one a long convex eye bead and the other a segmented eye bead. Twenty-seven faience beads were also recovered from this level. Similarly, the same period has also yielded bangles of glass as well as faience. All nine fragments of glass bangles and 78 portions of faience bangles have been found from sub-period IB. Glass bangles from this level are a remarkable find as it is by far the earliest finds. Glass bangles are found in blue, black and white in colour, and a few of them are polychromed with silver and golden colour. There are findings of translucent to semi-transparent and blue, black and green colour glasses at site too. Interesting is to note that though the faience marks its presence from the subperiod IA, its frequencies increase many folds in the sub-period IB whereas glass only enters in sub-period IB (Joshi, 1993: 117–125).

2.2 Maski Period I at Maski (Lingsugur Taluk of Raichur district in Karnataka) representing Chalcolithic phase at the site yielded four glass beads. Not much detail is available in the report; however, # 7 illustration of Pl. XXVII describes the bead as ‘Glass,

Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts

261

blue opaque, standard barrel circular from a late level of Period I (MSK-10, 1082)’ (Thapar, 1957: 108).

2.3 Balathal Balathal in Udaipur district of Rajasthan has yielded eight glass beads from layers 6, 11, 13, 14 and 16 all belonging to the Chalcolithic levels (Kanungo et al., 2007). Chalcolithic levels at Balathal also yielded seven beads of faience from layers 6, 7, 15, 16 and 17. Although the beads came from the Chalcolithic level, morphological features and the locus of finds created doubt, and the excavator expressed a possibility of beads percolating down from early historic deposits above. Similar doubts were cast on the finds of glass beads from Chalcolithic phase at Ahar, where due to disturbance on the top-level glass beads findings were considered to be not from the Chalcolithic phase (Sankalia et al., 1969). It was but obvious for the excavators to have doubt on Chalcolithic level glass finds, since neither the finds were in numbers and varieties (or with any glass chunks), nor there were antecedent finds from other sites or glass was reported from a respectable number of Chalcolithic sites. It is important to mention that in both cases excavators doubted the contexts of glass finds but not of faience. The presence of faience is acceptable as it has been in existence from Harappan sites in the form of beads from the early levels. Late twentieth-century excavations at Harappa have uncovered a settlement dating to 3300 BCE (Meadow & Kenoyer, 2001). The faience ornaments, figurines and vessels are more prevalent during urban phase of the Harappan Period, 2600–1900 BCE (Dales & Kenoyer, 1991; Kenoyer, 1991, 1994). Faience is seen as a precursor to or a proto glass as it is also one of the man-made vitreous products, production of which requires high temperature and melting of quartz.

3 The Vitreous Product Vitreous products are basically silica-based. For fusing, silica temperature of about 1700 °C is required; however, addition of sodium carbonate (Na2 CO3 , soda) lowers the glass transition temperature. But soda makes the glass water soluble; thus, a stabilizer such as lime (CaO, calcium oxide) is added to provide a better chemical durability. There are four vitreous products—glass, glaze, enamel and faience which consist of silica, alkali metal oxides and lime. Glass, glaze and enamel contain a large quantity of soda or other alkali metal oxide, whereas faience contains only a small quantity of alkali metal oxide.

262

B. Vikrama

3.1 Faience In the Indus Valley, silica-based glazing and faience started being produced around 2600 BCE. Faience is made from ground quartz which is melted and ground again to make a frit or glassy paste. This material can be coloured with copper to make a bluegreen or turquoise colour and then fired at high temperature to create a shiny glazed object. The valuable ornaments and decorations created from steatite and faience were used by Indus people as indicator of status and wealth (Kenoyer, 2005: 43). Indus Valley artisans were exclusive in utilizing efflorescence, a technique, wherein the colour of glaze and interior core are identical and the glaze is strongly bonded with underlying body (Kenoyer, 1994: 36). The first scientific study on faience from Mohenjo-daro was conducted by Md. Sana Ullah, who describes it as having ‘a hard, fine granular body, covered with glaze. The prevailing colours are bluish-green and greenish-blue, although white, chocolate and red specimens have also been found. The microscopic examination reveals a compact granular structure, comprised of angular quartz grains bound together with a transparent cement. Its chemical analysis shows also that silica is the chief constituent, forming about 90 percent of the total amount. From these facts it may be inferred that the original paste was composed of finely crushed quartz or pure white sand, a glassy flux and a colouring matter, if necessary…. Possibly, silicates of soda which forms a highly viscous solution with, was employed as a constituent of the paste and served to impart the desired property to the wet paste’ (Ullah, 1931: 686–687). He further records that ‘the glaze has perished mostly through decomposition and the material available is scanty for a complete chemical analysis. However, judging from its transparency, the nature of colouring matter, and the iridescent films on these objects, one can easily conclude that it is of the nature of glass’ (Ullah, 1931: 687). However close to glass it appeared, it was not glass. Glass is a transparent, hard substance suitable for forming objects like beads, vessels, window panes, and many other such articles. It is thought by some that the first glass was probably developed in the Mitannian or Hurrian region of Mesopotamia, possibly as an extension of the production of glazes (Rasmussen, 2012: 11).

3.2 Glaze, Glass and Enamel Glaze is a thin vitreous coating applied to another material to make it impermeable or to produce a shiny decorative appearance. Enamel resembles a glaze in that it is also fused to a body of a different material, usually, metal. The term glass is commonly applied to the transparent, brittle material used to form windows, vessels and many other objects. More accurately, glass refers to a state of matter with a disordered chemical structure, i.e. non-crystalline such as silicate glasses, which are inorganic products of fusion, cooled to a rigid condition without crystallizing. It requires hot working, while faience requires cold working. Chemically, glass, glaze and enamel

Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts

263

can be identical in composition, the fundamental difference being their method of use in antiquity. Glaze has been accepted at Harappan sites which looked like glass. Vitreous slabs have been reported from Mohenjo-daro which are described as having ‘a rounded upper surface; but its base is flat and it was evidently poured out or lifted when hot, on to a wire (?) screen, whose impression is clearly seen. Contraction would have separated the slab from the metal as it cooled’ (Marshall, 1931: 575). But no case of enamelling has been found from these sites. The production of faience involves only cold working and reduced temperature sintering of the raw materials. In contrast, the routine production of glass vessels and other objects involves the manipulation of hot, viscous fluids, a process that was more akin to metal working. The change from cold working for glazes and faience to hot working for glass may not have been a logical progression or an easy transition. Such a transition would most likely have required input from metal workers who were more familiar with such high-temperature manipulations. Thus, it can be argued that the discovery of the techniques necessary for hot-working glass was the result of interaction between the workers of glazed stone and faience and metal workers (Rasmussen, 2012: 18). Industrial production of glass vessels and other objects was of course late as the blow pipe technique was discovered only in late centuries of BCE. However, glazing and enamelling, which required cold working and sintering, were definitely possible but no such case is encountered beyond 1500 BCE.

4 Harinagar Evidence Recently, a large collection of copper objects including 55 vessels—weighing about 44.8 kg—along with other implements and weapons was discovered stashed in a large cauldron of copper during the levelling of agricultural field in the village of Harinagar (29° 11 17 N; 78° 10 28.1 E) in Tehsil Chandpur, district Bijnor, Uttar Pradesh (Vikrama & Pradhan, 2017, 2018; Vikrama & Vikrama, 2019). Typologically, contents of the big cauldron have parallels in the material exhumed from signature Harappan sites like Harappa, Mohenjo-daro, Chanhudaro, Lothal, etc. Finds from Harinagar are exciting on many accounts: first—as it is the largest hoard of copper objects (within or outside Indus Valley), second—it has yielded the largest number of copper vessels, i.e. 55, third—preserved in almost pristine condition after the last use as they remained in closed and packed environment inside the cauldron, barring the breakage due to forced extraction of pots by the locals, fourth—collection included a square section stiletto sword and a shouldered axe with four drill holes at the tang for fixing the handle, the shapes have not been encountered so far anywhere in Harappan domain, fifth—the soot from two of the pots has been dated to 2300–2200 BCE, and sixth—interestingly, on some vessels, apart from patina, smooth patches of glassy material of greenish hue are noticed sticking on the interior as well as exterior surfaces which are definitely not part of

264

B. Vikrama

patina layer forming naturally due to oxidization, rather appear more like a result of deliberate human effort (Fig. 1). The suspected glass on the pots raised a possibility that the manufacturer might have used enamel on the copper pots. Broken edges of the pots showed uniform line of greenish enamel/glassy layer on either side of the metal (Fig. 2). Many argue this to be as malachite, a corrosive by-product, which normally occurs in copper object of antiquity. However, the presence of layered soot on the cooking vessels which overlay the green layer (Fig. 3) enforced the possibility of an applied layer on the

Fig. 1 Green layer

Fig. 2 Uniform green line on either side of metal section

Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts

265

Fig. 3 Handi with soot and green layer underneath

metal surface of the pots. The green glassy layer shows a web of surface cracks (Fig. 4) which can be compared with cracks of porcelain and enamelled metal ware of modern times. On many pots, traces of bluish-green glossy surface can be seen in different degrees of decomposition (Fig. 5), and it appears that the glassy enamel has worn off due to decomposition or deterioration or devitrification. As the pots were buried in soil inside a big cauldron, soil got stuck to these pots. In some pots, where the layer of soil has come off (perhaps due to forced extraction), exposed layer shows green

Fig. 4 Web of surface cracks—crizzled surface

266

Fig. 5 Enamelled surface deteriorated to various degrees

Fig. 6 Green to pale green specs of crystalline appearance

B. Vikrama

Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts

267

Fig. 7 a Stuck pots due to long contact after separation and b crystalline formation at sandwiched portion

to pale green specs which appear to be of crystalline nature (Fig. 6). Some pots got stuck together due to decomposition of the glass layer while in long contact (Fig. 7a), and the separation peeled the layer off to show the crystalline formation between the two layers of metal (Fig. 7b). At places, rough (irregular grooved) metal surface is visible where from the glass layer it was removed; however, the deteriorated glass layer can be seen with lamellar structure overlaid with a crust of soil containing green

268

B. Vikrama

Fig. 8 Glass layer with lamellar structure overlaid with a crust of soil containing green and blue crystalline particles

and blue crystalline particles coming out of decomposed/devitrified glass (Fig. 8). At places where the finished glass surface survives, scratches appear as use marks (Fig. 9). Devitrification in the archaeological sense is a natural process that chemically degrades siliceous materials in the archaeological record (Romich, 2006: 164). Devitrification occurs when the surface of the glass becomes crystalline through the absorption of moisture from the environment. Deterioration process of glass is very complex involving diffusion of hydrogen ions from water within the glass network which is initiated by the moisture present in the immediate environment. Hydrogen ions (H+) from the water molecule displace the sodium (Na+) or potassium (K+) ions from the network, leaving a hydroxide ion (OH-). The by-product of this displacement is sodium or potassium hydroxide, both highly reactive alkalis. Alkaline corrosion products react with atmospheric gases to form sulphate and carbonate crystals. The net result is a gradual deterioration of the glass, usually accompanied by the simultaneous formation of a crust, normally composed of carbonates and sulphates and other opaque weathering products (Anderson, 1996; Griffiths & Feuerbach, 2001). A series of fine microcracks start to become visible (see Fig. 4). This stage is known as ‘crizzling’ and can eventually lead to the formation of flakes and pits on the surface (Fearn, 2002). Weathering layers are lamellar structures (see Fig. 8) within the glass structure with high crystallinity. There is often staining around fissures and milkiness. Weathering layers may be formed by leaching followed by the formation of heterogenous structures in leached layer (Anaf, 2010) which ultimately break away from the substrate and restart the cycle at the new surface. These layers may be transformed into a crystalline end product (see Figs. 7 and 8) (Griffiths & Feuerbach, 2001).

Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts

269

Fig. 9 Finished enamel surface in yellowish green and newly exposed surface in bluish-green with use marks

It is interesting to note that the surface below the glassy layer appears rugged and rough, which could be a deliberate effort to provide a suitable grip to the enamel as well as to compensate for the thermal expansion of the two materials (see Fig. 8). Gold and copper are the most suitable metals to receive and retain enamel. For enamelling, metal is formed into shape, then the surface is cleaned and roughened for better adhesion, and enamel powder or paste is then applied and fused to the metal in furnace at a temperature approximately ranging between 750 and 850 °C (Muros et al., 2012).

270

B. Vikrama

5 Conclusion In visual examination, the specimen from Harinagar qualifies as enamelled pots. However, further scientific investigations would be necessary (e.g. using scanning electron microscopy) to characterize fully this vitreous layer adhering to metal and confirm its nature. As of today, we do not have any concrete evidence to suggest pure glass objects were known to Indians before 1400 BCE, but the high-quality enameling on 2300 BCE Harinagar samples suggests that Indians had mastered this craft. Harinagar objects amply show different stages of deterioration of glassy surface and also bring forth other signs indicating sophistication and mastery over the enamel production. These finds cannot be treated as the only experimentation by the makers as no other site has produced similar evidence. The absence of evidence from other sites may be attributed to: 1. 2. 3. 4.

Circumstance of burial; Excavation process; Urge to chemically clean the objects for better preservation; and Lack of information on the decomposition or devitrification of glass.

Harinagar finds could retain such important evidences as all the objects were stored inside a large cauldron, which was buried in silt. A situation where internal environment was free from frequent changes. Further, the finds were not subjected to any kind of wet cleaning. Future excavation should take care while exposing metal objects, and excavator should also collect the immediate soil covering the vessels as it may contain some evidence of decomposed glass. The finds of Harinagar in the background of findings of glazed pottery at number of Harappan sites, stray glass find at Chalcolithic Balathal, Ahar and Maski, and late Harappan-PGW overlap phase at Bhagwanpura bring us to start looking for possibility of knowledge of glass in early 2nd millennium BCE India rather than concluding the glass in India is a 1st millennium BCE phenomenon.

References Anaf, W. (2010). Study on the formation of heterogeneous structures in leached layers during the corrosion process of glass. In CeROArt (Online) http://journals.openedition.org/ceroart/1561. https://doi.org/10.4000/ceroart.1561 Anderson, D. (1996). Stained glass and its decay. In The Conservation and Repair of Ecclesiastical Buildings. http://www.buildingconservation.com/articles/glassdecay/glassdecay.htm Dales, G. F., & Kenoyer, J. M. (1991). Summaries of five seasons of research at Harappa (District Sahiwal, Punjab, Pakistan) 1986–90. In R. H. Meadow (Ed.), Harappa excavation 1086–90 (pp. 185–262). Prehistory Press. Fearn, S. (2002). Continued studies in the deterioration of glass. Conservation Journal, 42. http://www.vam.ac.uk/content/journals/conservation-journal/issue-42/continued-studies-inthe-deterioration-of-glass/

Situating Harinagar Hoard Finds in Pre-Iron Age Glass Crafts

271

Griffiths, D. R., & Feuerbach, A. M. (2001). The Conservation of wet medieval window glass: A test using an ethanol and acetone mixed solvent system. Journal of the American Institute for Conservation, 40(2), 125–136. Joshi, J. P. (1993). Excavation at Bhagwanpura 1975–76: and other Explorations & Excavations 1975–81 in Haryana, Jammu & Kashmir and Punjab. New Delhi: Archaeological Survey of India. Kanungo, A. K., Mishra, V. N., & Shinde, V. S. (2007). Western Indian (Mewar) Chalcolithic Beads with Special Reference to Balathal. Beads: Journal of Society of Bead Researchers, 19, 42–57. Kenoyer, J. M. (2005). Steatite and Faience manufacturing at Harappa: New evidence from Mound E excavations 2000–2001. Museum Journal: National Museum of Pakistan, Karachi, 3–4, 43–56. Kenoyer, J. M. (1994). Faience from the Indus Valley Civilization. Ornament, 17(3), 36–39 & 95. Kenoyer, J. M. (1991). The Indus Valley tradition of Pakistan and Western India. Journal of World Prehistory, 5(4), 331–385. Kurkjian, C. R., & Prindle, W. R. (1998). Perspectives on the history of glass composition. Journal of American Ceramic Society, 81(4), 795–813. Marshall, J. (1931). Mohenjo-Daro and the Indus Civilization. A. Probsthain. Meadow, R. H., & Kenoyer, J. M. (2001). Recent discoveries and highlights from excavations at Harappa: 1998–2000. INDO-KOKO-KENKYU (indian Archaeological Studies), 22, 19–36. Muros, V., Gleeson, M., Dolph, B., Griswold, G., Mahony, C., North, A., Tzadik, C., & Fox, K. (2012). Glass. http://www.conservation-wiki.com/wiki/Glass#ref27 Rasmussen, S. C. (2012). How glass changed the world: The history and chemistry of glass from antiquity to the 13th century. SpringerLink. https://doi.org/10.1007/978-3-642-28183-9_2 Rehren, T. (2021). The origin of glass and the first glass industries. In A. K. Kanungo & L. Dussubieux (Eds.), Ancient Glass of South Asia—Archaeology, ethnography and global connection. Singapore/Gandhinagar: Springer Nature/IIT Gandhinagar. Romich, H. (2006). Glass and ceramics. In E. May & M. Jones (Eds.), Conservation science: Heritage materials (pp. 160–184). RSC Publishing. Sankalia, H. D., Deo, S. B., & Ansari, Z. D. (1969). Excavations at Ahar (Timbavati). Pune: Deccan College Post-Graduate & Research Institute. Thapar, B. K. (1957). Maski 1954: A chalcolithic site of Southern Deccan. Ancient India, 13, 4–142. Tiwari, R. (2003). The origins of iron-working in India: New evidence from the Central Ganga Plain and the Eastern Vindhyas. Antiquity, 77(297), 536–544. Ullah, K. B. M. S. (1931). Appendix I. Notes and analyses. In J. Marshall (Ed.), Mohenjo-Daro and the Indus Civilization (pp. 686–690). London: A. Probsthain. Vikrama, B., & Pradhan, A. (2018). Copper vessels in the Indus Valley Civilization: A case study in the light of Harinagar Hoard. Indian Journal of History of Science, 53(3), 271–278. Vikrama, B., & Pradhan, A. (2017). Implications of a recent Hoard of Copper objects from Harinagar District Bijnor, Uttar Pradesh. Puratattva, 47, 73–82. Vikrama, B., & Vikrama, S. M. (2019). Emerging scenario of early archaeological cultures in Upper Ganga Plain with special reference to Sakatpur (Saharanpur) and Harinagar (Bijnor). In V. Jaiswal (Ed.), Ganga in legend and history—Archaeology, literature and visual arts (pp. 1–11). Aryan Books International.

History of Glass Ornaments in Tamil Nadu, South India: Cultural Perspectives Veerasamy Selvakumar

Abstract Naturally available materials including plant leaves and seeds, stones, animal horns and bones were modified and utilized from the early times as ornaments. Beads of various semi-precious stones were used in Tamil Nadu from the Iron Age, and they might have been introduced in the earlier Neolithic period. The Iron Ageearly historic megalithic burials have beads and objects made of semi-precious stones, such as carnelian, quartz and other materials, placed as offerings for the dead. Glass beads and objects were perhaps widely introduced in the early historic period in South India with the development of long-distance trade, and strangely, they have not been found in the megalithic burials so far, and the excavated burials have mainly produced stone beads. In the later medieval period, glass bangles were used extensively and Valayalkarars specializing in glass bangle trade became a socially important community. Glass bangles became symbols of women, marginalized communities and even certain social groups. This chapter seeks to present a survey on the use of glass beads and glass ornaments from a cultural perspective, based on archaeological and textual sources.

1 Introduction Glass is a fabricated material that was used primarily for manufacturing various types of ornaments and ornament components such as beads, pendants, studs and cabochons, for a larger part of history, in India. Archaeological sites in Tamil Nadu have yielded evidence of glass beads from the early historic context, which can be tentatively placed between third century BCE and third century CE. There has been a continuous presence of glass ornaments from the early historic to the contemporary context in Tamil Nadu. Their distribution in the archaeological context suggests widespread use, but at varying intensity. This chapter presents an overview on the use of glass beads and bangles and their cultural contexts in historical Tamil Nadu. V. Selvakumar (B) Department of Maritime History and Marine Archaeology, Tamil University, Thanjavur, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_11

273

274

V. Selvakumar

Almost all the excavated sites in Tamil Nadu have produced glass beads in various cultural contexts, and some of these sites have yielded glass bangle fragments. However, their precise temporal use is unclear, due to the lack of complete data on their distribution in the published excavation reports. This chapter focuses on the excavated sites of Arikamedu and Kudikkadu for which well-documented data are available (Fig. 1), and it does not stress on the description of the material evidence from other excavated sites. Almost all the excavated sites in Tamil Nadu have produced glass beads and ornaments, and a study of these materials is a monumental research work. The study of beads, and glass beads, otherwise called ‘Indo-Pacific’ beads, began in India mainly in the second part of twentieth century (Abraham, 2016; Basa, 1992, 1993; Dikshit, 1969; Francis, 1990; Kanungo, 2004, 2016; Singh, 1989; Stern, 1987). The beginning of glass technology has been one of the research questions in focus among the researchers. Bhagwanpura is considered to have yielded glass in late second millennium BCE (Kanungo, 2008: 1023). Recent studies and excavation at Kopia by Alok Kanungo have produced evidence for glass bead making in first century BCE-CE context and presence of glass beads of identical composition in fifth century BCE (Kanungo, 2013; Kanungo & Brill, 2009). Apart from the origin and the earliest occurrence of glass and manufacturing technology, questions related

Fig. 1 Select early historic sites in Tamil Nadu with evidence of glass beads

History of Glass Ornaments in Tamil Nadu, South India …

275

to the mode of organization of bead-making craft, distribution of the produced goods, nature of the economy, the communities involved in the process (Francis, 1994) and the associated belief systems are also equally important in understanding the social dynamics of glass materials. One group of investigations has concentrated on the site of Arikamedu, mainly undertaken by Peter Francis Jr., whose studies are important in terms of understanding the bead production, techniques, furnaces and trade (Francis, 1987, 2002). Many of the studies on beads, before 1980s, focused mainly on typology rather than understanding the chaîne opératoire (Carter et al., 2016). Deviating from the old approaches, Peter Francis Jr. influenced bead studies using ethnographic approach. His ethnographic studies at Papanaidupet in Andhra Pradesh have offered clues to the reconstruction of the possible methods of glass bead and ornament production of the past societies (Francis, 1991). Scientific analyses of beads by several researchers have supplemented the understanding the glass technology (Lankton, 2011; Lankton & Dussubieux, 2006).

2 Glass Beads from Arikamedu and Other Sites in Tamil Nadu The process of urbanization that emerged in Tamil Nadu from 500 BCE led to the rise of major centres with industrial activities. Numerous urban and rural settlements existed (Fig. 1). The settlement of Keeladi (Keezhadi) has given a date of sixth century BCE for early settlement evidence (Sivanantham & Cheran, 2019). Arikamedu is an important excavated site with evidence for urban features and bead making. This site has Iron Age materials at the lowermost layers, and they have not been scientifically dated yet. This evidence might extend before 300 BCE. The site was continuously occupied till fifth century CE. Perhaps there was a gap in occupation till the tenth century, when the site witnessed occupation during the Chola times, perhaps till fourteenth century. Again, the site was in occupation after fifteenth century CE. Glass beads and bead wasters from Arikamedu have been extensively studied by researchers (Francis, 2004). R.V.G. Hancock has undertaken neutron activation analysis of the beads from Arikamedu (Francis, 2004: 521). Using ethnographic analogy, Peter Francis has demonstrated the possible method of manufacture of bead making (drawn beads) at Arikamedu. Peter Francis (1991, 2004: 451) finds correlation among the bead-making wasters from Papanaidupet, Arikamedu and several other sites. He suggested that Arikamedu glass beads were made locally and they were different from the Western glass materials (Francis, 2004: 472). Import of segmented glass beads, gold-glass and segmented melon beads, folded glass beads at Arikamedu has been suggested by Peter Francis (2004). According to him, the dark blue potash-manganese-cobalt colourant is unique to Arikamedu.

276

V. Selvakumar

Robert Brill, Laure Dussubieux, Bernard Gratuze, Alok Kanungo, James Lankton, Wood and others have scientifically studied the glass beads from India and other regions of Asia and Indian Ocean Rim using ‘LA-ICP-MS (Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry)’ technique (Lankton & Dussubieux, 2006; Wood et al., 2017; Siu et al. 2021). Their scientific investigations have added new dimensions to the bead studies. Lankton and Dussubieux (2006: 136), based on scientific analysis, have argued that most of the glass from Arikamedu belongs to ‘potash glass in dark blue, black, green and red’. They add that ‘composition of the Arikamedu potash glass, generally moderate in CaO and Al2 O3 , is similar to that of most of the potash glass samples found from South Asia to Korea, and it is possible that glassworkers at Arikamedu, or a related site nearby, produced potash glass’. They identify another variety which is of ‘the green, red and black soda glass’ variety (Lankton & Dussubieux, 2006: 138) at Arikamedu. James Lankton (2011) has suggested that glass production took place in India certainly by fifth century BCE in North India and later by third century BCE in South India, with Arikamedu represented as a major manufacturing site. The following summary is based on Lankton (2011), who observes that 13 compositional types of glass materials have been identified in India, as revealed by the analyses of Laure Dussubieux, Bernard Gratuze, James Lankton and Robert Brill. He adds that ‘mineral soda glass (m-Na-Al) type with high alumina, high barium and low uranium’ is found in South India and Sri Lanka (m-Na-Al 1), and similar material with ‘low barium and high uranium’ occurs in North India (m-Na-Al 3). ‘Potash glass with moderate lime and alumina (m–K-Ca-Al)’ forms an important component of Arikamedu beads from the Pondicherry museum collection. ‘Mineral soda-lime glass moderate in lime and alumina (m-Na-Ca-Al)’ is distributed in Tamil Nadu. Beads made of ‘mineral alkali glass obtained by mixing potash (m-K-Ca-Al) glass with soda glass (mNa-Ca-Al)’ are also reported at Arikamedu. ‘Soda glass with high alumina, lime and magnesia (v-mix-Ca-Al)’ has also been identified at this site. ‘Natron glass: mineral soda-lime glass with low alumina, potassium and uranium’ also occurs at Arikamedu. Arikamedu glass beads show similarities with the materials from Phu Khao Thong in Thailand suggesting strong linkages between these sites (Dussubieux et al., 2012). These studies suggest that Arikamedu was a major glass and glass bead producing centre, and possibly the technology, raw materials and beads were disseminated/traded/exchanged. Interestingly, a shipwreck from Godavaya in Sri Lanka has produced glass ingots (Muthucumarana et al., 2014) suggesting that glass as raw material was produced at a few centres and shipped and transported from the manufacturing centres to the consumption centres. This discovery reiterates the fact that all sites did not produce the raw glass necessary for bead and ornament making and the ingots were also traded. It explains the absence of glass bead manufacturing evidence at many sites in Tamil Nadu, while glass beads are widely distributed. From the archaeological excavations in Tamil Nadu, it is certain that the port towns and the urban centres of the early historic period located at the key nodes were producing raw materials in addition to the finished goods for trade and exchange. Thus, there existed a few primary production centres of glass and several settlements where beads were made and used.

History of Glass Ornaments in Tamil Nadu, South India …

277

2.1 Glass Evidence from Kudikkadu/Karaikkadu Kudikkadu, located 25 km south of Arikamedu, is a site contemporaneous with Arikamedu and may be considered as a satellite settlement forming a part of Arikamedu-based trade hub and network (Figs. 2, 3, 4). The site of Kudikkadu/Karaikkadu was excavated by the University of Madras in 1989 (IAR 1987–88: 103; Raman, 1991) and the author participated in the excavations as a post-graduate student. The site has about three-metre thick cultural deposit, dating to the early historic and medieval periods, possibly with a gap of occupation between these periods. The site might have been occupied by about third century BCE and was in occupation till the end of early historic period. Possibly, the site witnessed occupation during the Chola period and Late Medieval period. No radiocarbon dates are available for this site, which has produced one Roman amphora handle fragment and a large volume of roulette ware ceramics. Remarkably, a vast quantity of bead wasters, drawn tubes and finished glass beads were encountered at this site. However, the materials from this site have not been systematically studied. While glass beads are found in the early historic layers, glass bangle fragments were mainly confined to the upper layers of the late medieval period. Two trenches were excavated at the site, viz., KDU I and KDU II. Three furnaces were found in early historic context. Furnace 1 was at a depth of 1.79 m, built of mud, and it had evidence of glass slag. Furnace 2 at a depth of 1.5 m was made of brick-fragments, pottery and clay. It was 60 cm in length and 40 cm in width and was rectangular shape. Furnace 3 excavated in Trench KDU II was at a depth of 1.49 m. This furnace had an opening on the top with a diameter of 30 cm, and this context produced glass beads and a pot with traces of smoke. It was one metre long

Fig. 2 Stone, glass beads and other objects from Kudikkadu

278

V. Selvakumar

Fig. 3 Trench KDU II excavated at Kudikkadu

Fig. 4 Distribution of glass and stone beads from KDU II the numbers in X axis indicate layer numbers

History of Glass Ornaments in Tamil Nadu, South India …

279

and 0.16 m in thickness. And the fourth furnace excavated in Trench KDU II was found at a depth of 0.58 m, it was 0.70 m in diameter and 16 cm in depth, and its function was not certain. While the three furnaces excavated at the site belonged to early historic period, the one excavated in the upper layers of KDU II belonged to the medieval period. In Trench KDU I, furnaces were found at a depth of 149 to 179 cm and the frequency of the beads in the sediments increased from the depth of 170 cm, from Layer 5. The bangle fragments were found from Layer 3 and above the depth of 90 cm, but not below. Based on the glass wasters and beads, it is obvious that glass bead making existed at this site in the early historic period. Presence of glass bangle fragments in the upper layers at Kudikkadu suggests that the site was continued to be occupied and bead making might have taken place here in the medieval period too. The Chola period occupation is evidenced at Arikamedu, from tenth century CE, and similarly Kudikkadu was occupied around that time. The later phase at Kudikkadu could be contemporary to the site of Manikkollai (the name in Tamil means place of beads), which has possibly evidence for glass working such as glass wasters and chunks molten glass on ceramics. Manikkollai is located about 20 km further south of Kudikkadu. The glassmaking evidence at Manikkollai could probably date to seventeenth–eighteenth centuries CE based on associated ceramics. The site is considered to belong to early historic by a few researchers, but this statement needs further validation through excavation and ceramic studies. What emerges from the excavation at Kudikkadu is that this site has evidence of bead making, both glass and stone, in the early historic and medieval contexts. Glass bead making did not take place at Arikamedu in the medieval period, as the site was perhaps abandoned or had limited occupation after fifth century CE. Further scientific studies are needed on the bead and bangle collections from Arikamedu, Kudikkadu and Manikkollai to understand the glass-bead-making processes and the associated networks of glass trade.

2.2 Comments While glass beads and bangles are reported from several spatio-temporal contexts in Tamil Nadu, detailed studies have not been undertaken on these materials, except those from Arikamedu. Many of the sites have produced evidence of beads, but evidence of glass manufacture is uncertain, including the site of Pattanam (Abraham, 2013). From the observation of the available materials, it appears that evidence of glass manufacture is found mainly at Arikamedu and Kudikkadu in early historic context, and perhaps at Kudikkadu and Manikkollai in the medieval context. Probably Arikamedu and Kudikkadu could have served as a source of raw glass, in addition to supplying the finished glass beads in the early historic context. The finding of beads from surface context might not be useful in understanding the temporal development of glass beads, and at several sites, the beads found from the surface are from the

280

V. Selvakumar

later context, and hence, they may not also be useful for scientific studies seeking to understand the temporal pattern.

3 Glass Bangles in the Medieval Period In the archaeological excavations, glass bangles are found mainly in the contexts of the medieval period in South India, and they have not been reported from the contexts of early historic layers in clear terms, except as a few stray finds. Shell bangles, made of Turbinella Pyrum, were popular in the early historic period, and they are frequently mentioned in the Sangam Tamil texts. At several sites, for example at Arikamedu and Pattanam, no known finds of glass bangles have been noticed in the early historic context. Gangaikondacholapuram, the capital of the Cholas dateable to eleventh century CE, has evidence of glass bangle fragments of black, blue and yellow types (Vasanthi, 2009: 100). The archaeological site of Ambal (Nagapattinam dist., Tamil Nadu), excavated by the author, has produced glass bangles mainly from the late medieval context (Selvakumar et al., 2016). Glass bangles are considered to be rare in India till the fourteenth century CE by a few researchers (Sankalia, 1947; Francis, 2004: 519). However, the Bhagwanpura evidence points to the possible introduction of glass bangle as early as fourteenth century BCE (Kenoyer, 2005: 37) and glass bangles were perhaps used in northern part of India in the early historic times. This evidence needs to be researched further. Glass bangles definitely occur from the later half of 1st millennium CE in South India with a higher frequency. Bangles are reported to have been produced at Mantai from 700 to 1000 CE according to Peter Francis. At Alagankulam and Anuradhapura, Gunasena (2018: 337) states that glass bangles occur between 100 and 500 CE. Mantai also have bangles throughout the occupation, while the black bangles appear from third to seventh centuries CE. Bohingamuwa (2017) too argues for the early occurrence of glass bangles in Sri Lanka. Arikamedu has one specimen of bangle from Trench XI locus 046. It is not certain if it was a bangle or just a by-product of bead manufacturing activity. In the light of this data, further scrutiny of the excavated materials in Tamil Nadu is needed to understand the usage of glass bangles.

4 Glass in Literature References to glass are found in ancient Indian texts and Tamil literature also has references to glass or glass-like materials. The terms palingu and palikku found in Tamil literature could be interpreted as crystal quartz or glass. Among the texts of early historic period, Kurinjipattu 57–59, Malaipadukadam 515–518 and Akananuru 315, 10–12 have references to glass-like or crystal material. Kurinjipattu (57–58) refers to the flowing water of a spring as clear as glass. In the Tamil Ramayana of Kambar, dated to eleventh century CE, a vessel of glass is mentioned (Mithila

History of Glass Ornaments in Tamil Nadu, South India …

281

Section, poem, 14). Manimekalai, which is dated to early medieval period, has a reference to palikkarai, which could mean a room of quartz or glass (Manimekalai 40). Palingu is reported in Manimekalai, and palikku is another term that referred to glass. Manimekalai refers to aati which could be crystal in the context (Manimekalai 8.47). Thirukkural is a didactic work, and it refers to palingu as a transparent material (Thirukkural 706) meaning crystal or more particularly glass. These descriptions prove beyond doubt that people had the knowledge of glass.

5 Glass in Inscriptions Medieval inscriptions do have certain terms which could refer to glass, but they are very few. An inscription datable to CE 1127 found at Trisulam near Chennai mentions about a small tank used by glass bangle traders (Palukku Vaniyan Kazhuval kuzhi) while mentioning the land boundaries (SII VII, 540). An inscription found in Belgaum in Karnataka, dateable to CE 1264, refers to Kallakundarage and Nitturu as balagara sthala, which means centres of bangle manufacturers (Kambar, 2002: 187). Thanjavur Brihadiswara temple inscriptions, dateable to eleventh century (1014 CE), mention about palikku vayiram as a material associated with several ornaments (Selvakumar, 2017). It is not clear if palikku vayiram refers to a type of diamond or merely fabricated glass. According to the Tamil Lexicon, palikku vayiram is interpreted as an unpolished diamond variety. In the Glossary of Tamil Inscriptions (GTI, 2003), palukku is reported as glass. The ornament of Tiruvadinilai (sandal for gods), mentioned in the inscription of Thanjavur Brihadiswara temple inscription, had 10 quartz pieces and 38 palikku vayiram pieces, and 314 palikku vayiram pieces were found in a sandal. An unknown ornament had 77 palikku vayiram pieces weighing one mancati (a weight measure of 220 mg) and one kunri (a weight measure of 110 mg), and each piece weighed about 4.285 mg. Prishtakandigai (SII, II, 93. p. 431) had four palikku vayiram pieces; Sri Muti (crown) had palikku vayiram pieces, each weighing of 10.84 mg, while the same material in Tirumalai weighed 3.54 mg (Selvakumar, 2017) (Figs. 5, 6). Thanjavur temple inscription has reference to pothi, which might be also a type of glass (Selvakumar, 2017), and poth in Urdu means a small glass bead. Some of these beads might have come to South India through long-distance trade activities.

6 Craft Specialization and Bangle Making/Selling Castes During the late mediaeval period, there seem to have been a proliferation of crafts in South India because of transformation in the economic system, increase in domestic consumption and long-distance trade. The importance of crafts and the associated communities is attested by their upward social mobility; for example, the Kaikkolar,

282

V. Selvakumar

Fig. 5 Weight (in mg.) of palikku vayiram compared to other materials of Brihadiswara temple, Thanjavur

Fig. 6 Weight (in mg.) of palikku vayiram in different ornament types of Brihadiswara temple, Thanjavur (Uzhuttu and Sri Candam had identical stone weight in different ornaments, suggesting conscious selection of plaikku vayiram of identical size and weight categories)

who were textile weavers, became an important social segment in the later medieval times (Ramaswamy, 1985), and they actively participated in temple administration in South India. Apart from the production of glass and glass artefacts, organization of production, the market demand and sellers’ role in distributing the produced goods were also

History of Glass Ornaments in Tamil Nadu, South India …

283

important in an economic system. Specialization in glass bangle making and glass bangle selling was one of the features of material cultural production and distribution system of the medieval period. While several people took up the profession of selling glass bangles, a few specific groups specialized in this profession and gained social status. Among the Nayakkar or Nayudu caste of South India, one division is known as Valayalkara Nayakkar (Sankar, 2008). They were producing and selling glass bangles across South India. They are also known as Valayalkara Chettiyar. Valayalkar Kavarai people are recorded in the Census of India (CI 1881). In Telugu, gazula means bangles, and bangle dealers were known as Gazula/Gazulu Balijas in Andhra Pradesh and in the northern part of Madras presidency. Gazulu Lakshminarasu Chetty was the son of an indigo and textile merchant Gazulu Sidloo Chetty (Neild-Basu, 1984). He was the founder of the Madras Native Association. It appears that the influential bangle sellers shifted to trading in other commodities in the colonial times. In Karnataka, Banajigas were the bangle sellers. In Maharashtra, Kasars were traditionally makers of bronze objects and traded in the same material. They also made glass bangles at a later time (Mitragotri, 1992; Vaidya, 2018).

6.1 Origin of the Balijas On the origin of the Balija community, there are a few prevailing myths which have been documented by Edgar Thurston and Rangachari (1909). According to one version of the story, lord Siva asked his companion Parvati to emerge in front of him with glorious features. Then, she appeared in front of Siva, but Siva said that she was not very charming. Siva said that he could make her more appealing; then, she requested Siva to increase her charms. Later Siva, from his braid, created a being who came up with several bangles and turmeric paste. Parvati ornamented herself with these adornments. The person who was created by Siva is considered as the ancestor of the Balijas. In another version of the story, documented by Thurston, Parvati not happy with her appearance, prayed to Brahma, who asked her to perform penance and pray. She then prayed, and a person came from the fire leading a donkey with bangles, turmeric, ear ornaments of palm leaf rolls (pathra kundala), black beads, sandal powder, comb and perfume. This great being (maha purusa) is considered as the ancestor of the people of Balija community. Gazula is considered as a sub-division of Balija, Kapu and the Toreya community of South India. The Toreyas have a traditional belief that they originated from the bangles of Machyagandhi, the daughter of a fisherman on the Yamuna river, who married king Shantanu of Hastinapur (Thurston & Rangachari, 1909: 279–80). Bangle-makers and sellers migrated to various parts of Southern India and Deccan during the later medieval period, mainly after the thirteenth century. It appears that glass bangles became an important commodity associated with women and familyoriented and religious rituals from the thirteenth century CE. Selling of glass bangle

284

V. Selvakumar

was undertaken by itinerary merchants. Migration and settlement of this community led to the creation of bangle sellers’ street at many settlements of South India till the twentieth century. Almost all major and minor urban centres and major village settlements had the streets of Valayalkarars (Bangle sellers) in the twentieth century. These streets still bear their old names. The colonial records have references to the ubiquitous bangle sellers across the country in the eighteenth and nineteenth centuries CE. In the colonial times, the bangle merchants walked into the remote areas of the villages selling bangles (Fig. 7), and in the later context, they were moving with

Fig. 7 Colonial representation of an itinerant bangle seller, Victoria Albert Museum, London (after Archer & Parlett, 1992)

History of Glass Ornaments in Tamil Nadu, South India …

285

bicycles for reaching out to the rural market. Thus, they became part of the village landscapes of South India and are reflected in the literary narratives of modern period. The bangle sellers were considered as a good omen to be sighted in the settlements, since bangles had acquired an auspicious status. Sarojini Naidu (Naravane, 1996) has depicted the conditions of the bangle sellers in her poem. ‘The Bangle Sellers Bangle sellers are we who bear. Our shining loads to the temple fair… Who will buy these delicate, bright. Rainbow-tinted circles of light? Lustrous tokens of radiant lives, For happy daughters and happy wives. Some are meet for a maiden’s wrist, Silver and blue as the mountain mist, Some are flushed like the buds that dream. On the tranquil brow of a woodland stream, Some are aglow with the bloom that cleaves. To the limpid glory of new born leaves. Some are like fields of sunlit corn, Meet for a bride on her bridal morn, Some, like the flame of her marriage fire, Or, rich with the hue of her heart’s desire, Tinkling, luminous, tender, and clear, Like her bridal laughter and bridal tear. Some are purple and gold flecked grey. For she who has journeyed through life midway, Whose hands have cherished, whose love has blest, And cradled fair sons on her faithful breast, And serves her household in fruitful pride, And worships the gods at her husband’s side.’ Sarojini Naidu, 1912 (in Tharu & Lalita, 1991: 331).

7 Bead Production in Nagapattinam Nagapattinam was a medieval and modern period settlement, and evidence of beads and glass bangles has been reported from this site from an archaeological excavation (Jayakumar, 2009). The records of the Portuguese mention that the Portuguese purchased beads from Nagapattinam for African trade. Beads of green, yellow and blue were in demand among the natives of the African coast (van der Sleen, 1956; Wood et al., 2009.) Perhaps, the Valayalkara Chettiyars were involved in the bead

286

V. Selvakumar

production around Nagapattinam in the Lower Kaveri delta and the Coromandel coast.

8 Mounds of Bangle-Makers The Valayalkara Chettiyars are considered to have migrated from Andhra Pradesh into Tamil Nadu. The author has explored a few sites in Madurantakam Taluk near Chingleput, and they were known as Valayalkasan medu, which means ‘the mounds of bangle smelting’ in Tamil. The local people reported that the bangle-making chettis or chettiyars (merchants) were on the move, and they made bangles nearby the tanks in the remote areas. These tanks had alkali-rich salty soil with whitish appearance, and they also offered the water necessary for their camp activities. The author noticed about six sites from Madurantakam in the south to Uttiramerur in the north in Madurantakam Taluk which seem to be dated to the later medieval period. These sites had bangle wasters, beads and pottery fragments.

9 Thiruvilayadal Puranam and Shiva as Bangle Seller Thiruvilayadal Puranam is a Tamil literary work of sixteenth century CE composed by Paranjyothi Munivar. This purana text has stories about the playful divine acts of Siva who took human incarnation in the town of Madurai. According to one of the stories (32. Bangle Selling Canto), the women of Daruhavana were arrogant and Lord Shiva took the form of Bhikshadana and enticed these women to teach them a lesson. Because of the sin caused by the temptation, these women were born as daughters of merchants in their next birth. Then, Siva took the form of a bangle merchant and offered bangles to these women, and they were freed of their sins. The incorporation of this myth of Siva as a bangle seller in the purana suggests the importance of bangle merchants in the late medieval society, i.e., during the time of the Nayakas. Representation of Shiva as bangle seller is found as stucco images on the gopura of the Meenakshi temple at Madurai dateable to the Nayakka period, possibly sixteenth and seventeenth centuries (Fig. 8). In Madurai temple, this puranic story is enacted as a part of bangle festival and is being celebrated in the month of August–September (Avani festival).

10 Sacred Centres and Bangle Merchants The Sthalapuranas (stories on the origin of temples) of a few of the Amman temples in Tamil Nadu are invariably associated with bangle sellers. A few of these temples belong to the community of Kavarai Chettiyars who were bangle sellers. In the

History of Glass Ornaments in Tamil Nadu, South India …

287

Fig. 8 Siva as seller of glass bangles stucco figurine in Madurai temple (Source dinamalar.com)

sthalapurana stories, the bangle merchants are narrated as resting/sleeping under a tree, during the noon time. Selling bangles in the remote areas in the hot climate is a tough activity, and thus, they are narrated as resting under a tree after selling bangles, and then, divine intervention happens. At Periyapalayam near Chennai, a bangle merchant was selling bangles, and after lunch, he was resting on the bank of the Arani river and his bangles disappeared. Later, he saw the goddess in his dream and the goddess declared her existence in the village. Similar literary motifs are noticed in the stories on the origin of several temples of South India such as at Samayapuram and Uraiyur near Tiruchirappally and Natchiyarkovil in Tamil Nadu. On certain festive occasions, the goddesses are decorated with garlands of bangles (Fig. 9). This offering happens mainly in the month of Aadi in the pooram asterism (July–August). It is common to observe abundant offering of bangles at many of these temples. People offer glass bangles on the branches on temples expecting fulfilment or after the fulfilment of their wishes. Bangles are offered mainly to the shrines of local deities, and these offerings are secured to the trees nearby the shrines seeking welfare of the family and child birth. Many of the temples are also sites of festival markets where textile, pottery, iron wares and bangles are marketed. Such temples offered a large market during the festivals and the merchants used to sell their commodities. In some of the larger temples such as Sri Rangam, Madurai and Kumbakonam, market areas came up very close to the temple precincts.

288

V. Selvakumar

Fig. 9 Amman with decorated bangles. Source Temple Folks, Chennai

11 Narikkuravars as Bead Sellers Apart from the Balijas, who were influential bangle traders with a higher status in the social hierarchy, several nomadic communities were involved in itinerant bead and bangle selling too. The Narikkuravars, who are categorised as a Scheduled Tribe and

History of Glass Ornaments in Tamil Nadu, South India …

289

Fig. 10 Narikkuravar women meeting collectors at Ramanathapuram. Source dailythanthi.com

considered to belong to the poorer, marginalized strata of the society, are engaged in the selling of glass bangles and beads (Jackson, 1989; Vijay, 2014; Zafiu, 2017). These nomadic people, who are known as gypsies and speak Vagri Boli language, might have been engaged in this trade from the colonial time, and they were notified as criminal tribes by the colonial administration. Even today, the Narikkuravar women are involved in selling bangles, beads, necklace, safety pins, ribbons and other commodities to the common people. In 2018, the Narikkuravars appealed to the district collector seeking permission to sell glass beads and other ornaments in Rameswaram (Fig. 10). As an itinerant community, it was easier for them to move across the country and sell beads among the people in the remote villages. The network of bead-makers and traders used such communities to distribute their goods, and it was a mutually beneficial relationship. The transformation of Narikkuravars is one of the examples of nomadic, hunters groups adapting to the prevailing market system to earn their livelihood.

12 Baby Shower (‘Bangle Protection’) Ceremony Bangle is viewed as the symbol of women. In South India, pregnant women in their fifth or seventh or ninth months of pregnancy are wearing plenty of bangles as part of baby shower ritual called ‘bangle protection’ (or Valayal Kappu) for the first pregnancy (Fig. 11). The pregnant women are adorned with hand full of glass bangles.

290

Fig. 11 A Lady during baby shower ceremony Courtesy Mr. R. Karthikeyan

V. Selvakumar

History of Glass Ornaments in Tamil Nadu, South India …

291

It is an established practice and people believe that the sound of the glass bangles is good for the foetus, and it would protect them. Perhaps wearing such a large volume of bangles would prevent them from heavy household work in the later stages of their pregnancy. Irrespective of caste and economic status, all communities of Tamil region follow this ritual practice. Cultural and economic factors are intimately linked in the development of belief systems. There are no known documents on the practice, except the oral traditions; there may be records available, but they need to be located.

13 Bangle Offering to River Deity River is personified as women, and worship of river is undertaken in Tamil Nadu in the month of aadi (July–August). Small black bangles are offered in the river in the Tamil month of aadi along with leaf ornaments called ‘kaathu olai’. Kaathu olai is the leaf ornament called pathra kuntala associated with the male aspect of Siva. In kaathu olai karugamani, a rolled palm leaf is inserted into two black bangles of small size, normally about 3 cm in diameter, which is left in the river along with the flowers and other offerings.

14 Gender and Ornament Use Bangles were perhaps seen as a symbol of prosperous women and were one of the offerings in the temples of goddesses. The practice of offering glass bangles at the sacred sites was known during the Vijayanagar times (Sinopoli & Morrison, 1995). Glass has been a cheaper material than metal, and it could be produced in multiple colours, and thus, it became an attractive alternative to the shell bangles and metal ornaments. Wearing of bangle is considered as a ‘Saubhagya lakshana’ in Indian tradition (Mitragotri, 1992: 108), and this belief persists even today. Identifying ornaments with a particular state of womanhood is recorded in the Sangam Tamil texts. Some women were labelled as kazhi kala makalir (women without ornaments) (Purananuru 280) and todi kazhi makalir (women without bangle) (Purananuru 238), which mean widows who abandoned their ornaments. Bangle is identified with women, and during the national movement, this idea was used as well as challenged (Thakar 2002; Basner, 2010; Gupta, 2017). Women broke the bangles of Czechoslovakian origin, although breaking of bangle was considered as an unacceptable practice for women. For instance, in Karnataka, the Bedas used the slogan ‘Our hands are of fighting men not of the bangle wristed women’ (Patil, 1992: 74). Even today, some women in rural India ritually abandon their ornaments or break their glass bangles, as a symbolic expression of widowhood.

292

V. Selvakumar

15 Glass Artefacts as Social Markers: Ornaments of Common People and Subalterns Glass bangles and beads were common among certain castes and communities, and a few of the elite communities never used to wear anything other than gold, at least till the twentieth century. While some types of glass beads are associated with the poorer sections, some of the glass beads along with gold are used by people of higher social status. Certain glass beads are associated only with certain communities and act as a marker of social identity. Interestingly, the black glass beads are also worn by Muslim women. Mostly, Telugu-speaking people wear ornaments of black glass beads as thali (a symbol of marriage). Similarly, for infants, a string of microbeads is tied around the hip, and here, the smaller beads are chosen in order not to hurt the tender infants. Glass bangles and beads were more often worn by women of various castes with low family income, and the primary reason for such a usage was its affordability and colourful nature. Glass as a low-cost material fulfilled the requirements of the people from poorer communities and classes. Unlike precious metals, glass is affordable to many, and glass beads seem to have been more popular among the people from the late medieval times. Hence, glass material could be considered as a signifier of the subalterns in some contexts. There are exceptions to this rule, and during the occasion of rituals and festivals, women of every community wear glass bangles. The colonial records depict the nature of glass ornaments worn by various communities (Figs. 12, 13). Edgar Thurston while referring to ornaments of the Cheruman records about low cost of these ornaments and refers to their glass bead necklaces (Thurston & Rangachari, 1909: 63–64). The Gadaba tribes wear glass beads during the wedding (Thurston & Rangachari, 1909: 248). The practice of breaking of bangles by wife on the death of husband is not universal among the simple living communities. For example, among the Gadabas, the bangles of widows are broken on the death of the husband (Thurston & Rangachari, 1909: 255), whereas the Golla women do not break or remove bangles on the death of husband (Thurston & Rangachari, 1909: 287). Among the Gauda caste of Karnataka, there is a clan name of Malara which means a bundle of glass (Thurston & Rangachari, 1909: 270). The Irulas wear a mass of glass necklaces and glass bangle and glass beads (Thurston, 1897: 381). The use of a large volume of glass beads and bangles is noticed among some of the communities.

16 Competition in Glass Bangle Trade The colonial system as well as the development of industrial capitalism opened up the Indian market to the global players affecting the interests of the craft groups in nineteenth and twentieth centuries. The arrival of German glass bangles into Indian

History of Glass Ornaments in Tamil Nadu, South India …

293

Fig. 12 An Irula woman with garlands of beads (after Thurston, 1897, Pl 7)

market perhaps caused troubles to the local bangle-makers and sellers in the early twentieth century. Edgar Thurston records a rumour about an insect from the glass bangles of German origin affecting people (Thurston, 1912: 81). ‘In recent years, a scare has arisen in connection with an insect, which is said to take up its abode in imported German glass bangles, which compete with the indigenous industry of the G¯azulas. The insect is believed to lie low in the bangle till it is purchased, when it comes out and nips the wearer, after warning her to get her affairs in order before succumbing. A specimen of a broken bangle, from which the insect is stated to have burst forth and stung

294

V. Selvakumar

Fig. 13 Paliyans hunter-gatherers with glass beads (after Thurston & Rangachari, 1909: 469)

a girl in the wrist, was sent to me. But the insect was not forthcoming’. (Thurston & Rangachari, 1909: 95). Such a myth might have emerged due to the competition in the market and the worries associated with the future of the local craftsmen and traders. The glass bangle industry was further transformed in the late nineteenth and twentieth centuries. During the time of Great Depression in the twentieth century, the glass bangle industry of Madras Presidency faced stiff competition from the glass bangles from Czechoslovakia (Manikumar, 2003: 83). In 1940s, China glass was used for making glass bangles (FCW, 1941). From about the middle of twentieth century, measures were taken through the Board of Scientific and Industrial Research to modernize the glass industries.

History of Glass Ornaments in Tamil Nadu, South India …

295

17 Impact of Bangle Industry on the Society The glass bangle and bead industries transformed from a small-scale industry to a major cottage industry in the colonial period in India. It is not clear how the industries of the early historic periods functioned. There is a probability that the glass production was controlled by the guilds of merchant in the early historic context, since it was a fairly new technology. In the medieval times, specialised communities would have developed, and these activities were undertaken by these communities, since the inscriptions refer to the communities of glass traders. However, in the late medieval context because of the growing monetization and further development in the economy, specialised group, the Gazulu Balijas emerged, and under the influence of colonialism, this industry grew further. The ideological linkages and rituals might have developed in this period due to the aggressive marketing practices. The commodities were marketed by various communities, apart from the Gazulu Balijas. The wholesale dealers and itinerant traders and the communities of Narikkuravas took to glass bead and bangle trade. Thus, diverse communities benefitted out of the glass bead and bangle trade, and the network of these communities was essential for marketing the goods at the grass roots. The processes shift from agrarian to commodity production and marketing might have emerged in the late medieval period, where we find the movement of cotton textile weavers, silk weavers and bangle-makers moving to the urban, semi-urban and rural centres of South India.

18 Discussion and Conclusions The Iron Age, which is placed between 1000 and 500 BCE in South India, is characterized by the megalithic burials and the use of iron, copper–bronze and other material cultural remains such as black ware and black and red ware. Ornaments of the Iron Age (within the megalithic burials) were mainly made of carnelian and quartz, and at a few sites, chank (Turbinella Pyrum) shell ornaments have been found. The predominant presence of crystal and carnelian beads in the megalithic burials reveal that stones were the most preferred materials for making ornaments in the Iron Age. People perhaps began to use glass materials mainly from the early historic context, since no clear trace of glass beads has been noticed from the Megalithic burials of the Iron Age context so far. The recent radiocarbon dates have pushed back the beginning of early historic to fifth or sixth century BCE in Tamil Nadu; however, a detailed study of the distribution of glass and other material cultural remains is necessary to understand the precise chronology of introduction of glass bead making and other technologies in South India. Glass beads were in use continuously from the early historic period to the contemporary period in Tamil Nadu. It appears that glass bangles became popular after thirteenth century CE as of now, although they might have been introduced in the early historic times. Glass beads were perhaps used as a cost-effective medium of

296

V. Selvakumar

exchange by the merchants in the early historic times for some of the goods. Usage of these artificial glass beads was confined to only certain section of the people, and they were not considered as a status symbol unlike other precious materials such as carnelian or other stones. Ornaments carry symbolic meanings and they could convey various social, economic and cultural ideas and identities. Glass beads and bangles were the ornaments of the common people, and one could argue that they were a part of subaltern material culture at least in the early part of twentieth century. Their cost-effective nature offered the poor people to adorn themselves, and they perhaps enjoyed a lower exchange rate. Only in the post-globalization era glass and plastic ornaments are more acceptable to many of the elite communities, since their diverse colour range attracted almost all sections of the society. Glass bangles became important and popular because of the commercialisation after thirteenth century, active marketing strategies and creation of rituals and belief systems. The specialized caste groups involved in the glass trade and their social status reflect the profitability of this trade. The glass bangles began to occupy an important place in the rituals and festivals, and they were considered as auspicious materials (mangala or sowbhagya) to be offered to women and goddesses. The glass objects symbolised women in some contexts and subalterns in some contexts. They were sacred in some contexts and were also associated with certain elite communities. The community of bangle-makers could carve a market for their goods from the time of merchant capitalism to the industrial capitalism from 13th to twentieth centuries CE. They adapted to the plastic-induced bangle industry, and the liberalization related developments in the late twentieth century brought major change in their profession. The historical trajectory of the bangle-making craft groups requires a more detailed study. The merchant capitalism and industrial capitalism had contributed to the nature and growth of the ornamental glass industry in the colonial period. It appears that the monetization of economy under the colonial rule, the political economy and the social transformations influenced the development of glass industry and bangle sellers. The bangle-makers/sellers used bangle industry as a means of subsistence, and the Narikkuravars took up this occupation to earn their basic livelihood. The bangle industry offered differing benefits to those involved in the commerce, with the major traders earning a lot of wealth, while the marginalised groups barely meeting their daily needs. It appears that the development of shrines, creation of ritual and religious practices as part of the belief systems and the linkages with the glass bangles and bangle-making Chettiyars point to the impact of the market on craft production in a monetised economic system of the colonial times. The social and economic contexts of the glass bangle industry reveal how material cultural when embedded within ideology and belief systems leads to the creation of market and transformations of the fortunes and social status of the communities concerned. The story of glass bangle industry suggests that sustainability of crafts is influenced by the market factors which in turn were influenced by the political economy and the social factors. Moreover, it is obvious that consumerism powered the wealth generation in many contexts.

History of Glass Ornaments in Tamil Nadu, South India …

297

Acknowledgements I would like to thank Alok Kanungo and Laure Dussubieux for the support and S.B. Darsana for the suggestions on the manuscript.

References Abraham, S. A. (2013). In search of craft and society: The glass beads of early historic Tamil South India. In S. A. Abraham, P. Gullapalli, T. P. Raczek, & U. R. Rizwi (Eds.), Connections and complexity: New approaches to the archaeology of South Asia (pp. 239–262). Left Coast Press. Abraham, S. A. (2016). Glass beads and glass production in early South India: Contextualizing Indo-Pacific bead manufacture. Archaeological Research in Asia, 6, 4–15. https://doi.org/10. 1016/j.ara.2016.02.003 Archer, M., & Parlett, G. (1992). Company paintings: Indian paintings of the British period. Victoria and Albert Museum / Mapin Publishing Pvt. Ltd. Basa, K. K. (1992). Early glass beads in India. South Asian Studies, 8, 91–104. Basa, K. K. (1993). Manufacturing methods of monochrome glass beads in Southeast Asia. Man and Environment, 18(1), 93–100. Basner, K. S. (2010). Bangles duty and transgression in North India. Semantic Scholar. Corpus ID: 67768045. https://www.semanticscholar.org/paper/Bangles-%2C-Duty-and-Transgressionin-North-India-Basner Bohingamuwa, W. (2017). Sri Lanka and the Indian Ocean contacts: Internal networks and external connections. PhD Thesis. University of Oxford. Carter, A. K., Abraham, S. A., & Kelly, G. O. (2016). Updating Asia’s maritime bead trade: An introduction. Archaeological Research in Asia, 6, 1–3. CI = Various Census of India .(1881). Report of the Census of British India Taken on 17 Feb 1881 (Vol 3. pp. 22). London. Dikshit, M. G. (1969). History of Indian glass. University of Bombay. Dussubieux, L., Lankton, J., Bellina-Pryce, B., & Chaisuwan, B. (2012). Early glass trade in South and Southeast Asia: New insights from two coastal sites, Phu Khao Thong in Thailand and Arikamedu in South India. In Crossing borders in Southeast Asian archaeology. Selected papers from the 13th International Conference of the European Association of Southeast Asian Archaeologists, Berlin, 2010. (Vol 2, pp. 307–328). Singapore: NUS Press. FCW. (1941, August 30). Glass and products, Foreign Commerce Weekly, 4(9): 26. Francis, P. Jr. (1987). Bead emporium: A guide to the beads from Arikamedu in the Pondicherry Museum. Pondicherry: The Pondicherry Museum. Francis, P. Jr. (1990). Glass Beads in Asia, part II, Indo-Pacific beads. Asian Perspectives, 29(1), 1–23. Francis, P. Jr. (1991). Beadmaking at Arikamedu and beyond. World Archaeology, 23, 28–43. Francis, P. Jr. (1994). Towards a social history of beadmakers. Beads, 6, 61–80. Francis, P. Jr. (2002). Asia’s maritime bead trade 300 B.C. to the Present. Honolulu: University of Hawai’i Press. Francis, P. Jr. (2004). Beads and selected small finds from the 1989–92 excavations. In V. Begley, N. Karashima, K. V. Raman, S. E. Sidebotham, & E. L. Will (Eds.), The ancient port of Arikamedu: New excavations and researches 1989–1992 (pp. 447–604). École Française d’Extrême Orient. GTI 2003 = Glossary of Tamil Inscriptions. Chennai, Santhi Sathana. Gunasena, K. G. (2018). Interactions between Sri Lanka and South India in the early and middle historic through the perspective of personal adornment. PhD Thesis. University of Exeter. Gupta, R. (2017). Satyagraha and the feminizing of the Indian national movement. ANTYAJAA: Indian journal of Women and Social Change, 1(2), 133–142. https://doi.org/10.1177/245563271 7694240 IAR = Indian Archaeology—A review. New Delhi: Archaeological Survey of India.

298

V. Selvakumar

Jackson, W. (1989). Rituals of a ‘Gypsy’ tribe: The Vagri or Narikuravar”, presentation at Indianapolis Regional American Academy of Religion Conference, Bloomington, April 8. https:// www.academia.edu/4413853/Rituals_of_a_Gypsy_Tribe_The_Vagri_or_Narikuravar Jayakumar, P. (2009). Nagapattinam Akzhayvu. Avanam Journal of Tamil Nadu Archaeological Society [tamil], 19, 118–122. Kambar, J. (2002). History of Belgaum district up to 17th century. PhD Thesis. Kolhapur: Shivaji University. http://hdl.handle.net/10603/138413. Kanungo, A. K. (2004). In Glass beads in ancient India: An ethnoarchaeological approach. Oxford: BAR International Series, 1242, John and Erica Hedges. Kanungo, A. K. (2008). Glass in India. In H. Selin (Ed.), In Encyclopaedia of the history of science, technology, and medicine in non-western cultures (pp. 1023–1033). https://doi.org/10.1007/9781-4020-4425-0_9743. Kanungo, A. K. (2013). In Glass in ancient India: Excavations at Kopia. Thiruvananthapuram: Kerala Council for Historical Research. Kanungo, A. K. (2016). Mapping Indo-Pacific beads vis-à-vis Papanaidupet. Aryan Books International / International Commission on Glass. Kanungo, A. K., & Brill, R. H. (2009). Kopia. India’s First Glassmaking Site: Dating and Chemical Analysis, Journal of Glass Studies, 51, 11–25. Kenoyer, J. M. (2005). Culture change during the Late Harappan period at Harappa: New insights on Vedic Aryan issues. In E. F. Bryant & L. L. Patto (Eds.), The Indo-Aryan controversy: Evidence and inference in Indian history (pp. 33–61). Routledge. Lankton, J. (2011). Glass technology of the past preparing for the future, Paper Presented in the ICTS Workshop on the future of the past, Mangalore, 23rd November. Tata Institute for Fundamental Research. Lankton, J. W., & Dussubieux, L. (2006). Early glass in Asian maritime trade: A review and an interpretation of compositional analyses. Journal of Glass Studies, 48, 121–144. Manikumar, K. A. (2003). A colonial economy in the great depression, Madras (1929–1937). Hyderabad: Orient Longman Ltd. Mitragotri, V. R. (1992). A Socio-cultural history of Goa from the Bhojas to the Vijayanagara. PhD Thesis. Goa: University of Goa. Muthucumarana, R., Gaur, A. S., Chandraratne, W. M., Manders, M., Ramlingeswara Rao, B., Ravi Bhushan, Khedekar, V. D. & A.M.A. Dayananda (2014). An early historic assemblage offshore of Godawaya. Sri Lanka: Evidence for Early Regional Seafaring in South Asia, Journal of Marine Archaeology, 9(1), 41–58. Naravane, V. S. (1996). In Sarojini Nadu, her life, work and poetry. Orient Blackswan. Neild-Basu, S. (1984). The dubashes of Madras. Modern Asian Studies, 18(1), 1–31. https://www. jstor.org/stable/312381. Patil, V. S. (1992). Role of the Bijapur district Karnatak in the Indian freedom struggle. PhD Thesis. Karnataka University. http://hdl.handle.net/10603/94927. Raman, K. V. (1991). Further evidence of roman trade from coastal sites in Tamil Nadu. In V. Begley & De. Puma (Eds.), Rome and India: The ancient Sea trade (pp. 125–133). University of Wisconsin Press. Ramaswamy, V. (1985). Textiles and weavers in medieval South India. Oxford University Press. Sankalia, H. D. (1947). The antiquity of glass bangles in India. Bulletin of the Deccan College Post-Graduate and Research Institute, 8, 252–259. Sankar, R. (2008). Sangakala Makkalin Valviyalum Tharkala Nilaiyum. PhD Thesis. Bharathiar University. http://hdl.handle.net/10603/105898. Selvakumar, V. (2017). Rattanattin Tiruvabharanangal in the inscriptions of Brihatiswara Temple Tancavur. In A. K. Kanungo (Ed.), Stone beads of South and Southeast Asia: Archaeology, ethnography and global connections (pp. 49–114). IIT Gandhi Nagar / Aryan Books International. Selvakumar, V., Mathivanan, K., Parandaman, V., Jaseera Majeed, A., Raja, S., & Karthikeyan, R. (2016). Iron age-early historic and medieval settlements in the lower Kaveri Valley: A preliminary

History of Glass Ornaments in Tamil Nadu, South India …

299

report of the archaeological excavation at Ambal (2015–16). Nagapattinam District, Tamil Nadu, Puratattva, 46, 168–174. Siu, I., Henderson, J., Qin, D., Ding, Y., & Cui, J. (2021). A study of 11th–15th centuries AD glass beads from Mambrui, Kenya: An archaeological and chemical approach. Journal of Archaeological Science: Reports, 36, 1–10. https://doi.org/10.1016/j.jasrep.2020.102750 SII = South Indian Inscriptions. New Delhi, Archaeological Survey of India. Singh, R. (1989). Ancient Indian glass: Archaeology and technology. Parimal Publications. Sinopoli, C. M., & Morrison, K. D. (1995). Dimensions of imperial control the Vijayanagara capital. American Anthropologist (new Series), 97(1), 83–96. Sivanantham, R., & Cheran, M. (2019). Keeladi An urban Settlement of Sangam Age on the banks of the River Vaigai. Chennai: Tamil Nadu State Archaeology Department. Stern, E. M. (1987). Secret of Papanaidupet. Glastechnische Berichte, 60(10), 346–351. Thakkar, U. (2002). Breaking the shackles: Gandhi’s Views on Women. In M. P. Mathai, M. S. John, & S. K. Joseph (Eds.), Meditations on Gandhi: A Ravindra Varma Festschrift (pp. 179–188). Concept Publishing Company. Tharu, S. J., & Lalita, K. (Eds.). (1991). Women writing in India: 600 B.C. to the early twentieth century. The Feminist Press. Thurston, E. (1897). Badagas and Irulas of Nilgiris, Madras Government Museum. Bulletin, 2, 1–7. Thurston, E., & Rangachari, K. (1909). Castes and Tribes of Southern India. Madras: Government Press Thurston, E. (1912). Omens and superstitions of Southern India. T.F. Unwin. Vaidya, M. (2018). The consolidation of Jati in colonial India. In R. Seshan & S. Kumbhojkar (Eds.), Re-searching transitions in Indian history (pp. 167–179). Routledge. Vasanthi, S. (2009). Excavation at Gangaikonda Cholapuram, the imperial capital of Rajendra Chola, and its significance. In H. Kulke, K. Kesavapany & V. Sakhuja (Eds.), Nagapattinam to Suvarnadwip: Reflections on the chola naval expeditions to Southeast Asia (pp. 96–100). Singapore: Institute of south-east Asian Studies. van der Sleen, W. G. N. (1956). Trade-wind beads. Man, 56, 27–29. Vijay, S. (2014). A socio-economic perspective on the occupations of Narikkuravars in Tamil Nadu. Proceedings of the Indian History Congress, 75, 1208–1214. https://www.jstor.org/stable/441 58511. Wood, M., Dussubieux, L., & Wadley, L. (2009). A cache of ~5000 glass beads from the Sibudu cave iron age occupation. Southern African Humanities, 21, 239–261. http://www.sahumanities. org.za. Wood, M., Panighello, S., Orsega, E. F., Robertshaw, P., van Elteren, J. T., Crowther, A., Horton, M., & Boivin, N. (2017). Zanzibar and Indian Ocean trade in the first millennium CE: The glass bead evidence. Archaeological and Anthropological Science, 9(5), 879–901. https://doi.org/10. 1007/s12520-015-0310-z Zafiu, M. (2017). Health access for vulnerable groups: A study on the “Gypsy” Narikuravar community in Tamil Nadu, India, Joseph Wharton Scholars. http://repository.upenn.edu/joseph_wha rton_scholars/36.

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India and an Outline of the Use Jan Kock and Torben Sode

Abstract The production of lead-backed mirrors in India dates back to the early Mughal period (1526–1857), when the craft was introduced by Persian craftsmen. In the seventeenth century, the Mughals used them to decorate their palaces, a tradition that continued until the nineteenth century. In Kutch, Gujarat, the locals have decorated the walls and interiors of their mud houses with glass mirrors for centuries. The use of mirrors is still quite widespread among the people living in the rural areas of Gujarat and in southwestern Rajasthan that incorporate them in embroideries. The people of the Banjara community, who live mainly in the central part of India, are known for using relatively large glass mirrors, up to 10 cm in diameter, in their embroidered textiles. In Gujarat and Rajasthan, the mirror pieces are also used to decorate boxes, furniture, small figures and other objects.

This article describes the manufacturing of such mirrors in a still active workshop located in the town of Kapadvanj in Gujarat. There, glassworkers produce a glass sphere with very thin walls in which is poured liquid metal, a mix of lead with a small quantity of tin and zinc, to coat the inside of the sphere. The glass sphere is then broken into mirror fragments that are packed and then shipped to places such as the town of Limbdi where women will cut the mirrors at the required size before being sold to craftsmen producing embroidered textiles.

J. Kock (B) Aarhus University, Hojbjerg, Denmark e-mail: [email protected] T. Sode Glass Historian Torben Sode, Brønshøj, Denmark e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_12

301

302

J. Kock and T. Sode

1 Mirror Making at Kapadvanj The last working furnace which produces blown mirrors coated with lead is in the town of Kapadvanj in the state of Gujarat in Western India. This is the southern workshop in this paper. Until recently, there were two functional glass workshops situated close to each other in the village. The northern workshop, on which this article is based, was closed down at the end of the first decade of this century. For the last many years, these two glass workshops have been producing nothing but mirrors backed with lead, with a technique that has not changed for centuries. Before the northern workshop was shut, both workshops were working only at half capacity and the mirrors were sold at such low prices that the production was hardly worthwhile. The competition from mirrored flat glass and mirrored plastic is becoming untenable. Also, the fashion of using mirrors in dresses, and other textiles is declining. The southern glass workshop operates now when there are orders. Until the middle of the twentieth century, the glass workshops also produced bottles, bracelets and other kinds of glassware. Thanks to the proprietor of northern workshop, Mohammed Siddiq Shisgar (Fig. 1), we could carry out the detailed documentation of the workshop in 1997 and 1999. In 1999, he was 92-year-old, living in the nearby city of Ahmedabad and used to come down to Kapadvanj covering a distance of more than 60 km. He was determined to continue the production as long as he lived in spite of the bad financial circumstances. The name Shisgar means maker of glass. Mr. Shisgar was a Sunni Muslim, and by tradition, they have dominated the glass production in Kapadvanj. According to Mohammed Siddiq Shisgar, his family had produced glass mirrors in Kapadvanj for more than 200 years, but in the past, glass mirrors were also made in other workshops in northern India and in today’s Pakistan. Rudyard Kipling wrote in the end of nineteenth century, that glass mirrors used for ornamental plaster work and in embroideries were manufactured in Karnal, north of Delhi (Mukherji, 1888: 296–97). One of the main reasons why glass and mirrors were produced in Kapadvanj is that the river Tabuka that flows through the city has deposited large quantities of alkaline sediments, one of the major ingredients for glassmaking. Locally, this alkali is called ush, and the right to collect it from the river was sold on licence annually. Furthermore, fuel was easy to come by locally in the form of wood. To melt the raw glass, the necessary sand, the so-called black sand, was for a long time imported from the neighbouring state of Rajasthan. The sand must have contained the necessary needed calcium, most probably in the form of shells. Kapadvanj have for centuries been one of the important glass centres in western India with a large production of different kind of glass vessels, bottles, sprinklers, bangles and of course mirrors (Dikshit, 1969: 133–139). In the end of the nineteenth century, Mr. B.A. Gupte wrote on the manufacturing of glass in the Bombay Presidency: ‘Very little glass is produced in the Presidency, and that principally is manufactured at Kapadvanj in the Kaira District and is remarkable for iridescent properties and good colour resembling old Venetian. The shapes too of the little vessels and cups are very quaint and beautiful. At Kapadvanj the workers are Muslim, numbering about 70, who have

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

303

Fig. 1 92 years proprietor of the northern glass works, Mohammed Siddiq Shisgar photographed in 1999

been following this craft for several generations. Most of them are poor and have frequently to borrow money for the purpose of carrying on their work. The material for making glass are ush an alkaline earth obtained locally, impure carbonate of soda sajikhàr, and a variety of dark, flinty sand from Jaipur. The glass is made in large earthy furnaces, in form like huge slipper baths. The floor sloping towards the holes prepared to receive the melted glass. The materials are mixed together in certain proportions, placed in the furnace, and when raised to a white heat the molten mass

304

J. Kock and T. Sode

is run into a trench, where it remains till it is cool. It is then broken up into small pieces, re-melted and shaped into bangles and small vessels’ (in Maconochie, 1895).

2 The Glass Works and the Furnace The active glass workshop of Kapadvanj and the ruin of the second one are on a small hill at the outskirts of the city. The hill is elevated more than two metres above the surrounding area, and a qualified guess is that the hill is mainly artificial and consists of waste from the glass produced for centuries. Even today, workers dump waste on the slope to the east. On the premises of both glassworks, there are ruins of old tank furnaces to be seen (Fig. 2). The premises of the now disappeared glass workshop was surrounded by a fence with the workshop itself built to the north, while in the southwest corner a bricked up building was used as office and included a room for the supervisor. A quite large courtyard was used for storing firewood and coal for the furnace, as well as pieces of broken glass to be reused as raw glass. Attached to the glass workshop building itself were some rooms to store raw material, boxes for packing and ready production among others. In the courtyard, the dirty recycled glass was cleaned by shovelling it through a sieve, and then, it was either stored in the workshop building ready for use or brought

Fig. 2 An abandoned furnace at the premises of the southern glass works

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

305

immediately to the furnace. The internal transport of broken glasses was done in an iron pan carried on the shoulder. The glass workshop building was bricked up in a way that left many small openings which allowed sufficient air into the hall with the furnace. The roof was a light construction to protect mainly from the sun and the rain. On both sides of the main hall, there were small rooms which were partitioned by approximately 1.5-m-high walls. Some of these rooms were used for raw materials and reusable waster from the previous production. The cheapest raw material to buy today is burned out light bulbs and fluorescent tubes. They are mixed with waste from the previous production. This means broken pieces of mirrors often with thick layer of lead are incorporated in the glass batch. So, if one measures the composition of the finished glass mirror it will contain some lead, which is added to the glass by accident and was accepted because it did not raise problems for the ongoing production. In the past, a similar lead contamination was certainly happening when the workers mixed raw materials from the river bank and sand from Jaipur with reused wasters from the previous production. The furnace was a typical tank furnace and same has been in use for centuries (Fig. 3). The furnace was situated in the centre of the building and was mainly constructed with blocks of sandstone and is covered by a rather flat dome. The furnace was laid out with two chambers separated by a wall with a height of a little more than half a meter (Fig. 4). To the east was the stoke chamber with a stoke hole on the southern long side. To the east there was a deep hole used when removing ashes. The stoke chamber must have been modified with a grate and beneath that with a pit for slag and ash of coal. Grate is not required for a wood fired furnace, as evident from the other in-use native wood fired furnaces in India. This change in the firing (addition of charcoal) was introduced in Kapadvanj nearly hundred years ago. The flames had to pass over the separating wall and heat the top of the glass batch. To the west was the melting chamber with a floor nearly levelled or only slightly lower compared with the floor in the hall of the glass workshop. On the northern long side of the furnace, there were two working places adjacent to two apertures in the furnace. The western working place was not operating at the time of the visit, and the corresponding aperture was closed with bricks and mortar. When in use, this opening served as a glory hole where also, raw glass could be shovelled into the melting chamber. The aperture could be closed by a big rolling shutter of stone, served by the person sitting at the corner of the furnace. The eastern working place, marked with ‘1’ on the plan, was in use and constructed as a slender vertical slot. This opening was partly bricked up until a little above the surface of the molten glass. The bricks were removed one by one as the level of the glass batch went down as the result of the glass being used, to maintain the bottom of the working slot near the level of the surface of the glass to keep the access to the glass convenient for the glassblower. When not in use, the slot was closed by a shutter that was operated by a worker sitting at the north-west corner of the furnace. At the western end, there was also a working place where there was a big opening into the furnace. At the east corner of the southern long side, marked with a ‘4’ on the plan, a small opening in the furnace allowed for the melting of small portions of lead.

306

J. Kock and T. Sode

Fig. 3 Plan and cross section of glass furnace in the northern glass works in Kapadvanj. Drawing: Sven Kaae

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

307

Fig. 4 Interior of the tank furnace in Kapadvanj. The melting chamber in the foreground

In the furnace, was melted about three tons of glass at a time. The melting takes approximately 36 h. Of course, the melting time was longer, about three to four days, when the furnace was heated for the first time after it had been closed down for repair or at the beginning of the blowing season. Glass was blown continuously for six days or until the furnace was empty. Two glassblowers at a time were working in shifts of four hours. In one shift, they would produce 60–70 globular pieces of glass. When the blowing was finished after six days, the furnace was allowed to cool down for four hours before new raw material was shovelled into the melting chamber. Two tons of firewood and one ton of coal were needed to heat the furnace each day at a cost (in 1999) of, respectively, INR 1500/- and 2500/-.

3 Blowing the Glass Mirrors Although the northern workshop was closed, the technology was preserved and is still practiced in the second workshop. The blowing of a glass sphere is the first step for the fabrication of the mirror. The process starts in the working area next to the furnace, marked ‘1’ on the plan of the furnace. The worker is sitting to the left of the opening and has a marver to the right of it (Fig. 5). Before gathering the glass, the worker has to make sure the surface of the glass batch is clean, if not, he will remove debris using a tool in shape of an angled iron on a long iron stick.

308

J. Kock and T. Sode

Fig. 5 Glassblower has gathered glass on the blowpipe from the slender opening at the side of the furnace

First, the glassblower gathers a small parison on a blowing pipe and then handles it over to the master glassblower at the front of the furnace who places the pipe near the opening of the furnace to keep it warm. It will be used later on as a pontil. Then, the blowing of the globe itself can start. Little by little, glass is gathered on a second glassblower’s pipe until there is enough glass for blowing a glass sphere with a diameter of 50–60 cm (Fig. 6). The pipe with the parison that weights a little more than one kilo is now handed over to the master glassblower who sits at the front of the furnace (marked ‘2’) on the elevated platform, leaning a little back. This way he gets a convenient working position for manipulating and blowing the globe as well as for using the vertical and the oblique marver opposite from the rolling shutter. First, the glassblower heats the parison in the glory hole and in several steps blows the glass globe to its final size (Fig. 7). The pipe is then carefully broken off the globe, which leaves a small opening where the blowing pipe was. The globe, resting on the platform before the glory hole, is then turned 180 degrees, and the prepared blowing pipe acting as pontil is attached to the opposite side of the opening in the globe (Fig. 8). The globe is reheated once to smooth the sharp edges where the pipe was broken off. Now, the globe is ready to be coated with lead on the inside. The molten lead is in a crucible, placed next to a small opening of the furnace, marked ‘4’ on the plan. By the furnace, a pile of round ingots of lead weighting one kilogram, each one of them sufficient for mirroring exactly one globe, are ready to be used. The lead ingots are melted one by one in the crucible.

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

309

Fig. 6 Glassblower sitting on an elevated platform blowing a half-finished globe. He is leaning back slightly in characteristic working posture

Fig. 7 Reheated globe is blown to desired size. To the right is the hot pontil ready for use

310

J. Kock and T. Sode

Fig. 8 Blower attaches the pontil opposite the hole visible after the blowing pipe is detached

The lead-backing, or mirroring, takes place in front of the furnace in the area marked ‘3’ on the plan. To allow a convenient working position, the pontil with the globe still attached to it is sunken half the way into a hole in the floor (Fig. 9). The lead is poured into the hot globe while it is rotated continuously (Fig. 10). When the pouring is finished, the pontil with the globe is removed from the hole in the floor, and now the glassmaker rotates the globe slowly to coat the inside of the sphere with the lead. It is easy to handle the hot glass globe and the lead with the iron pontil. When the entire globe is coated with lead, it is placed on a tripod with the opening facing down above the cup used earlier to melt the lead in order for the excess metal to be collected. It is later recycled (Fig. 11). According to the owner of the glass works, the metal used is remolten recycled lead to which 50 g of zinc and 50 g of tin are added per kilo of lead. Then, the pontil is broken off. It does not matter if the globe coated with lead cracks because the globe has to be broken into relatively large pieces when cooled. The production does not leave much waste because the pieces which are too thick or have not been sufficiently or too heavily coated with lead are re-molten. Usually, glass must be annealed in a furnace; but in this case, it is not necessary as the glass sphere was blown so thin. Among a lot of other things, the very knowledgeable owner tells us that they hardly ever fail when they make mirrors by coating the inside of a glass bubble with lead; but they have not been very successful when it comes to backing plane glass with lead.

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

311

Fig. 9 Blower rotates the globe while the assistant carefully pours molten lead into it

The lead-coated glass is broken or cut into pieces and packed in cardboard boxes and then transported nearly hundred kms towards the west to the town of Limbdi where it is sold to merchants (Fig. 12). The pieces of glass mirrors are also sold to merchants in Bhuj in Gujarat and in Hyderabad in Andhra Pradesh as well as to village people in Gujarat who then cut them into round-shaped mirror pieces. The merchant that we visited and interviewed in 1997 used to buy five tons of mirrored glass four to five times a year. He had to pay rupees 30/- a kilogram and then pay for the transportation and 14% of sale tax. In Limbdi, able women carry out the cutting of mirrors (River, 1993a). First the glass is cut with an agate stone known as a khambat into longish pieces or strips varying in size from one to a few cms (Fig. 13). The width of the strips gives the diameter of the finished mirrors. Afterwards, the strips are cut into squares. The small square pieces are then given the final round shape by using an iron scissor (Fig. 14). It is quite surprising that they cut the mirrors with large loosely riveted pairs of scissors, which look like trimmers. Depending on the size and the wanted quality, a woman can cut approximately 2 kg in one hour. The cut pieces of mirrors are sold according to quality all over the Indian subcontinent. In 1999, in Limbdi, about 150 women were engaged in cutting mirrors. The small convex pieces of glass mirrors are finally sold to the end users, who use the mirrors in traditional mirrorwork and in textiles with mirror embroideries known as abhala bharat 1 (Fig.15) (Rivers, 1993b). Among the locals of Kutch, such mirrors decorate 1

It was once common that many tribes, and among others the Rabaries, had their own design, and for couples, as part of their wedding, to get dress and equipment with many mirrors stitched on by woman from the family or the village. The standard of the handicraft was very high. They were proud of their skills.

312

J. Kock and T. Sode

Fig. 10 Globe is rotated systematically for lead coating

The traditional use of this is quickly diminishing, and the modern international culture is taking over. Various museums have collected good examples and is now presenting the phenomenon. Also there are many private collectors as well as shops specialized in collecting and selling these things. Today at the night market in Ahmedabad, you will find quantities of new but traditional textile. Often they are of a very poor quality to keep the prices modest and the earning as high as possible. Old pieces are rare. A group of woman from Kutch has a row of stalls near the Tibetan Market at Janpath in New Delhi selling old textile with mirrors and beadworks from Kutch. A part of it is old genuine pieces, others only look old but are new and finally there are new but traditional textiles with mirrors.

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

313

Fig. 11 Excess lead is collected after the globe has been coated

Fig. 12 Produced globes are broken into big pieces, sorted and packed in cardboard boxes. Unusable pieces are re-melted

314

J. Kock and T. Sode

Fig. 13 a Mirror strips, b cut into squares, c trimmed mirrors, d a perfect trimmed mirror, e a mirror in profile with the characteristic curved surface line

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

315

Fig. 14 Woman in the bazar (market) of Limbdi shaping blanks to mirrors in desired size with a heavy scissor

316

J. Kock and T. Sode

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

317

Fig. 15 Textiles and other items with embroidered or attached mirrors (a interior of a house in Bhuj, Kutch with a traditional textile with mirrors for the door; b mud and mirror work in a traditional round house of the Rabari tribe. Reconstructed in The Folk Art Museum in Bhuj, Gujarat; c a shop in Bhuj specialised in selling old genuine stitched textile; d Detail of Fig. c; e a mirror embroidered piece for the wall; f wall decorations around the window with white plaster and mirrors in a traditional circular house in Kutch; g mirror worked blouse at Kutch; h textile with mirrors for the wall, bought at the Tibetan Market in New Delhi; i hat with sparkling mirrors at Kutch; j a woman stitching mirrors on a piece of textile in Bhuj, Kutch; k moderate stitching; l fashion with mirrors in a shop in Gujarat; m touristic textile with mirrors at the Night Market in Ahmedabad; n camel decorated with mostly modern flat glass, camel parade in Jaisalmer, Rajasthan. Photos: Lis Kobbelgaard)

the walls (Fischer, 1995; Shah, 2013), and in late medieval time, they were used in decorating palaces in western India.2 For instance, in Rajasthan the palace of mirrors (Sheesh Mahal) in Amber was built by Mirza Raja Jai Singh in 1639. In the nineteenth century, decorating the interiors of the palaces with gilt stucco finely modelled around mirror of glass and coloured tinsel covered with mica (Fig.16) became a new normal among the maharajas and the noblemen (Taschen, 1999).

4 New Knowledge The study of the glass mirror makers in Kapadvanj has been an important supplement to our knowledge about mirror making in medieval Europe. This research is part of an extensive ethno-archaeological project to study and document traditional glassmaking techniques. It is often the case that glassmaking techniques that once were used in Europe are still executed in traditional glass workshops in India. The final goal of this project is to understand the applied technologies from our past better and at the same time to document them in places where they have been kept alive while it A traditional culture is disappearing and with this the need of mirrors from Kapadvanj. Also it is at the same time easier to use mirrors of flat glass from modern industry and mirrored plastic. 2 The architectural use of mirror work originated from seventeenth century India and was introduced in India during the reign of the Mughal Empire. The luxurious tastes of the Mughals found expression in the glittering mirror rooms with which they furnished their palaces. As with many of India’s intricate crafts, the techniques used in mirror work were passed down from one generation to another. Both Mughal and Rajput palaces include apartments known as shish mahals, or mirror pavilions, intended for visual delight. The walls, vaults and domes are most often entirely covered with small convex mirrors set in stucco in geometric patterns. The ceilings, niches and windows are in the same way decorated with mirrors together with pieces of coloured glass and other decorative elements. The Amber Palace was built by king Man Singh in sixteenth century and completed in 1727. The Sheesh Mahal was reserved for personal use by the imperial family. By lighting a few candles, the reflection converts the light into the impression of thousand stars. The central part of the mirror pavilion is ornamented with mirrorwork, flat European glass with reverse glass painting and marble panels often with floral decorations. The walls, vaults, niches and ceilings are adorned with convex mirror work and colourful glass, set in geometric patterns to create dazzling effects. Larger European flat mirror glass was used in walls and ceilings as well.

318

Fig. 15 (continued)

J. Kock and T. Sode

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

319

Fig. 16 Mirror pavilions in Mughal Palaces (a the fine mirror works in geometric pattern in the Sheesh Mahal in Amber Fort; b mirrored wall decoration in the Jai Mandir, the hall of Public Audience, Amber Fort; c the glittering convex mirrors in the dome of Sheesh Mahal; d mirrored ceiling in the Sheesh Mahal; e the central mirror room in Samode Palace, ornamented with mirrorwork set in stucco, reverse glass painting and white marble panels; f details of mirror work on the walls and ceiling in Samode Palace; g niche with mirrored vaults in Samode Palace and h mirrored archway decoration, detail from the City Palace, Jaipur)

320

J. Kock and T. Sode

Fig. 16 (continued)

is still possible (Kock & Sode, 1995: 2). This ethno-archaeological research can be used with great advantage to interpret the crafts of the past to compare them critically with archaeological finds and early written sources. As a result, it is possible to set up models of ancient and medieval manufacturing methods which otherwise could not be illustrated. In this connection, the study of the glass mirror makers in Kapadvanj is interesting. During several visits to the site, it has been possible to study and document the manufacturing process in detail. It has given us a better understanding of the manufacturing techniques for the blowing and the inner coating a glass globes with hot molten lead in order to make mirrors. We now know that this is not a problem as long as it is done, while the glass sphere is still warm but not too hot. Another realization is that the thin glass mirror can simply be cut with a pair of scissors. The small convex mirrors from the Viking Age and Early Medieval Time show that this was done in the past also in Europe and the Near East (Kruger 1990: 292; Kock, 1999; Sode & Kock, 2001). Acknowledgements We want to thank warmly the proprietor Mohammed Siddiq Shisgar and all the staffs of the northern glass works for their enthusiasm and hospitality. The same also for the

Traditional Glass Mirror Making in Kapadvanj, Gujarat, India …

321

owner of the southern glass work who opened the door for us in 2019. In the same way, we want to thank our colleague at the National Science Centre, Delhi and Ahmedabad. A special thanks goes to Mrs. Sonal Metha, Exe. Director Eklavya Foundation, Ahmedabad, a socially engaged connoisseur of the ethnic and nomadic textile and craft traditions in rural Gujarat. Thanks to Dr. Bob Brill, Corning Museum of Glass, N.Y., USA, for sharing with us his knowledge on traditional Indian glass manufacturing.

References Dikshit, M. G. (1969). History of Indian glass. University of Bombay. Fischer, N. (Ed.). (1995). (reprint). Mud mirror and thread: Folk traditions of rural India. Grantha Corporation / Museum of New Mexico Press. Kock, J. (1999). Spejlet i middelalderen. In T. Madsen og Jens Vellev (Ed.), Ole Højris, Hans Jørgen Madsen (pp. 216–220). Menneskelivets mangfoldighed. Arkæologisk og antropologisk forskning på Moesgård. Højbjerg. Kock, J., & Sode, T. (1995). In Glass, glassbeads and glassmakers in Northern India. Vanløse. Krueger, I. (1990). Glasspiegel im Mittelalter Fakten, Funde und Fragen (pp. 292). Bonner Jahrbücher 190, Bonn. Maconochie, E. (1895). A monograph on the pottery and glass-ware of the Bombay presidency. Government Central Press. Mukharji, T. N. (1888). Art-manufactures of India: Specially compiled for the glasgow international exhibition. Supt. Of Govt. of India Print. Rivers, V. Z. (1993a). Glass art from India, The India Magazine, 13. Rivers, V. Z. (1993). Indian mirror embroidery from Gujarat. Ornament, 16(3), 66–69. Shah, A. (2013). Shifting Sands, Kutch - A Land in Transition. Bandhej Books. Sode, T., & Kock, J. (2001). Traditional raw glass production in Northern India: The final stage of an ancient technology. Journal of Glass Studies, 43, 155–169. Taschen, A. (1999). Indian interiors. Taschen Verlag.

Glass Products in South Asia

Glass Beads of Eastern India (Early Historic Period) Sharmi Chakraborty

Abstract The archaeological record of beads from eastern India has been sketchy and unstandardized. Here, an attempt is made to present a cohesive picture emphasizing regional patterns and preferences. When possible, glass beads are chronologically situated on the basis of excavated sites with firm chronological sequence. A general distribution of the beads is given according to colour, shape and manufacturing method when available. It appears from the current dataset that there is a distinction between the coastal sites and the inland sites. While the inland areas have more of green and blue glass, the coastal area has opaque red, yellow and orange. The barrel shapes predominate but an important proportion of cylinder circular and disc beads was also noted. In contrast, facetted beads are few. An interesting feature for eastern India is the presence of two types of composite glass bead that tend to imitate agate.

1 Introduction Beads are units that make up stringed ornaments and are used in the decoration from headgear to footwear. Their variety and availability increase many folds in the early historic phase. The early historic phase, as widely accepted, encompasses the second phase of urbanization of the Indian subcontinent. While beads may not be a major criterion for urbanization, the demand for beads of various kinds can be associated with prosperity and display of such. In the context of ancient India, this form of behaviour acquires a special significance. The literature of this period glorifies as well as sets the standard of an ideal urban man the ‘nagaraka’ and his ‘nayika’ (his beloved). They are expected to be an expert in the sixty-four arts which includes the art of grooming. Grooming and dressing up are not complete without ornaments, and beads appear to have been a prominent part of this display. The stone sculptures and terracotta plaques of the period testify to this (Fig. 1). These beads,

S. Chakraborty (B) Centre for Archaeological Studies and Training, Eastern India, Calcutta, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_13

325

326

S. Chakraborty

Fig. 1 Beaded female figure with attendant (private collection)

as archaeological evidence suggests, fall into five different categories—stone, glass, terracotta, bone/ivory and seeds. This paper deals with only glass beads. The eastern part of the subcontinent has been variously defined, with the outer limits or boundaries varying with various kingdoms. For the early historic period, one may try to delineate linguistically a predominantly Pali speaking east against a Prakrit speaking region. For this paper, the states of Bihar, West Bengal and Odisha

Glass Beads of Eastern India (Early Historic Period)

327

have been included. The north-eastern states are excluded due to lack of properly documented archaeological sites of early historic period. This includes Assam where chronology and periodization are yet to be ascertained and scientific documentation of beads is not found in publications. There are hardly any publications with a detailed account of beads at early historic sites in Jharkhand, as well as the Andaman and Nicobar Islands. Bangladesh is included for being a part of ancient India, focusing on two important sites of the early historic period viz., Mahasthan and Wari Bateswar (Fig. 2). There is no recorded glass bead from eastern India prior to the early historic period.

Fig. 2 Important glass bead yielding sites of ancient India discussed in the paper

328

S. Chakraborty

The distribution of glass beads reflects exchange networks and choice. Due to the wide availability of raw material for glass (except the colouring agents), glass is generally produced at a regional level rather than brought from afar. At the same time, since glassmaking and glass working were relatively new technologies in the early historic period, manufacturing units were limited. While distribution patterns, as representation of trade and exchange networks, have often been evaluated and discussed, the role of choice and preference is little understood. The following discussion explores both the issues in the context of eastern India and their implications.

2 Problems of Tracing Ancient Distribution Archaeology is the means of reconstructing the past from the material that is intentionally deposited, abandoned, or lost and which has been preserved from the ravages of time. These materials are discovered by archaeologists either through exploration or excavation of a selected area. However, artefacts available for study are only part of what people used in the past. Data on glass beads from the eastern region of India found in available publications (excavations reports) lack a standard method of documentation, absolute dates and sometimes even have no mention of the cultural phase of the beads. Although, the shapes, sizes and colours throw light on ancient sensibilities and practices, shapes are poorly recorded and Munsell charts are hardly used for description of colours. H. C. Beck’s chart (1928) is followed very loosely with no reference to size which is important to ascertain the exact shape. Beads are often relegated to the position of ‘minor antiquities’ thus neglected as a source of information. Therefore, we most often have only a presence/absence type of data with no finer classification and chronological division. The following preliminary analysis and hypotheses are limited to colours, shapes and techniques and are derived within the backdrop of the above-mentioned limitations.

3 Eastern Indian Glass Beads The glass beads from this region can firstly be grouped under monochrome and composite varieties. The second grouping can be done on the basis of colour, though at times colour of the glass is not mentioned in the reports.

3.1 Colour 3.1.1. The green glass beads are more common (Fig. 3). In the larger context of

Glass Beads of Eastern India (Early Historic Period)

329

Fig. 3 Opaque green bead from Deulpota (scale in cm), Courtesy Bernard Gratuze

general distribution in India, the inland sites have yielded more green glass beads than the coastal sites. From Mahasthan (Boussac & Alam, 2001), there is reference to dark and light green glass. From the Gangetic delta green glass beads are found in semi-transparent glass with fewer numbers are opaque. A rare pale translucent green variety of long hexagonal cylinder bead was found at Harinarayanpur. 3.1.2. Blue glass beads are present inland but are more common in the Gangetic delta. Blue glass has a few shades—light blue, blue, deep blue and blue green (turquoise—as mentioned in Mahasthan report) (Fig. 4). But only blue is mentioned from Kumrahar (Altekar & Mishra, 1959), Sonpur (Sinha & Verma, 1977), Vaisali (Krishna Deva

Fig. 4 Blue beads from Mangalkot (scale in cm), Courtesy Bernard Gratuze

330

S. Chakraborty

& Mishra, 1961) and Pataliputra (Sinha & Roy, 1969) in Bihar as well as Wari Bateswar (Rahman, 2003) in Bangladesh. A few cobalt blue glass beads are noticed from Chandraketugarh. 3.1.3. There are other colours which are fewer in number like opaque white, yellow, purple and black (Fig. 5). Fifty per cent of beads found at Chirand are black in colour (Verma, 2007). 3.1.4. There is another group consisting of the drawn glass beads with ocherous colour—ocherous red (mentioned henceforth as Indian red), yellow ochre and orange ochre (Fig. 6). These are known from Mahasthan, Wari Bateswar (Rahman, 2003), Tamluk (Gangopadhyay, 2003), Chandraketugarh, Mangalkot, Harinarayanpur and Deulpota (Chakraborty, 2000). Excavations at Sisupalgarh yielded respectable number of Indo-Pacific bead of Indian red colour (personal observation during 2004–05 excavation season). Identical beads with black interior are noticed at Chandraketugarh, Directorate of Archaeology and Museum West Bengal (DAWB) collection. Fig. 5 Black bead from Mahasthan (scale in cm), Courtesy Bernard Gratuze

Glass Beads of Eastern India (Early Historic Period)

331

Fig. 6 Assortment of Indian red and orange ochre beads from Mahasthan (scale in cm), Courtesy Bernard Gratuze

3.1.5. Composite glass beads are fewer in number and this is true for all the sites. The Pataliputra excavation report mentions a single case, which could possibly be a millefiori bead (Sinha & Roy, 1969). Millefiori was also found from Mahasthan (Boussac & Alam, 2001) (Fig. 7) as well as Deulpota (Chakraborty, 2000). The colour is green with red, black yellow and white ladder type design. Recently, Kanungo (2019) has interpreted these to be mosaic patterned beads rather than millefiori. Identical beads were found at Kopia (BCE-CE transition period) in Sant Kabir Nagar district of Uttar Pradesh (Kanungo, 2013: 332) and at Khlong Thom in Thailand (second-seventh century CE) (http://www.thebeadsite.com/b-lthai.htm). From Mahasthan, a few more types of composite glass are found—marine blue with white vertical stripes, white with white vertical stripes, white and black—all in a short barrel shape. The site has also revealed a triangular piriform-shaped opaque blackish green glass bead with three eyes (Boussac & Alam, 2001). 3.1.6. A convex cylinder bead with etched band is reported from Kumrahar (Altekar & Mishra, 1959) which appears as a mimic agate bead. Mahasthan (Boussac & Alam, 2001), Chandraketugarh (Chakraborty, 2000) and Tamluk (Gangopadhyay, 2003) have also yielded similar type in black glass, shape being barrel circular or oblate, on which there is inlaid spiral design in white glass (Fig. 8). Picture (Plate LXXXI: 3, 4) from Kumrahar suggests that they are of same type. Banded black glass bead has also been reported from Chirand (Verma, 2007).

332

S. Chakraborty

Fig. 7 Ladder patterned bead from Mahasthan (scale in cm), Courtesy Bernard Gratuze

3.1.7. Sandwich beads of white wavy band on black/dark violet mimicking agate was observed in Chandraketugarh (Chakraborty, 2000), Wari Bateswar (Haque et al., 2001), Mahasthan (Boussac & Alam, 2001), Kumrahar (Altekar & Mishra, 1959) and Harinarayanpur (Fig. 9). A blue and light blue sandwich bead was also noticed at Chandraketugarh. 3.1.8. The site of Kumrahar (Altekar & Mishra, 1959) has also yielded gold foil beads, which are found in many sites of North India as well as in eastern India (Fig. 10). 3.1.9. Only a single cemented eye bead is known from this region, and it was found from Tamluk (Gangopadhyay, 2003). It appears that the ochre-coloured beads of yellow, orange and red are more common in the coastal sites compared to further inland. The other important group is represented by blue, green and blue green, which are more common in inland. Green is more prominent among these varieties. However, in this group, some are also drawn glass bead. Composite beads are more common in inland, though they are present in coastal sites.

Glass Beads of Eastern India (Early Historic Period)

333

Fig. 8 Black bead with white spiral inlay from Mahasthan (scale in cm), Courtesy Bernard Gratuze

Fig. 9 Broken purple and white sandwich beads found at Harinarayanpur (scale in cm), Courtesy Bernard Gratuze

334

S. Chakraborty

Fig. 10 Gold foil bead with a collar barrel circular shape from Chandraketugarh (private collection) Courtesy Bernard Gratuze

3.2 Shape In the published literature as well as museum and private collections, short circular barrel beads are the most common shape across the different colours. Possibly because this shape is easier and faster to obtain in manufacturing. The circular cylinder and circular disc beads are also found in good numbers. Later, drawn glass beads of Indian red, orange and yellow ochre colours are mainly found (Figs. 11, 12 and 13). Facetted beads are very few in number. The most common shape in this group is a hexagonal long cylinder. This is found in blue and green colours. A pale blue variety almost transparent with a thin perforation has been found. These belonged to a later period (post CE 200) as suggested by the Mahasthan excavation where it was found from levels 12 and 13 (Boussac & Alam, 2001). The other site from which this variety is known is Chandraketugarh (Chakraborty, 2000) and possibly also from Sonpur (Sinha & Verma, 1977). These hexagonal-shaped beads, particularly of the green variety, are significant among beryl beads of South India (Francis, 2000–2001). These might have been imitation of those beads.

Glass Beads of Eastern India (Early Historic Period)

335

Fig. 11 Number of cylinder circular beads from lower Bengal by colour and by site. HNP stands for Harinarayanpur, CKG for Chandraketugarh, DPT for Deulpota and MGT for Mangalkot

Fig. 12 Number of short barrel beads from lower Bengal by colour and by site

3.3 Shapes Found in Different Colour Groups 3.3.1. Green beads from the Kumrahar excavation are found in a variety of shapes, including: tablet, spherical, hemispherical, elliptical and long barrel collared.

336

S. Chakraborty

Fig. 13 Number of circular disc beads from lower Bengal by colour and by site

(Altekar & Mishra, 1959). In Sonpur (Sinha & Verma, 1977) oval, cornerless hexagonal, long hexagonal cylinder shapes were found. From Chirand hemispherical, cornerless hexagonal, lug collared, elliptical, short cylinder and bicone shapes are found (Verma, 2007). Mahasthan (Boussac & Alam, 2001) has yielded spherical, collared spherical, prismatic (Fig. 14), facetted sphere, cornerless cube, tubular, annular, diamond, etc... beads. But from lower Bengal shapes are usually simple like short barrel, cylinder circular and spherical, though collared beads are also known (Fig. 15). From Chandraketugarh (Chakraborty, 2000), Mahasthan (Boussac & Alam, 2001) and Wari Bateswar (Rahman, 2003) ‘unpierced’ spherical beads were noticed (Fig. 16). Gadrooned bead is available from Mangalkot (Chakraborty, 2000). Bangarh (Goswami, 1948) yielded a flat barrel bead. 3.3.2. Black glass beads have been found in Mahasthan (Boussac & Alam, 2001), in circular, oblong, ovoid, spherical facetted, conical, annular and tubular shapes. Shapes are not mentioned for the Wari Bateswar site (Haque et al., 2001). Chandraketugarh has yielded circular, long barrel and facetted circular shapes and Deulpota yielded oblong and oblate beads (Chakraborty, 2000). From Tamluk (Gangopadhyay, 2003), the short barrel shape has been reported. Chirand (Verma, 2007) excavation has revealed a good number of black beads in hemispherical, long barrel, short barrel and elliptical shapes. Most of these are from the Kushana1 phase.

1

The Kushana kingdom dates from c. CE 30 to c. CE 375, until the invasions of the Kidarites. They ruled around the same time as the Western Satraps, the Satavahanas, and the first Gupta Empire rulers.

Glass Beads of Eastern India (Early Historic Period) Fig. 14 Prismatic opaque green bead from Mahasthan (scale in cm), Courtesy Bernard Gratuze

Fig. 15 Broken collared tablet bead from Deulpota (scale in cm), Courtesy Bernard Gratuze

337

338

S. Chakraborty

Fig. 16 ‘Unpierced’ bead from Mahasthan (scale in cm), Courtesy Bernard Gratuze

3.3.3. Purple glass beads are rare. These have been reported from Chandraketugarh (Chakraborty, 2000) where they have different shapes—square barrel, circular ellipsoid, hexagonal ellipsoid and collared ellipsoid. From Mahasthan (Boussac & Alam, 2001), spherical, spherical facetted, annular, collared oblate, tabular cylinder shapes are found. 3.3.4. White opaque beads are usually found in the forms of short barrels. A single specimen of segmented variety is found at Deulpota (Chakraborty, 2000). Another segmented bead is reported from Vaisali (Sinha & Roy, 1969) but the colour is not mentioned. 3.3.5. Yellow beads are also mostly found in the circular short barrel type, although some are a circular cylinder shape. They are found at the sites of Mahasthan (Boussac & Alam, 2001), Chandraketugarh, Deulpota, Harinarayanpur (Chakraborty, 2000) and Tamluk (Gangopadhyay, 2003). 3.3.6. There are four shades of blue beads—blue, light blue, deep blue and blue green. In all four varieties, the circular short barrel shape predominates followed by circular cylinders. In blue and blue green, a few facetted shapes are noticed. At Chandraketugarh, Harinarayanpur (Chakraborty, 2000) and Wari Bateswar (Rahman, 2003), cornerless cubes were found among the blue beads. A collared barrel was found in Chandraketugarh and Deulpota and short hexagonal bicone from Harinarayanpur in blue-green variety (Chakraborty, 2000). From Mahasthan (Boussac & Alam, 2001), shapes like oblong, ovoid facetted, conical, tubular, collared oblate, facetted tubular, melon shaped, etc. are found. In Mahasthan, although the number of blue beads is less than what is found at the Gangetic delta region, a larger variety of shapes was recorded. Blue and light blue glass beads are reported from Bangarh (Goswami,

Glass Beads of Eastern India (Early Historic Period)

339

1948) and all in globular shape. An unperforated blue glass bead was found from Chandraketugarh. 3.3.7. Almost all drawn glass beads of Indian red, yellow ochre and orange ochre have circular section. In this variety, cylinder and disc shapes are also well represented (Figs. 12 and 13). 3.3.8. Collared beads are found in green, blue-green, yellow and violet colours. These are in long barrel circular, standard triangular, standard rectangular and standard tubular circular shape. Collared glass beads have been found from Mahasthan (Boussac & Alam, 2001), Tamluk (Gangopadhyay, 2003), Chandraketugarh (Chakraborty, 2000), and lug collared variety from Chirand (Verma, 2007) and Deulpota. The Vaisali excavations (Sinha & Roy, 1969) do not mention the colour of the glass but bead shapes include long barrel cylinder, short cylinder circular, lenticular, oval flat, single collared gadrooned, diamond faceted cylinder and segmented.

4 Chronology For the chronology of the various types of beads, Mahasthan (Breuil et al., 2001) is taken as type site. The individual layers are well dated with a number of carbon dates. The earliest level for glass beads is level 5 which saw radical changes in the settlement according to the excavators. The charcoal samples dated between 366 and 162 BCE. The level 7 is dated to 173 BCE. Level 11 has been dated between CE 60 and 172. Level 12 is dated to CE 82 to 242 of which latest date have been considered more appropriate. Charcoal from pit of level 13 suggests ante quem c. CE 600. Level 14 is dated to a period between sixth and tenth century CE. Level 15 is dated to sixth to twelfth century CE. Level 16–18 have been dated between sixteenth and eighteenth century CE. Red, yellow and orange opaque is found from level 6 to 15. These are possibly part of Indo-Pacific bead group. Black opaque is available from level 6 to 13. Turquoise blue found from level 8 to 18. Deep blue found from level 5 to 14. Green glass is available from 5 to 12. Light green from 5 to 14. A few violet colour beads are found, and they are distributed in levels 8, 11 and 12. A gold foil bead appears in level 8 and millefiori from level 12. Barrel beads with spiral bands are found from levels 6, 7 and 8 but also later in level 11 (Boussac & Alam, 2001). The glass beads in this site originate in the fourth century BCE (Mauryan phase) and continued till early mediaeval period (Pala-Sena phase). It is important to note that violet colour beads are found only from the late levels of early historic phase. From Vaisali (Krishna Deva & Mishra, 1961), glass beads are available from Period III dated between 200 BCE to CE 200. At Kumrahar (Altekar & Mishra, 1959), Period III dated between 100 BCE and CE 300 yielded green, blue and black beads. At Sonpur (Sinha & Roy, 1969), glass is present in Period II which dates to the Northern Black Polished Ware (NBPW) phase and has been dated between 650

340

S. Chakraborty

and 200 BCE. The total number of beads available from this phase is 31. Period III dated between 200 BCE and CE 200 yielded 41 beads. Only a few beads have been illustrated and described. Green and black glass beads are dated from Period II while green, blue and black are from Period III. At Chirand (Verma, 2007), most of the glass beads are available from Period IV which is Kushana period. If we compare with at Sonkh (Hertel, 1993), perhaps the best dated site of North India, we find that green, blue green and blue glass appear in Period II dated to 2454 ± 124 BP. Yellow beads appear in the Kshatrapa phase and black, gold foil and millefiori in Period V, which corresponds to Kushana period. This generally agrees with the excavated finds from eastern India, except for the stray millefiori at Pataliputra dated to c. CE 1750, which must have been due to some disturbance or reuse of an earlier deposit. Sonkh excavation also shows that the Kushana phase is the richest in terms of number of glass beads. This is true for Mahasthan also, but unlike Sonkh most of the bead varieties continue into the later period. This may be due to continue of importance of Mahasthan in the post Kushana period.

5 Situating Eastern Indian Beads in Indian Context A better understanding of the glass beads from eastern India can be inferred by comparison with other bead corpuses from the rest of the country. Also, chemical analysis can help understand the provenance or place of production of the glass used for the glass beads found in this area. In absence of a numerical dataset from key sites like Patna, Vaisali and Sisupalgarh, the results from the statistical analysis are not very inclusive. Yet the author tries to see how the sites are aligned in a pan-Indian scale to ascertain if there is any ‘eastern’ quality in the distribution. Figure 17 shows a Pearson’s correlation2 cluster analysis, which includes sites from which numerical data are available: Rajghat, Sonkh, Ahichchhatra, Mangalkot, Deulpota, Chandraketugarh, Harinayanpur, Mahasthan, Veerapuram, Nevasa and Tamluk. The hierarchical clustering with Pearson’s correlation coefficient (Fig. 17) shows that the sites of South Bengal has clustered together while Mahasthan is with the rest of the sites. The cluster profile created from K Clustering (Fig. 18) clumps Mahasthan with the rest of India. We can see that these two groups diverge from the mean in terms of two distinct varieties of beads. Group one representing lower Bengal has more of the Indo-Pacific beads of Indian red, orange and yellow while the second group has more of blue, green, black, grey and cream. The gold foil and composite beads are above the general mean for the second group. Arikamedu which has a strong presence of Indo-Pacific beads is not included here because the descriptions available in the report are in a different format (Francis, 2004). 2

Pearson’s correlation coefficient is a technique for investigating the relationship between quantitative and continuous variables.

Glass Beads of Eastern India (Early Historic Period)

341

Fig. 17 Cluster analysis of beads found from important archaeological sites.

Fig. 18 K Cluster plot of the beads

Chemical analysis of beads from West Bengal, India and Bangladesh has revealed very interesting facts (Gratuze et al., 2015). The type of glass for beads found in South Bengal shows a higher percentage of glass from South India (m-Na-Al 1) while all the sites have glass from North India (m-Na-Al 3). Some gold foil beads have the same composition as that of Bara in Pakistan. There are Sassanian type of glass and

342

S. Chakraborty

Table 1 Distribution of glass types at sites in North India and Bengal ( taken from Gratuze et al., 2015) Mahastan Wari Harinarayanpur Mangalkot Deulpota Chandraketugarh Bateshwar Mixed-alkali 4 Cu disc

2

Mixed-alkali 8 Cu other

1

m-K-Ca-Al

1

6

m-Na-Al 1 m-Na-Al 3

19

m-Na-Ca-Al 1 v-Na-Ca-Al Bara

4 1

1

3

1 1

1

1

7

4

7

6

3

6

4

3

2

2

v-Na-Ca Sasanian

1

m-Na-Ca-Al or v-Na-Ca-Al Arikamedu Venitian or modern

1

2

8

3

2

Venetian glass from Deulpota and Harinarayanpur, which belongs to later phase of history, early mediaeval to colonial. While Chandraketugarh has yielded glass with a slightly different composition which may indicate local production or something to do with context of the finds, the authors feel that Mahasthan might have produced glass beads as well. A summary of the results of the beads analysed by Prof. Bernard Gratuze (CNRS Orléans, France) is given in Table 1.

6 Concluding Remarks In spite of the variable nature of the data coming from the sites of eastern India, the following can be surmised. 1. 2. 3.

Second century BCE to second century CE appears to be the most prolific. All the sites of Bengal make a distinct group due to the presence of Indo-Pacific beads of Indian red, orange and yellow ochre. The ‘interior’ regions like those of Bihar and Mahasthan (the latter was much nearer to shore than at present) have more green glass while blue is more common to sites in the Gangetic delta and Tamluk.

Glass Beads of Eastern India (Early Historic Period)

4. 5. 6.

343

The inland sites have revealed more variety in shapes for green and blue glass beads. Two varieties of composite glass are typical to Bengal and Bihar region—the spiral inlaid on black ovoid and wavy white sandwiched in blackish violet. Very pale blue and light green beads of transparent glass found in hexagonal cylinders and with thin perforation were found from Mahasthan (third century CE) and Chandraketugarh.

Coming to the question of distribution as a reflection of choice, the only strong case can be made for the two composite varieties mentioned above (No. 5). The apparent similarity with agate makes this proposition more interesting as agate is the most common stone used in making beads of West Bengal and Bangladesh. The preponderance of the short barrel type followed by disc beads also suggests preference for small and simple beads, which may be due to the ease of manufacturing or even depend on the way coloured beads were mixed and matched when an ornament was made. This at present is difficult to reconstruct. But if the terracotta plaques showing various beaded ornament including beaded decoration in the footwear (See Fig. 1) are any indicator, there seem to be a vibrant display of colour and shapes in which glass beads must have played a prominent part. The study of beads could become a window to many aspects of past culture. They can say something about a group’s aesthetics as a component of bodily decoration. They may be markers of identity. They also inform us about the technical capabilities of a particular culture. The network of procurement of raw materials and distribution of finished products also can be studied. Therefore, we need to refine our process of documentation. A standardized description of colour and shape is necessary. Beck’s chart of shapes and Munsell’s colour chart could be followed. The excavation reports need to incorporate the stratigraphic as well as spatial details of the beads. Then the laboratory analysis of samples and ethnographic studies of tradition craft could be undertaken to understand technology and exchange network. Acknowledgements I thank IIT Gandhinagar for giving me an opportunity to take a fresh look at the glass beads; Piyali Sengupta, Director DAWB, for permitting me to study beads in their collection; Chinmay Kulkarni, IITGN for making the Map (Fig. 2); Dr. Sharmila Saha and Sumita Guha for their help.

References Altekar, A. S., & Mishra, V. (1959). Report on Kumrahar Excavation 1951–55. Historical Research Series vol. III. K.V. Jayaswal Research Institute. Beck, H. C. (1928). Classification and nomenclature of beads and pendants. Society of Antiquaries. Boussac, M. -F., & Alam, Md. S. (2001). Beads. In Md. S. Alam, & J. -F. Salles (Eds.), FranceBangladesh joint Venture Excavations at Mahasthan: First Interim Report 1993–1999 (pp. 427– 495). Department of Archaeology, Ministry of Cultural Affairs, Government of the People’s Republic of Bangladesh and Mission Francaise de coopération archéologique au Bangladesh, Maison de I’Orient Méditerranéen- Jean Pouilloux, Lyon, Ministry of Foreign Affairs.

344

S. Chakraborty

Breuil, J. -Y., Salles J. -F., Rahman Md. H., & Shafikul, Md. A., (2001). Radiocarbon Dates from Mahasthan. In Md. S. Alam, & J. -F. Salles (Eds.), France-Bangladesh joint Venture Excavations at Mahasthan: First Interim Report 1993–1999 (pp. 219–228). Department of Archaeology, Ministry of Cultural Affairs, Government of the People’s Republic of Bangladesh and Mission Française de coopération archéologique au Bangladesh, Maison de I’Orient Méditerranéen- Jean Pouilloux, Lyon, Ministry of Foreign Affairs. Chakraborty, S. (2000). Chandraketugarh: An archaeological and cultural study (500BC to 500AD). PhD Thesis. Pune: Deccan College Post-Graduate & Research Institute. Deva, K., & Mishra, V. (1961). Vaishali excavation. Vaishali Sangh. Francis, P., Jr. (2000–2001). The Stone bead industry of South India. Beads: Journal of the Society of Bead Researchers 12–13, 49–62. Francis, P., Jr. (2004). Beads and selected small finds from the 1989–92 Excavation. In V. Begley, P. Francis Jr., N. Karashima, S.E. Sidibotham, & E.L. Will (Eds.), Ancient port of Arikamedu: New excavations and researches 1989–92 (pp. 447–604). Mémoires Archéologiques 22.2. Ecole Française d’Extrême-Orient. Gangopadhyay, K. (2003). Archaeology of early historic period in coastal West Bengal with special reference to Medinipur district. PhD Thesis. Bhubaneswar: Utkal University. Goswami, K. G. (1948). Excavation at Bangarh 1938–41. Calcutta: Asutosh Museum Memoir 1. Gratuze, B., Lankton, J., Islam, Md. S., & Boussac, M.-F. (2015). Glass beads of Mahasthan: Evidence for production and exchange. In Mahasthan II Fouilles du Rampart Est: Études Archéologiques under the direction of Jean François Salles. Brepol Publishers Turnhout. Haque, E., Rahman, S. S. M., & Ahsan, S. M. K. (2001). A preliminary report on Wari Bateswar trial excavation by ICSBA. In E. Haque (Ed.), Excavation at Wari-Bateswar: A preliminary study (pp. 11–41). The International Centre for the Study of Bengal Art. Hartel, H. (1993). Excavation at Sonkh, 2500 years of a town in Mathura district. Dietrich Reimar Verlag. Kanungo, A. K. (2013). Glass beads & bangles from Kopia. In A. K. Kanungo (Ed.) Glass in Ancient India: Excavations at Kopia. Kerala Council for Historical Research Kanungo, A. K. (2019). Chevron and Millefiori in India. Journal of the Borneo International Beads Conference, 2019, 69–88. Rahman, S. S. M. (2003). Wari Batesware Prapta Kancher Punti: Ekti Pratnatattvik Samiksha (in Bengali). Pratnatattva, 9, 1–10. Sinha, B. P., & Roy, S. R. (1969). Vaisali Excavation 1958–62. Patna: Directorate of Archaeological Museum Bihar. Sinha, B. P., & Verma, B. S. (1977). Sonpur Excavation 1956 and 1959–62. Patna: Directorate of Archaeology and Museum Bihar. Verma, B. S. (2007). Chirand Excavation Report: 1961–1964 and 1967–1970. Patna: Directorate of Archaeology (Department of Youth and Cultural Affairs, Govt. of Bihar).

A Review of Selected Glass Bead Types from the 2007–2009 Seasons of Excavation at Pattanam, India Shinu Anna Abraham

Abstract Over the last decade, excavations at the archaeological site at Pattanam in Kerala, India, have produced a range of material data related to South India’s involvement in early Indian Ocean exchange networks. One such category is glass beads, an assemblage that now numbers over 100,000 beads. By far the most common type of glass bead from Pattanam is the Indo-Pacific bead, a small monochrome drawn bead that was likely produced in South Indian or Sri Lankan glass bead workshops. But along with Indo-Pacific beads are a variety of other distinct glass bead types, most of which were not made by the drawing method and whose production sites are less certain. A number of them have parallels from other sites in South Asia and elsewhere along the Indian Ocean littoral. These bead categories include gold glass beads; large faceted types that may imitate beryl gemstone beads; polychrome ‘zone’ beads that mimic banded agate beads; large drawn barrel beads and terracotta disc beads. This chapter focuses on the glass bead assemblages from the 2007, 2008 and 2009 excavation seasons. Each bead type is briefly described in terms of their typology and known regional/chronological distributions, as a first step towards understanding Pattanam’s position within a complex suite of inland and maritime exchange networks.

1 Introduction Pattanam is a 45 ha port site located along India’s southwestern coast in the state of Kerala. It was first identified through extensive surface remains (Shajan et al., 2004) and trial trenches in 2004 (Selvakumar et al., 2005). Formal horizontal excavations began in 2007 under the guidance of Dr. P.J. Cherian, then-director of the Kerala Council of Historical Research (KCHR), and continued for nine seasons until 2015. (Excavation has recently resumed in 2020.) Stratigraphy is considerably disturbed at the site, but an array of material remains suggests a cultural sequence spanning from the fifth century BCE to c. fifteenth century CE (the later medieval period). The S. A. Abraham (B) St Lawrence University, New York, USA e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_14

345

346

S. A. Abraham

phase most intensively represented appears to be the early historic, roughly from the third century BCE to the fifth century CE (Cherian et al., 2016: 37). The site has produced extensive evidence for contacts with sites outside of South Asia around the western Indian Ocean, including Roman Egypt and western Asia (Cherian, 2012; Shajan et al., 2005; Sidebotham, 2011).

2 The Pattanam Glass Bead Corpus The Pattanam glass bead (PGB) project was developed to catalogue the glass beads recovered from the Pattanam excavations. As mentioned, nine years of meticulous excavation at Pattanam have produced an assemblage of over 100,000 glass beads. In terms of both quantity and curation, this makes the PGB corpus one of the largest known to date from an archaeological context, at least for the greater Indian Ocean region. So far, a little over 8100 of the Pattanam glass beads have been catalogued, from the 2007, 2008 and a portion of the 2009 seasons. In addition to excavation data (trench, layer, depth, chronological phase), recorded data include stylistic attributes (colour, opacity), formal attributes (shape, size, condition) and technological attributes (manufacturing methods, post-manufacture modifications). From a material perspective, beads have a range of traits that lend themselves well to study: ubiquity, durability, quantity and variability. The various attributes of glass beads also have analytic value—size, colour, opacity, shape, manufacture, postmanufacturing modification and chemical composition, just to name a few. But the complications of bead study are as numerous as their advantages, and it is the same qualities which make them a valuable resource that also confound the archaeologist: among other concerns, their durability and persistence raise issues of chronological reliability, their quantities create debate about appropriate scales of analysis, their small size can defy even the strictest field collection strategies, and their aesthetic appeal makes them a common target for looters. Nevertheless, as a proxy for the materiality of trade and exchange in the Indian Ocean, glass beads have become an invaluable source for new insights. They are also helpful for shedding light on the socio-economic landscapes of early Tamil South India. By building on what is known about glass production, glass bead making and exchange within South India and throughout the Indian Ocean world, the documentation and categorisation of the Pattanam glass beads may help refine our understanding of the South Indian contexts in which these beads were found.

2.1 Indo-Pacific Beads from Pattanam The majority (up to 75%) of the glass beads from Pattanam are what are commonly referred to as Indo-Pacific beads (Fig. 1), which have been discussed extensively elsewhere (Abraham, 2016; Abraham & Christie, 2010; Abraham et al., 2011; Carter,

A Review of Selected Glass Bead Types from the 2007–2009 …

347

Fig. 1 Indo-Pacific Beads

2016; Francis, 2002; Kanungo, 2016; Wood, 2016). Indo-Pacific beads are small monochrome glass beads, made by a drawn tube ‘lada’ method, where a thin tube of glass is pulled or drawn from a furnace-heated glass gather and then sliced into beads that are most often less than 6 mm in diameter and length, and very often much smaller. The technique produces enormous quantities of very small and very similar beads. The colours have a limited range, and the most common at Pattanam are opaque black, opaque red, translucent blue and translucent green. The lada drawn tube method likely has its origins in South India and/or Sri Lanka (Dussubieux et al., 2010), but Indo-Pacific beads have been recorded from a variety of sites within South Asia and across the Indian Ocean littoral and beyond, including western Europe and Japan (Katsuhiko & Gupta, 2000; Lankton & Dussubieux, 2013; Lischi, 2018; Pion & Gratuze, 2016; Wood, 2011, 2016). Dating Indo-Pacific beads continues to be a vexing issue, both because their stratigraphic integrity cannot always be ascertained and because techniques, styles, colours and even their use seem to have persisted for centuries. From Pattanam, broadly speaking, the bulk of Indo-Pacific beads seem to come from layers labelled as Early Historic or Medieval.

2.2 Other Bead Types from Pattanam The purpose of this essay, however, is to introduce other glass bead types from Pattanam. Altogether they constitute about a quarter of the beads recovered. Unfortunately, we still lack chemical data for Pattanam beads, so this essay will concentrate on describing the key morphological types within the bead assemblage, as a way to begin making material connections between Pattanam and other sites producing similar bead styles. Several recurring types are introduced here: gold glass beads; large faceted types that may imitate beryl gemstone beads; polychrome ‘zone’ beads that mimic banded agate beads; large drawn barrel beads and terracotta disc beads. Although a variety of Indian Ocean glass bead scholars have been working on glass beads from different sub-regions, a standardized classification system is still elusive.

348

S. A. Abraham

Glass beads can be subdivided according to any combination of attributes—manufacturing method, colour, opacity, size, shape and chemical composition. Because many of these traits range along a continuum rather than appearing as discrete categories, it can be challenging to settle on a single universal sorting system. Here, I begin by using bead terminology that commonly appears in research articles that discuss glass beads, and then, I present a preliminary overview of other beads whose classification is less clear.

2.2.1

Gold Glass Beads

Given their striking appearance, gold glass beads are among the best studied glass bead type in Indian Ocean and Old World assemblages (Fig. 2). Also referred to goldin-glass beads or gold foil beads, they are usually created by sandwiching a thin layer of gold foil in between two tubes of clear glass (Spaer, 1993: 10). The bead would be then heated and shaped. Origins of this bead are unclear. Boon assumed the technique spread from Egypt to become a Hellenistic and then a Roman tradition (Boon, 1977: 194). Spaer discusses possible Egyptian or Nubian origins (Spaer, 1993: 18), and others have posited an Indian origin (Singh, 1980). The Hellenistic site of Rhodes, with evidence for glass bead manufacture, has produced a large number of gold glass beads (Spaer, 1993: 18). Francis (2002: 93) argued that they reached India, China and Korea via land and maritime routes in Roman times, and in Islamic periods, they were exported to Srivijaya. Gold glass beads continued to be produced in medieval Europe (Spaer, 1993: 20) and so have a long history.

Fig. 2 Gold glass beads

A Review of Selected Glass Bead Types from the 2007–2009 …

349

According to Dikshit (1969: 56), the introduction of gold glass bead in India coincides with the adaptation of the technique in the early centuries of the Christian era. Dikshit also mentioned a vague reference in the Arthasastra suggesting that the technique may have been known on the subcontinent during the Mauryan age. Dikshit cites examples within that time range from sites like Nasik, Kondapur Ter, Chandravalli, Arikamedu and Kaundinyapur. The site of Bara in Pakistan also reported both gold foil and silver foil beads and was possibly a site of manufacture (Dussubieux & Gratuze, 2003). However, beads from Arikamedu in southern India and Mahasthan in Bangladesh appear to have Mediterranean, possibly Egyptian, origins—highlighting the challenges in ascertaining the chronology and distribution of this glass bead type (Gratuze et al., 2015; Tomber, 2013: 95–96). From the 2007, 2008 and a portion of the 2009 trenches, 28 gold glass beads or bead fragments were recorded. Although they were recovered throughout trench depths at Pattanam, such beads are most often associated with strata dating from third century BCE to third century CE. This timing overlaps with the occurrence of a variety of gold glass beads from both and late early phase periods at the Red Sea port Berenike in Egypt (Then-Obłuska, 2015). Gold foil beads dominate among metal in glass beads during Early Roman period in Egypt and the Meroitic period in Nubia (Then-Obłuska, 2016: 42). They are also widespread at Mantai in Sri Lanka, where they date from the early historic period possibly up to early medieval period or later (Francis, 2013: 356). Gold glass beads are also known from well-known sites in Southeast Asia such as Oc Eo in Vietnam and Kuala Selinsing in Malaysia, and even China (Francis, 2002: 92), and are most recently documented from sites in Thailand (Carter, 2013: 339–340).

2.2.2

‘False Beryl’ Beads (Large Hexagonal Faceted Paddle Drawn Beads)

Peter Francis (2002: 43) describes these beads as hexagonal beads made from reheated tubes, nearly always translucent blue or translucent green (Fig. 3). Based on their shape and colour, and comparing to samples from Arikamedu, Francis suggested that these beads may have been produced to replicate beryl beads and exported to Rome—hence the nickname ‘false beryl’. His idea is inspired by a comment from Pliny: ‘The Indians have found a way of counterfeiting various precious stones, and beryls in particular, by staining rock-crystal’. They are also found at Mantai in Sri Lanka (Francis, 2013: 353) and Oc Eo in Vietnam (Malleret, 1962, as cited in Francis, 2013: 353). Francis viewed them as imports from India, but other than their presence at Arikamedu, it is not clear on what basis Francis preferred Arikamedu as their manufacturing site. From Pattanam, 57 beads or bead fragments could be identified as false beryls. Most were partial, and a number were broken in half crosswise (Fig. 3)—a possible deliberate action rather than a by-product of formation processes, perhaps for display purposes.

350

S. A. Abraham

Fig. 3 False beryl beads

2.2.3

Red–Orange Disc Beads

These opaque red–orange disc-shaped beads were made by cutting very thin slices from solid tubes and then drilling each slice to create the perforation (Fig. 4). Examining similar bead types from Mantai, Francis argued that these disc beads are a subsidiary of the main Indo-Pacific bead industry—that these beads were made from thick Indo-Pacific glass tubes, which were cut into discs but hardly reheated. They have not been found at other Indo-Pacific bead-making sites (Francis, 2013: 364).

Fig. 4 Disc beads

A Review of Selected Glass Bead Types from the 2007–2009 …

351

Most commonly orange and red in colour, red–orange disc beads are found in India, Sri Lanka and Southeast Asia (e.g. Ban Chiang, Thailand). However, samples studied by Dussubieux and Gratuze (2013: 404–405) show them to be made of a mixed alkali glass, a mixed potash-soda flux, which seems to argue against them being an Indo-Pacific bead subsidiary. In Sri Lanka, they are also known from sites like Ridiyagama, Kelaniya, Tissamaharama and are reportedly the most numerous beads found in the Gedige area of Anuradhapura (Francis, 2002: 136–137). In India, they are reported from Dulhikotta, Kalahandi, Alagankulam, Kodumanal, and Karaikadu (Raman, 1991). Similar, though not identical, beads are reported by Carter from sites in Cambodia and Thailand (Carter, 2013: 350), but those appear to be wrapped rather than drawn and pierced like the Pattanam samples.

2.2.4

Collar Beads

Horace Beck (1926: 4) identified a collar as any attachment at the end of the perforation designed to strengthen the margin or reduce friction (Fig. 5). Francis (1986: 117; 2002: 42–43) said that collar beads have extra material surrounding both apertures. Francis also described collar beads as one of the most distinctive Indian developments during the early historic period (having been recovered from sites like Taxila, Vaisali, Nevasa), but that their numbers waned by the third century CE (Francis,

Fig. 5 Collar beads

352

S. A. Abraham

2013: 353). However, the category collar beads can be rather misleading, since it encompasses beads that otherwise vary a great deal in terms of style, colour and size. Gold glass beads from Pattanam for instance also appear with collars (see Fig. 2). Because it is so common in the glass bead literature, I include this category here, but it requires much more refinement to be useful.

2.2.5

Monochrome Wound or Pierced Beads

Monochrome beads that do not obviously identify as Indo-Pacific beads appear in the Pattanam assemblage in a range of sizes, shapes and colours. They include short bicone beads (Fig. 6); long faceted bicone beads (Fig. 7); large globular beads (Fig. 8) and tabular wound flattened oblate beads (Fig. 9).

Fig. 6 Short bicone beads

A Review of Selected Glass Bead Types from the 2007–2009 …

353

Fig. 7 Long faceted bicone beads

2.2.6

Polychrome Beads

This is also a broad category of glass beads, mostly (but not all) wound, and comprised of more than one colour glass, with a secondary colour or colours applied as a line around the bead, either perpendicular or parallel to the perforation. The shapes vary, and the secondary colour(s) are integrated in different ways. Polychrome beads appear regularly in the Pattanam assemblage; a few key sub-types are presented here. One of the most distinct is a finely made elongated banded bead that may have meant to imitated banded agate (Fig. 10). These may also parallel what used to be commonly called ‘zone beads’—beads that came in a variety of shapes but were distinguished by Beck (1926: 46) by a line decoration that divided the bead ‘around the perimeter concentric with the axis’. They may also parallel the beads described by Francis (2002: 93) as folded glass beads that imitated onyx—made from black or dark glass with one or two white stripes. A visually similar kind of bead was found at the site of Promtin Tai in Thailand (where Indo-Pacific beads and a gold glass bead were also found). This example also appears to imitate a striped agate, and its composition may link it back to south India, perhaps even Arikamedu (Carter, 2013: 339).

354

S. A. Abraham

Fig. 8 Large globular beads

Other examples from Pattanam include beads made in different ways (drawn, wound, folded, rod pierced) that include ‘trail’ decorations (Fig. 11) which appear similar to certain glass beads from Berenike (Then-Obłuska, 2015, 2017) and Mantai (Francis, 2013: 358). These include small drawn black beads with longitudinal white stripes, which have parallels with beads in late phase Berenike assemblages (ThenObłuska, 2015: 753, 755 (Figs. 4.31 and 4.32)), and a few Egyptian examples with white and red stripes (Then-Obłuska, 2017: 724, 725 (Fig. 1.14)). Another Pattanam polychrome style is the folded striped bead with a deep irregular white trail (Fig. 12), which is also visually akin to late phase Berenike examples [Then-Obłuska, 2015: 760, 759 (Figs. 5.24 and 5.28), 760] which had red or blue matrices. The Mantai report includes similar blue glass beads with a deep white zone (Francis, 2013: 353). The Pattanam beads are clearly folded and then pierced, as can be seen in the cross section (Fig. 12). Francis (2013: 353) cites parallels with a dark blue matrix from the Gangetic Valley and at Satavahana sites on the Deccan Plateau.

A Review of Selected Glass Bead Types from the 2007–2009 …

Fig. 9 Wound flattened oblate beads

Fig. 10 Elongated banded beads

355

356

S. A. Abraham

Fig. 11 Drawn trail decorated beads

3 Conclusion This is a preliminary but exclusive account of selected recurring non-Indo-Pacific beads from Pattanam. Pending the availability of chemical compositional data for the Pattanam assemblage, it is hoped that this macroscopic overview will facilitate discussions about Pattanam’s links with other sites from South Asia, the Indian Ocean littoral and beyond.

A Review of Selected Glass Bead Types from the 2007–2009 …

357

Fig. 12 Folded Stripe Pierced Beads

References Abraham, S. A., & Christie, H. (2010). The Indian Ocean and Indo-Pacific Bead: Mapping a Key Artifact Category from the Pattanam Excavations. Paper presented at the 20th Conference of European Association of South Asian Archaeology and Art, 4th–9th July. Vienna, Austria. Abraham, S. A., Cherian, P. J., & Christie, H. (2011). Pattanam/Muziris: The glass bead corpus from an Indian Ocean Port Site on the Malabar Coast of Kerala, India. Paper presented at the Society for American Archaeology 76th Annual Meeting, March 30th–April 3rd. Sacramento, California. Abraham, S. A. (2016). Glass beads and glass production in early South India: Contextualizing Indo-Pacific bead manufacture. Archaeological Research in Asia, 6, 4–15.

358

S. A. Abraham

Beck, H. C. (1928). Archaeologia 77, 1–76. Boon, G. C. (1977). Gold-in-Glass Beads from the Ancient World. Britannia, 8, 193–207. Carter, A. K. (2013). Trade, Exchange, and Sociopolitical Development in Iron Age (500 BC–AD 500) Mainland Southeast Asia: An Examination of Stone and Glass Beads from Cambodia and Thailand. PhD Thesis. Madison: University of Wisconsin-Madison. Carter, A. K. (2016). The production and exchange of glass and stone and beads in Southeast Asia from 500 BCE to the early second millennium CE: An assessment of the work of Peter Francis in light of research. Archaeological Research in Asia, 6, 16–29. Cherian, P. J. (2012). Pattanam archaeological site: Evidence of maritime exchanges. Tamil Civilisation, 12(1–2), 22–31. Cherian, P. J., Tomber, R., Abraham, S. A., Giumlia-Mair, A., Kelly, G. O., & Nayar, P. (2016). Items of personal adornment from Pattanam, Kerala. Journal of Indian Ocean Archaeology, 12, 35–63. Dikshit, M. G. (1969). History of Indian Glass. University of Bombay. Dussubieux, L., & Gratuze, B. (2013). Glass in South Asia. In K. Janssens (Ed.), Modern methods for analysing archaeological and historical glass (Vol. I, pp. 400–413). Wiley. Dussubieux, L., & Gratuze, B. (2003). Nature and Origine des Objets en Verre Retrouvés à Begram (Afghanistan) et à Bara (Pakistan), In O. Bopearachchi, C. Landes, & C. Sachs (Eds.), De l’Indus à l’Oxus: Archéologie de l’Asie Central, pp. 315–351. Lattes: Musée de Lattes. Dussubieux, L., Gratuze, B., & Blet-Lemarquand, M. (2010). Mineral Soda alumina glass: Occurrence and meaning. Journal of Archaeological Science, 37(7), 1645–1655. Francis, P. (1986). Collar beads: A new typology and a new perspective on ancient Indian beadmaking. Bulletin of the Deccan College Post-Graduate & Research Institute, 45, 117–121. Francis, P. (2002). Asia’s Maritime Bead Trade: 300 BC to the Present. University of Hawaii Press. Francis, P. (2013). The Beads. In J. Carswell, S. Deraniyagala, & A. Graham (Eds.), Mantai: City by the Sea (pp. 349–373). Linden Soft Verlag. Gratuze, B., Lankton, J. W., Alam, Md S., Boussac, M.-F. (2015). Glass beads from Mahasthan— Evidence for production and exchange. In J.-F. Salles (Ed.), Mahastan II Fouilles du Rempart Est, Etudes archéologiques (pp 357–374), Brepols, Turnhout, Belgique Kanungo, A. K. (2016). Mapping the Indo-Pacific Beads Vis-à-vis Papanaidupet. Aryan Books International / International Commission on Glass. Katsuhiko, O., & Gupta, S. (2000). The far east, southeast and South Asia: Indo-Pacific beads from Yayoi Tombs as indicators of early maritime exchange. South Asian Studies, 16(1), 73–88. Lankton, J. W., & Dussubieux, L. (2013). Early Glass in Southeast Asia. In K. Janssens (Ed.), Modern methods for analysing archaeological and historical glass (Vol. I, pp. 415–443). Wiley. Lischi, S. (2018). Macroscopic analysis of the bead assemblage from the South Arabian Port of Sumhuran, Oman (Seasons 2000–2013). Arabian Archaeology and Epigraphy, 29, 65–92. Malleret, L. (1962). L’Archéologie du Delta du Mekong. Ecole Française d’Extrême-Orient. Pion, C., & Gratuze, B. (2016). Indo-Pacific glass beads from the Indian subcontinent in early Merovingian Graves (5th –6th centuries AD). Archaeological Research in Asia, 6, 51–64. Raman, K. V. (1991). Further evidence of Roman trade from coastal sites in Tamil Nadu. In V. Begley & R. Daniel (Eds.), Rome and India: The ancient Sea trade (pp. 125–133). University of Wisconsin Press. Selvakumar, V., Gopi, P. K., & Shajan, K. P. (2005). Trial Excavations at Pattanam: A preliminary report. Journal of the Centre of Heritage Studies, 2, 57–66. Shajan, K. P., Selvakumar, V., & Tomber, R. (2005). Was Pattanam ancient Muziris? Man and Environment, 30(2), 66–73. Shajan, K. P., Tomber, R., Selvakumar, V., & Cherian, P. J. (2004). Locating the ancient port of Muziris: Fresh findings from Pattanam. Journal of Roman Archaeology, 17, 312–320. Sidebotham, S. (2011). Berenike and the ancient maritime spice route. University of California Press. Singh, R. N. (1980). The antiquity of gold-glasses in India. Puratattva, 12, 157–159.

A Review of Selected Glass Bead Types from the 2007–2009 …

359

Spaer, M. (1993). Gold-glass beads: A review of the evidence. Beads: A Journal of the Society of Bead Researchers, 5, 9–25. Then-Obłuska, J. (2015). Cross-cultural encounters at the Red Sea Port of Berenike, Egypt: Preliminary assessment (Seasons 2009–2012). Polish Archaeology in the Mediterranean, 24(1), 735–777. Then-Obłuska, J. (2016). Beads and pendants from the Tumuli cemeteries at Wadi Qitna and Kalabsha-South, Nubia. Beads: A Journal of the Society of Bead Researchers, 28, 38–49. Then-Obłuska, J. (2017). Between the Nile and the Ocean: The bead assemblage from Shenshef in the Eastern Desert (4th –6th centuries AD). Polish Archaeology in the Mediterranean, 26(1), 719–747. Tomber, R. (2013). Pots, coins, and trinkets in Rome’s trade with the east. Rome beyond Its Frontiers: Journal of Roman Archaeology (supplementary Series), 95, 87–104. Wood, M. (2011). A glass bead sequence for Southern Africa from the 8th to the 16th Century AD. Journal of African Archaeology, 9(1), 67–84. Wood, M. (2016). Eastern Africa and the Indian Ocean world in the first millennium CE: The glass bead evidence. In G. Campbell (Ed.), Early exchange between Africa and the wider Indian Ocean (pp. 173–194). Palgrave Macmillan.

Glass Bangles in South Asia: Production, Variability and Historicity Mudit Trivedi

Abstract The glass bangle in South Asia is a widely distributed artefact with recognized spatio-temporal variation. Often considered only as an addendum to widely traded glass beads, it remains an understudied class of archaeologically recovered glass objects. This paper argues there is much to be learnt from the historicity of glass bangles beyond the present debate that has dwelled on their ‘antiquity’. Towards this end, the paper provides a review of the techniques by which bangles were produced in South Asia. It argues that markedly different sociologies of artisan-group organization, scale of production, patron-craftsmen relations, and workshop location including itinerant production, distinguish the two major techniques, namely: the ‘Khalbut’ and the ‘two-mandrel’ methods. Building on this discussion, the paper provides a short archaeological history of the glass bangle up to the early medieval Period. It highlights the historically specific nature of marked patterns of regional, typological and chemical variation each indicating shifts in glass bangle production, use and valuation.

1 Introduction The glass bangle is arguably the least studied class of glass objects in South Asia. We understand little of how glass bangles change typologically, technologically or in terms of their chemistry. We struggle to discern when they are found whether they are early historic, early medieval or even later. Beyond elementary questions of chronology and typology, we do not understand how, when and under what cultural conditions, bangles, long valued and produced in diverse media-shell, faience, stoneware, lac, a range of metals and alloys and even precious stones came to be principally an item conceived, produced and used in glass. As Kenoyer (2003: 52) has noted ‘The use of bangles or bracelets as a form of ornamentation is common throughout the world, but in South Asia bangles appear to have taken a specific role as a symbol of socioritual status and ethnic identity’. M. Trivedi (B) Department of Anthropology, Stanford University, Stanford, USA e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_15

361

362

M. Trivedi

Indeed, bangles do not only serve as ornaments they simultaneously classify those who wear them. They mark persons in several ways signalling ethnicity, un/married or widowed status and the divides of ritual impurity and rites of passage (Gaborieau, 1977). Beyond these subtle symbolisms, glass bangles bear great potential for further inquiry into the glass cultures of South Asia. They hold the potential to reveal unique aspects of production, exchange and use—in patterns that are enduringly different to those of beads, vessels and other glass objects. Understanding the specificity of glass bangle producing communities and the networks that connected them to their diverse users are significant open questions in the archaeological history of Indian glass. This paper first provides an account of the several techniques by which glass bangles were historically made in South Asia. It draws attention to how these varied techniques rested upon very different sociologies of craft expertise, residence and mobility, exchange and patronage relations. It draws out the importance of distinguishing between them for archaeology. The paper suggests on this basis that a focus away from questions of antiquity to those of the marked historicity of the glass bangle is needed. The paper then provides a short archaeological history of the glass bangles up to the end of the early medieval period. It demonstrates how in every major period of South Asian history the glass bangle has much to teach us, if only we do not assimilate it to the study of glass beads. In conclusion, the paper lays out questions and strategies for much needed future research.

2 Attending to Process: Different Paths to the Bangle Form in South Asia The simplest way of understanding bangle production sequences is through the problem of enlargement. A small gob of molten glass, once formed into an initial ‘bead’ then has to be modified into a bangle of the desired diameter, and in doing so, the glass mass thins as the perforation widens. The specific actions and gestures used by the craftsperson, the sequence of steps and the tools used all shape and limit the possibilities for further shaping, decorating and distinguishing the base bangle into the desired ornament. As a result, specific types of bangles need certain key steps that allow modifications of shape and decoration, and these are more or less feasible in different methods of enlargement. This section outlines how not one, but at least three techniques of making a bangle are evidenced in South Asia. Each technique has its own histories in the subcontinent and beyond, each indexes a different craft-history and shifting communities of practice, and each related to very different sociological networks that underlay the bangle workshop and linked craftsperson to user.

Glass Bangles in South Asia: Production, Variability …

363

2.1 The Khalbut: Enlargement on a Cone/Conical Mandrel The most widely reported method of manufacturing bangles and the one assumed as default in most colonial and archaeological literature involves the use of a specifically fashioned clay cone attached to a mandrel (Fig. 1). The initial pull of glass is shaped into a bead, usually then marvered1 and pre-formed, and the bead then minimally enlarged and made free from its mandrel. As soon as the opening has enlarged to a centimetre, it is transferred to the tip of a cone attached to another mandrel. A thin pontil is inserted between the red-hot but now cooling glass bead and the conical mandrel rotated—the centrifugal torque exerted by the rotation expands the bead into a bracelet. The cone, which notably preserves across South Asia a Persian name— Khalbut,2 bears on it several grooves that allow the process to be stopped at different desired diameters, i.e. bangle sizes. The entire process, as noted by several observers, is designedly efficient and a craftsman takes less than a minute from start to finish for a simple bangle (Francis, 1982, 2013; Kanungo, 2004, forthcoming). If the craftsman plans to decorate the base bangle so produced, then the process has to be modified and several steps are added. As this militates against the rationale of the Khalbut process that is geared towards increased efficiency and mass production, the seamless and regularly round bangle so produced by the Khalbut technique is usually left unornamented. If decorations are added to the bead before, it is transferred to the cone and they bear some trace of the rapid rotation they have undergone. Alternatively, bangles produced by the Khalbut may be selectively deformed into innovative shapes, thus decorating the bangle without further addition of coloured glass but by using plastic means of modification. The colonial ethnography of crafts in India was conditioned by state attention towards the documentation of those industries that were considered of economic potential. In the context of surveys of ‘pottery and glassware’, these late nineteenthcentury texts thus provide accounts of proto-industrial bangle workshops where around a large kiln several men each produced up to a thousand bangles a day using the Khalbut method (Dobbs, 1895; Hallifax, 1892). Simultaneously, the policies of the colonial state, especially of allowing the mass import of Japanese and Czechoslovakian bangles, eviscerated the market-share of Indian bangle-makers. Thus, increasingly large multiple Khalbut workshops were not timeless forms of bangle production but responses to the early twentieth-century economy and common until a new Industrial method to make the process even more efficient was devised.3 1

The term marvering refers to the use of a flat clean surface or tool to roll warm glass on during the manufacturing process. The process allows for the craftsperson to shape the glass. Once additional coloured or prepared glasses are added to the pre-form being worked, it can be marvered again and this simple combination of steps is used to create a range of effects. 2 As an exception Dikshit (1969) records that at Pemgiri the bangle-making community he met with referred to this tool as the sundari. 3 This involving pulling not one bead, but ‘drawing’ an entire spiral from the melt onto a rotating cylinder/belan. This was achieved at first, it seems beginning in the 1920s, by using a team of four workers, and then later by the 1950s with the use of a mechanized rotating cylinder. The spiral is then cut, and the resulting open rings are individually and labouriously joined on a flame/gas

364

Fig. 1 Enlargement of a bangle using a Khalbut

M. Trivedi

Glass Bangles in South Asia: Production, Variability …

365

Even keen twentieth-century observers of Indian crafts, such as Peter Francis Jr., assumed that the principal historical technique of making bangles in South Asia had relied on the Khalbut. This is unfortunate and ironic, as the production model assumed in archaeological interpretation is but the most recent of workshops and methods, and one devised to increase productivity but reduce the complex possibilities of the object. By all indications, the Khalbut is a medieval era technological development. The history of its introduction, its uptake and use in parallel with other techniques remains uninvestigated and no Khalbut has yet been recovered in any South Asian archaeological context. We must question the abiding assumptions that the Khalbut method was the principal technique of production across periods in South Asia. In fact, there is ample evidence to the contrary. In addition, we must question the resultant expectations we have of the archaeological record—for other forms of bangle production, as discussed below are likely to have left very different and more ephemeral traces as compared to our expectations of the large multi-Khalbut workshop.

2.2 Balanced on Two-Mandrels: The Example of the Nepali Churihars Marc Gaborieau, French ethnographer of Indian Islam, encountered in the 1960s a caste of churihars (bangle-makers) who served many communities in jajmani (patronclient) relationships in the hills of western Nepal. While his central project was a detailed ethnography of the social status of this caste (Gaborieau, 1993), archaeologists are indebted to Gaborieau as he also provided a detailed account of the chaine operatoire of the distinct two-mandrel technique used by these churihars (Gaborieau, 1977). Using a much smaller and readily devised two-chamber kiln, the churihars drew a small ‘bead’, then shaped and marvered it on a stone, shaping the bead into the desired profile for the bangle. While the bead was still on the mandrel, the churihar reintroduced it to heat at the kiln’s mouth and if desired embellished its arris4 with the addition of canes and prunts of variably coloured glass. Notably, this combination of actions—the marvering, addition and layering of coloured glasses, shaping and reheating—can be iterated and repeated, and they allow the craftsman to variously combine simple techniques into a wide range of different end products (Fig. 2). Once the churihar is satisfied with such additions, the mandrel is struck repeatedly to detach the initial bead and begin the enlargement. The slightly enlarged ring is lamp (often by children and women). While the labour force required for the stage of jud.ai is thus enlarged, the total volume produced offsets this and keeps wages exploitatively low. Up to 50,000 bangles can be produced at each factory a day, and the endless capacity for the exploitation of labor has ensured that these bangles are the definitive twentieth-century type. 4 The term arris is typically defined as ‘a sharp edge formed by the meeting of two flat or curved surfaces’. It is introduced here in this technical sense to draw attention to the dexterity and skill which making bangles of particular cross sections were demanded. A common example in the South Asian record, are ‘triangular’ bangles defined by such sharp arris (see Fig. 4, e.g. from KSK).

366

M. Trivedi

Fig. 2 Enlargement of the bangle using the ‘two-mandrel’ technique as observed by Gaborieau at Kurahan (1. the churihar draws an initial gob or wind, 2. it is first shaped on a flat spatula-like tool, 3. then it is marvered and shaped on the marvering stone. Previously prepared trails or twisted canes are melted at the kiln mouth, where they are added as desired on to the arris of the bangle, 4. the enlargement then takes place at the kiln’s mouth in a zone of moderate heat, with the bangle balanced on two mandrels, each dexterously rotated in ways so as to utilize the plasticity of the cooling glass to attain the desired aperture and form 5 and 6) (Redrawn after Gaborieau, 1977)

Glass Bangles in South Asia: Production, Variability …

367

transferred to a second mandrel, which is thickened for a small portion of its length with clay. Through a dexterous series of rotations of these mandrels in different directions, while holding the bangle at the kiln mouth to keep it at the desired temperature and plasticity, the bangle is enlarged. This process affects enlargement without a cone, and thus relies on the skill of the craftsman to arrive at a circular form. While this is regularly achieved, the process also leads regularly to kinds of deformities that may be taken as traces of the technique. This ‘two-mandrel technique’, to give it a name, allows for the most extensive modification of the three-dimensional form, or profile of the bangle, especially as compared to the Khalbut technique. It also allows for a range of decorations to be added, which in the Khalbut technique would be deformed under the centrifugal force of the rotating cone. Beyond these design possibilities, as Gaborieau’s detailed ethnoarchaeology makes clear, the two-mandrel technique is associated with a much smaller kiln, one that could be readily constructed. In Gaborieau’s example, the technique is clearly linked to a hereditary community of caste specialists who were tied in jajmani relations to patrons in several villages. The social relations of the Nepali churihar clearly associate this technique with itinerant production. This picture of bangle production radically affects our expectations of the archaeological find of a glass bangle workshop, as the evidence that Gaborieau presents is starkly different from the large Khalbut workshops (Table 1). The archaeological trace of a small kiln, likely constructed on the outskirts of a village and abandoned after only a few days of use would after varied taphonomic processes leave more or a smear of debris than the traces of a workshop in accordance with our preconceived expectations. Finds of misshapen bangles, melt bowls and small furnace fragments Table 1 A comparison of the Khalbut and two-mandrel techniques Khalbut Kiln size (l × w × h)

Two Mandrel

95 cm × 35 cm × 55 cm (kiln 50 cm × 30 cm × 20 cm only)

Number of engaged craftsmen Usually 6–20

As low as 2

Number produced daily per craftsman

600–1200

50

Time per bangle

38–45 s

2–10 min

Geared towards

Pre-industrial mass production; Handwork addition reduces output

Designedly geared towards highly embellished

Sociology of production and exchange

Intergenerational and hereditary family workshops which employ many beyond kin and produce for market

Kin group, sometimes solitary craftsman and partner who hawks, mediates with female clients who are often jajman (patron). Also known to hawk own wares at local markets and fairs

368

M. Trivedi

in association, even in surface finds, are not uncommon but are regularly taken to be waste associated with glazed ware production (for one example from a surface survey at Kausambi see Trivedi, 2008: 454–459). In this technique, the scale of production is notably different, as is the care given to each artefact as the time invested to richly ornament each piece is greater. The labour group is smaller, and the melt usually worked by a solitary patriarch and staffed by his immediate kin group. Gaborieau noted production averaged at 40–50 pieces a day. The two principal production techniques described here are strikingly different in the socio-economic worlds that stand behind them and in the traces we might expect of them in the archaeological record. Yet, our extant knowledge and modes of studying glass bangles in South Asia leave us in a position where we can only assert the most limited understanding of these distinct histories.

2.3 Round and Seamed: Joined Cane Bangles Both the production techniques described above result in seamless bangles, as enlarged from an initial bead. On account of their handling, they also bear bases that are at least partly flat and thus bangles that are round in cross section cannot be produced by these methods. Round (in profile) glass bangles of a number of distinctive types have been produced, globally, since at least Roman times and are widely distributed. Their production technique involves the use of a pre-formed drawn glass cane that is then reheated to malleable softness and shaped over a tool with the ends joined together and seamed. Archaeologically recovered bangles often display this seam. This technique allows for a range of distinctive decorative techniques that prominently include the pre-formed cane being twisted on itself thus producing a twisted bangle, common in Roman assemblages. Different coloured trails can be applied to the base bangle, which then twisted produced a series of interlinear twisted coloured effects. Bangles of this type appear to have been distinctively produced in Palestine and the Levant, where they were produced in a range of greens and blues prior to the rise of Islam and subsequent to that in a range of coloured canes as added to a mostly opaque base bangle (usually black or green) (Spaer, 1988). It has been regularly assumed that such bangles were not made in South Asia, yet they do appear in different periods of the archaeological record in marked regional trends. The evidence from early historic Indo-Greek Barikot (Bettineschi, 2019) suggests that such twisted bangles were a prominent and valued part of the monochrome assemblages of the era. Whether such bangles were made in ancient South Asia remains a matter open to investigation (Fig. 3). Since at least the nineteenth century a particular kind of bangle, preferentially made of a transparent glass within which a series of coloured canes was carefully layered within the bangle has been produced in round sections in north India, Maharashtra and possibly the north-west Provinces in Pakistan. Their production involved a careful process of preparing the base melt and separately the coloured canes. The latter were carefully layered, impressed and marvered into the melt, which was then

Glass Bangles in South Asia: Production, Variability …

369

Fig. 3 Process for making bangles with ‘inside trails’ requires the glass melt to be prepared and then trails to be pressed into this. After this long canes are skilfully drawn in the manner shown. Redrawn based on documentation done as part of the Handicrafts series of the National Census in 1960

shaped into a cone. From this cone, long canes of up to 100 feet were drawn. The canes were maintained at specific and regular thicknesses and after cooling were broken into a pieces corresponding to the desired size of bangles. These were then reheated, joined over a Khalbut or another tool and then lain to cool. (Maharashtra Census Office, 1968) The immobilization at the time of joining the seams and the tools used to join the ends and the cooling of the bangle introduce small asymmetries that may help discern a range of distinct techniques used to produce such bangles. Depending on the intended effect and the skill of the craftsmen, some of these bangles are more oblong than the nearly symmetrically round bangles that were produced in Palestine and the Levant.

3 From Antiquity to History: Interpreting the Glass Bangle The archaeological study of glass bangles in South Asia has hitherto been dominated by a question of its ‘antiquity’. This is largely due to a foundational debate that unfortunately has continued to frame and limit their study. In this debate H.D. Sankalia made two distinct statements, an article on ‘The antiquity of glass bangles (1947)’ and

370

M. Trivedi

then laudably revising his opinion in light of new accumulating evidence in subsequent excavation reports. At the Kolhapur excavations, faced with major evidence of a medieval bangle workshop in his trenches Sankalia initially argued ‘glass bangles probably first came to [be] made in the Deccan during the Bahmani period’. This first statement is driven by what appears to him as a puzzle: Sankalia was troubled by the question of how did so recent an artefact, and one which came from a ‘craft [that] was mostly in the hand of Muslim families who practice it as a cottage industry’ had ‘seem[ed] to have acquired in Hindu social life such a high position during the Muslim and post-Muslim period’ (1947: 253). Subsequently, when Sankalia qualified his statement he asserted that glass bangles did exist in the pre-medieval archaeological record, and he did so by arguing that these earlier bangles were monochrome and undecorated (1952, 1960). Since this time, a principal thrust of studies commenting on glass bangles has been to seek to establish their pre-medieval antiquity and to do so for all categories of bangles—monochrome, bichrome and polychrome (see Lal 1952a, 1952b; Lal 1954–55: 90–91; Thapar, 1957: 111–112, 1967: 116–117; Sahi, 1988; Mishra et al., 2010). The number of colours a bangle bears is a crude measure of the additive complexity of how it has been decorated but it is a woefully inadequate description of the bangle itself. Tracking variation at the level of number of colours cannot address how the bangle was made, specifically what kind of plastic or additive decoration was effected nor relate these broad features to specific types produced by craftsmen using one or the other method of production as discussed above. Sankalia’s own reports had been invested in detailed studies of the bangles and described variability for several traits—the number of colours, cross-sectional shapes and descriptive accounts (and illustrations) of different decorations was provided. Yet owing to the debate, variation across these traits has rarely been correlated towards defining specific types of regional and historical salience. Without such a synthetic typology, a patently multi-variate problem is provided with partial univariate descriptions that occlude historical patterns of change. Attention to the production techniques of the bangle opens up the artefact to questions beyond its antiquity. Instead of focusing on the date of the first bangles, we can begin to ask a range of questions: what types they are, how they change between periods and whether there is any evidence on the artefacts themselves to establish what technique(s) they were made by. Through such typo-technological archaeologies of the glass bangle, we can begin to devise better typologies and identify shifts in production techniques, decorative repertoires of craftsmen and shifts in the cultural preferences of users can be identified. As the next section demonstrates, we can use the wealth of accumulated knowledge in South Asian archaeology to begin to build towards a diachronic and synchronic examination of how bangles were made, used and traded in different periods and regions and what we can learn from this about glass technology, social organization of the craft and the shifting significance of these objects in social life. Much can be gained from the published data when we let go of a quest for antiquity and pursue the variability of the glass bangle beyond using the number of colours as the prime variable. Even more can be gained from a rigorous examination of glass

Glass Bangles in South Asia: Production, Variability …

371

bangle assemblages themselves. In my research at the site of Indor (Dist. Alwar) Rajasthan, through a detailed multi-variate analysis of an assemblage of over 3900 glass bangle fragments dating between the fourteenth and twentieth centuries CE, it has been possible to pursue some of these possibilities. At Indor, a robust typology for the medieval glass bangle has been developed. By correlating the cross sections of the glass bangles with the kinds of tool-traces (ventral ferric sputter, bulges, crimps, tugs and trails marks) left on the bangle with the kinds of decoration made on them, it has been possible to define clear types of bangles as well as series in which they group within periods and between them. These series allow for the study of how distinct types of bangles were conceptualized, executed and extensively used at this medieval city. It is possible to track variation in excavated assemblages at a centennial scale. At Indor, we recovered what appears to be traces of itinerant production using small kilns of the kind Gaborieau described and distinctive wastes of the production process. The Indor study has also allowed for rigorous comparison of South Asian glass bangles with the tenth through twentieth-century CE phenomena of the ‘Islamic glass bangle’ as known from archaeological assemblages drawn from multiple contexts ranging from North Africa, Palestine and the Levant to Iran and Central Asia (Spaer, 1992). The glass bangle assemblages at Indor, a city uniquely associated with the conversion of a local lineage to Islam, thus open up the practices and uses that underlay cultures of the glass bangle to an archaeological study far more complex than that framed as a matter of medieval or pre-medieval, Muslim or Hindu (Trivedi, 2020; Trivedi & RDAM, 2016). The discussion below argues that alertness to typological distinctiveness enables us to historicize the bangle in every period of South Asian history. The subsequent sections provide examples for each major period up to 1250 CE. It makes clear that once we stop subsuming the bangle within the study of glass beads, the specificity of the bangle and its worlds emerges. Alongside independent communities of practice and expertise, unexpected patterns of cultural value and geographies of exchange come into view.5

4 A Short Archaeological History of the Glass Bangle in South Asia 4.1 The Earliest Finds: New Evidence for Production and Exchange Networks The earliest glass bangles in India appear to be those reported as rare finds in the context of the Painted Grey ware culture of Northern India from sites like Ropar, 5

This section is summarized from a longer reconsideration of the archaeological history of the glass bangle under preparation. Trivedi 2020 provides a wider historiographic argument and a detailed account of medieval bangles not covered here.

372

M. Trivedi

Hastinapur and Alamgirpur. Within the debated chronology of this period (c. 1100– 600 BCE), the specific moment(s) of their appearance has not been established, nor do we presently have a secure understanding of the authorship, typology or chemistry of these early glass objects. These bangles appear to be fashioned out of black, green and blue glass and were shaped into small plano-convex and a wide triangular forms. They occur alongside equally scarce bangles of agate and seem to be one element among inadequately understood early north Indian glass assemblages. The glass bangles that accompany the subsequent period of the ‘second urbanization’ and the associated expansion of material culture of the Northern Black Polished Ware period (conventionally 600 BCE onwards) appear to show a broadly similar rare distribution of bangles, largely of green and blue glass, undecorated by any additive techniques and made in a limited range of cross sections. The lack of volumetric quantifications of small finds precludes statistical comparisons across periods and while glass bangles become more widely distributed it is relatively unclear whether they are common as an ornament. (cf. Mishra et al., 2010). The most remarkable advance in the archaeology of early glass bangles comes from the excavation of Khao Sam Kaeo in the upper Thai peninsula (hereafter KSK). At KSK ‘a mature industry’ manufacturing distinctive glass bangles, displaying a ‘high level of organization’ has been excavated and dated to the early period of the site, between the fourth and second centuries BCE. The remarkable bangles at KSK are fashioned principally of a distinctive translucent green glass. Furthermore, this glass has been established as of north Indian composition (specifically of the mineral soda alumina glass group 3, hereafter m-Na-Al glass (Dussubieux et al., 2010)). In addition to the finds of over 300 bracelets, the excavations yielded definitive hotworked waste in this distinctive glass. Importantly, clear trends suggest that glass bangles and bead-making activities at the site were spatially and chronologically distinct. The monochrome green bangles from the site come in a range of distinctive shapes (Fig. 4). Dussubieux has described their relative abundance and noted these bangles, including the unique ‘House type’ or pentagonal profile bangle, are part of a South China Sea trade network with finds at sites in Cambodia and Vietnam (Dussubieux & Bellina, 2017: 565). A hitherto unnoticed connection is the occurrence as rare finds of the distinctive pentagonal type in translucent green glass in the early assemblages of Prakash and Kaundinyapur in the Deccan (Thapar, 1964–65: 116). That the evidence of the earliest glass bangle-manufacturing workshop known anywhere in South and Southeast Asia comes from KSK poses a range of new questions. That the KSK bangles are fashioned from a glass likely made in North India, at and in the region around Kopia (Kanungo, 2013; Kanungo & Brill, 2009), implicates these objects in complex networks where the movement of raw glass, expertise and personnel are all involved. The connections suggested by the distribution of m-Na-Al-3 glass point to how the origins of glass bangle-making expertise is not well understood. That problem requires further typological and chemical studies of bangle assemblages of this period across different regions of South Asia. Further detailed studies of the KSK glass bangles in terms of their manufacturing technique, tool traces and distinctive waste products will no doubt help reveal further aspects of this early complex.

Glass Bangles in South Asia: Production, Variability …

373

Fig. 4 Bangle profile shapes as classified at KSK and Khao Sek (Redrawn after Dussubieux & Bellina, 2018, Fig. 3)

Considering the earliest glass bangles across South Asia requires us to broaden our geographical horizons, decentre the question of local authorship and question the insistence on ‘indigenous origins’ of the artefact when the earliest networks point to complex networks of expertise and exchange across North India, the Deccan, the Thai peninsula and beyond. The earliest glass bangle evidence across South Asia thus suggests its status as a highly valued object, one that was traded long-distances and whose value arose in a cultural geography and exchange networks between early city states like KSK. It is likely that glass bangles were items originating from a handful of locations and entering a number of different exchange networks in the 1st millennium BCE. Wider recognition of their rarity, the few and distinctive types marked in both the kinds of glass and the shaped artefact, will allow for studies of the glass bangle trade and holds considerable promise for the study of this period.

4.2 Regional Trends in the Early Historic: The North-West, Deccan and Beyond The North-west part of the subcontinent shows clear evidence for a different kind of emergence of the glass bangle. Here bangles made in a range of distinctive blue glasses are subject to a distinct series of plastic modifications. These kinds, which generally include a prominent central ridge on an oblong bangle, are known from Taxila and other sites. The detailed study of bangles at Barikot by Bettineschi et al. (2019) clarifies this picture considerably. They track the evolution of bangles from the Acheamenid period to the Kusana, with a distinctive peak in quantity and new types—including a round twisted type—in the Saka Parthian period (Fig. 5).

374

M. Trivedi

Fig. 5 Principal types of the north-west early historic. Top: Central ridge in a deep rich blue known from Taxila. Central pointed arris in blue and black is also known from the European context as is the twisted bangle, known in a dark opaque black and a blue (Twisted redrawn after Karwowski, 2010)

Across north India and Bengal, a group of black bangles with a triangular shape and asymmetrical ridges is perhaps the most widespread type and clearest marker of the period. Both of these distinct types bear broad similarities in form to Roman and post-Roman regionalized styles of bangles, as known from Eastern Europe and the Levant (Karwowski, 2010; Spaer, 1988). In those regions, the directed production of black glass (Fe2 O3 > 5%) allowed for bangles that seem to imitate prior types made from jet. While several types of black bangles in the subcontinent continue into later periods, some are specific to the early historic and compare well with some of these decorative techniques. Detailed studies of early historic bangle assemblages, with close attention to the kinds of tooling traces, production marks and their distribution within South Asia are needed to understand their typology, production and internal differentiation. The evidence in the Deccan and South India for these periods is diverse and challenging. Black and green bangles have been described for the early periods of Maski, Prakash and other key sites (Thapar, 1957, 1967), but the continuity of megalithic and early historic assemblages, the lack of single-locus excavation methods and radiocarbon chronologies complicate any discrete chronological attribution to the bangles.

Glass Bangles in South Asia: Production, Variability …

375

What we can establish across sites like Nevasa, Kondapur and Brahmagiri is that the early material is rare in occurrence, and chemically varied in nature. Revisiting the published Nevasa analysis, the glass bangles analysed from the ‘Indo-Roman’ assemblage at that site include both South Asian mineral soda alumina glasses but also plant ash glasses of a type known from diverse Indian Ocean networks (v-Na-Ca glasses) (Varshneya et al., 1988). The evidence from early historic Mantai suggested to Francis a picture of bangles being available in a range of glasses (on the basis of colour and visual examination) that were distinct from the locally produced beads (2013: 382). The combined picture, as can be best ascertained, suggests a limited occurrence of bangles overall and that likely many of these were being traded over long-distances. We need more chemical analyses for securely dated finds from this period across several types to begin to understand whether it is exchange or local production that authors these bangles in this period. Notably, none of the bangles tested so far from South Asia plot to known Roman or Levantine Black bangle compositions. The clearest trend in this period is the marked distinction between bangles and the explosive growth in glass beads, especially the hundreds of thousands of ‘IndoPacific’ types that are recovered regularly from sites in South India and Sri Lanka. At Pattanam for example, only 51 glass bangles were recovered in total as compared to the 98,000 plus glass beads. Of these 51, the early Historic ones are black and are succeeded by a medieval deposit of different colours and decorations. That the glass bangles appear to be items of local use, unlike the bulk of beads meant for export (Cherian et al., 2016), suggests that we urgently need to disentangle these two trajectories in the use of glass adornments and study them both as closely as possible.

4.3 Early Medieval Bangles The well-established problems of the archaeology of the early medieval period preclude anything more than a scattered sense of trends (Hawkes, 2014a, 2014b). Nonetheless, it appears that we are at the cusp of delineating types and series within this period. At sites like Prakash, Mantai and Vadnagar, it appears that a complex and intricate industry of shell bangles of ever increasing complexity of elaborately carved types proliferated in the early medieval period (Ambekar et al., 2019; Francis, 2013; Thapar, 1967). It is within this world that glass bangles slowly come to be an increasingly common ornament, a process that likely bore significantly different regional histories of its uptake and growth in frequency. The first of such circulation spheres connects the western Deccan, south India and Sri Lanka. A range of distinctive types can be discerned as fashioned out of a glass that appears black and opaque in most cases, but with hues to brown on regular occasions. This distinct group bears additive decorations with stylized yellow circles arranged in a set of regularly repeated patterns (Fig. 6). Dikshit describes this group as bearing ‘irregularly made circlets, volutes, chevrons or zig-zag lines’, but he remained conflicted about whether the evidence attributed them to the ‘MuslimMaratha period’ or earlier (Dikshit, 1969: 68). Such bangles appear to have reached

376

M. Trivedi

Fig. 6 Principal types ascribed to the early medieval period

sites in all these three regions (Ujjain, Nevasa, Prakash, Yeleswaram, Mantai). At Mantai, Peter Francis Jr. allocated these, and simpler variants of black bangles with a yellow trail decoration, to the period broadly between the eighth and eleventh centuries (2013: 383). Another distinctive black bangle that is oblong in profile and bears a series of parallel grooves on its upper face—veritably, the ‘Mantai combed type’ also seems to be typical of this period and is also found at sites like Nevasa. Its unique morphology suggests that this form may be the result of a common grooved shell bangle type that comes to be translated into glass. A group of broad oblong black glass bangles recovered from the Sri Lankan site of Vankalai, dated to the period 1200–1250, was included among Brill’s South Asian glass samples and they appear to form an independent cluster within known mineral soda alumina compositions (Brill, 1999: 332–340). A concerted programme of analysis from several peninsular

Glass Bangles in South Asia: Production, Variability …

377

sites of these black bangles is likely to yield several technologically distinct series that may further be separable in time or chemistry. Elements from this complex of dark bangles with yellow/white additive decorations reach the site of Sirpur (Chhattisgarh), where excavations yielded a unique glass bangle assemblage. While no detailed study had been published, Dikshit rightly recognized the uniqueness of this assemblage, which can be placed likely in the late twelfth and early thirteenth century. It appears that a bangle-making workshop was also excavated ‘in a house’. The bangles from Sirpur display the varied use of marvered and unmarvered trails, towards some designs that are unique, and others that are very rare elsewhere. These include ‘figure of 8’ designs, well-executed and left unmarvered. The more intriguing and unique realization of an additive possibility is what Dikshit described as ‘wound’ and involved over a base bangle the addition of another ‘rod … by coiling it round and round and thus producing an irregular ropelike pattern’ (Plate III-D in Dikshit, 1969). Dikshit’s assessments of the comparative uniqueness of the assemblage and the evidence for on-site production make clear that a typological and compositional analysis of the Sirpur glass bangle assemblage and production waste promises to teach us much about Indian glass at a time of marked cultural and technological transitions. Another network, connecting the Tamil region with the wider world is extremely significant, for it provides a temporally specific anchor in this period and raises questions of exchange, craft and mobility. Dussubieux’s (2009) analyses of glass samples from the site of Lobu Tua in Sumatra include a few black bracelets. Lobu Tua has been known since 1892 to Indian audiences as the find-spot of an inscription, dated Saka 1010, CE 1080, of the trade guild known as ‘the Five Hundred of the Thousand directions’ or the Ayyavole 500 (Abraham, 1988; Subbarayalu, 1998). Recent excavations have revealed the site to be a briefly inhabited trading entrepot, occupied between the mid-ninth to end eleventh centuries. In this period, the site presents distinct enclaves of non-local trader settlement with material repertoires distinct from those of indigenous habitation sites. In terms of the glass alone, the site bears 9000 fragments of mostly blown glass, items travelling along the maritime silk route between China and the Islamic Middle East in both directions (Guillot et al., 2003: 223–271). The black, mostly plano-ovate bangles are few and represent wellmade types within the early medieval spectrum. As Guillot notes, their interpretation presents a distinct challenge, one which is heightened by the enduring identification of these bangles as Chettiyar (or Tamil Merchant) bangles into the present day in the nearby Aceh region (2003: 267). Dussubieux’s analysis demonstrates three different phases in the compositions and probable origins of the glass bangles of the Sumatran region. At Loba Tua, these c. eleventh-century bangles, with a high likelihood of association with Tamil merchants, are not South Asian mineral soda alumina (m-Na-Al) glass; they are rather a high soda vegetal glass (or v-Na-Ca), a composition thought to originate from the Middle East. The presence of migrant trading groups in combination with the small frequencies of these artefacts suggests that at least some of this evidence relates to migrant trader families moving with their cherished ornaments which themselves were drawn from complex networks of the v-Na-Ca glasses.

378

M. Trivedi

The most detailed published chemical analysis of glass bangles in India conducted so far provides significant contextual evidence for how and why these v-Na-Ca glass bangles might be associated with peninsular traders. This is the analysis of glass bangles from the ‘Megalithic’ and later occupations at the site of Hirapur (Pawar et al., 2014). At Hirapur, in a sequence that is not published with any associated radiocarbon absolute chronology, a range of plant ash glasses (v-Na-Ca) and two different types mineral soda Alumina (m-Na-Al groups 2 and 4) glasses were used to fashion annular black bangles. Two remarkable trends emerge from this analysis, first, that at one site, visually indistinguishable artefacts were fashioned from very different raw glasses. Second, that this aspect of local production endured over long periods of time suggests the networks of raw glass supply to bangle-makers likely involved distinct raw glass producers using different recipes and also likely involved recycling by the bangle-maker. This complexity merits much further study and only when we have analysis from a range of well-dated sites and a typology to describe variation in the bangles themselves can we begin to isolate patterns. Recent analysis of glass vessels and bangles from eleventh-century Ghazni in Afghanistan has isolated one component of that assemblage as another v-Na-Ca glass (Fiorentino et al., 2019). Surprisingly, one of the v-Na-Ca Ghazni bangles demonstrates significant association with a group of the Hirapur bangles indicating how much we have to learn about the glass bangle networks of exchange and value. Between the tenth century and the end of the twelfth century, ornament cultures in different parts of the South Asian subcontinent transitioned to include glass bangles more and more regularly. How and when and where these transitions took place are probably very varied and have regionally specific histories in different parts of South Asia. We know of only the Sirpur ‘factory’ attributed to this period, but it is highly probable that smaller kilns of itinerant producers have not been recognized, nor been looked for, nor recorded or analysed seriously when seen (Selvakumar, 2021). Without detailed study of significant assemblages, such as those from Lalkot at Delhi, it is hard to understand how these changes relate to older glass bangles in India, and what the nature of this change, its pace and tempo, or what its social and cultural significance is. It is in this period that the bangle-maker, as an artisan with a caste-identity clearly emerges. An inscription dating 1261 CE from Belgaum appears to be one of the earliest epigraphic references that provides us a sense of the distinctive identity of bangle production in medieval Karnataka. The inscription, a land grant which trades shares in a village, refers to this deal being as that on same terms as ‘that of Senalli, Kallakundarage and Nittiru, which were Balagara-Sthala (centers of Banglemanufacture)’ (Panchamukhi, 1941: 139). The inscription demands consideration of the imbrication into wider social relations of the bal.eg¯ar and how deeply marked places associated with them were. The inscription demands that we historically and archaeologically consider the formation of the bangle-maker into a caste. We must consider not only the consolidation of this social role and group, but also their presence in networks of client and patron and through such places in the landscape and the tensions that likely existed between itinerancy and being rooted in rural communities.

Glass Bangles in South Asia: Production, Variability …

379

5 Conclusion: Enduring Questions in the Study of South Asian Glass Bangles At different times in South Asian history, the glass bangle has been rare and uncommon, at others ordinary and at yet others deeply valued for reasons beyond its being an ornament alone. In addition to this, the production techniques, the decorative grammar, the sociology of its craftspersons and their kin and labour organization, the relationships to client and markets, the scales at which it was produced, the distances it was traded and the combined ways in which bangles were valued—these are elements of the bangle’s complex material histories that we are only now at the cusp of beginning to disentangle. In conclusion, this chapter points to the research agenda necessary for the further study of the glass bangle in south and Southeast Asia. Detailed multi-variate study of the earliest assemblages, which builds towards the definition of types is important and essential. It is through these studies alone that regional scale analysis will be possible of the evidence from the BCE centuries in different regions. The situation for the early historic and early medieval is similarly unclear, and in all of these phases attention to defining specific series of ‘monochrome’ types will aid the essential tasks of chronological clarification. The production techniques listed here: the two-mandrel, the drawn round rod, the Khalbut are likely only three points in a complex array of techniques used to author bangles in the region’s archaeological history. Only through careful assemblage level observation of plastic and chemical traces, inclusions and accretions will we be able to arrive at a fine-grained sense of when which technique was introduced, how and where they flourished and came to run in parallel or declined. A definition of series of types for the last millennia is needed for each region and every major assemblage. Chemical analysis is essential for it will yield key insights into the ways in which glass bangle craftsmen were reliant upon raw glass producers. Studies of assemblages may tell us much about the possibilities of whether bangle-makers coloured their base glass to draw the trails and canes they desire. Attention to not just the bangles, but the deformed rejects, the distinctive waste of the small kilns of the itinerant churihar and the waste products will help us establish the kinds of glass these craftsmen carried with them, how they worked it and coloured it and what networks of knowledge materials and praxis they so constituted. Preliminary evidence suggests that most glass bangle assemblages from the last millennium will have a range of contributions from both m-Na-Al and v-Na-Al glasses of likely South Asian origin, but also a range of others, traded in from the Middle East (v-Na-Ca) and even those from further afield. This complexity is only multiplied in the most recent past two centuries when industrial glasses of various origins, both foreign and South Asian come to be widely distributed as bangles. Implicitly it has been assumed that bangles are of less interest than glass beads, for example, as the latter come to be produced in large quantities and are traded long distances. Yet, precisely because of these more locally enmeshed mechanisms of itinerant production and exchange, the glass bangle affords critical opportunities for the archaeometry of South Asian glasses.

380

M. Trivedi

In contrast in the study of La Tene pre-Roman Iron Age European glass bangles, typological debate has extended over two generations (Gebhard, 1989; Haevernick, 1960). Building on these studies, in the last decade a truly multidisciplinary project combining experimental replication, trace element chemistry and regionallevel comparisons has broken new ground and sets an exemplary model for the study of glass bangles (Rolland, 2017). The experimental replication of glass bangles, using the two-mandrel technique, was conducted by building the smaller two-chamber temporary kiln Gaborieau documented at Kurahan in Nepal. This has led to insights in the technical expertise needed to fire this kiln, maintain an adequate temperature, work the glass(es), form the initial bead—enlarge the bangle and attempt the distinct range of plastic and additive decorations that were distinct to the La Tene era. Rolland’s studies track how the technical gestures and requisite expertise to make a bangle relate to the typological development and its regional variations. Through these combined means Rolland establishes for La Tene era Europe, a model of itinerant bangle markedly focused away from large urban centres. Across the European landscape, ideas of style, craft and change therein can be attended to with a specificity that is multi-variate and temporally specific to the scale of 25–50 years (Rolland, 2017). The study of South Asian glass bangles must draw much inspiration from Rolland’s research programme. While glass beads have been given far more of our attention, glass bangles, on account of their technical diversity, their demonstrable chemical variation, the typological problems which remain open and the intimate integration of the production of this ornament within wider socio-economic structures over the past 1000 years—holds the promise of teaching us far more about the glass-worlds of the subcontinent. On the basis of this preliminary synthesis, they seem to be differently distributed, change repeatedly in marked and typological ways and articulate networks of expertise and exchange that dendritically fan out from the most local of scales to those across the Indian ocean worlds. The jingling of glass bangles as Dikshit noted was in use since c. 1250 CE as a metaphor; we are, slowly but surely now, closer to hearing its distinct notes.

References Abraham, M. (1988). Two medieval merchant guilds of South India. Manohar Publications. Ambekar, A., Deshpande-Mukherjee, A., & Maity, P. (2019). The conch bangle industry through the ages at ancient Vadnagar: Preliminary observations. Archaeo + Malacology Group Newsletter, 31, 4–8. Bettineschi, C., Angelini, I., & Vidale, M. (2019). A first look at the development of glass and glass materials industries in the sequence of the early historic site of Barikot (Swat, Khyber Pakhtunkhwa, Pakistan), paper presented in the Conference cum Workshop on History, Science & Technology of Ancient Indian Glass, 21-25 January 2019, IIT Gandhinagar, India. Brill, R. H. (1999). Chemical analyses of early glasses (Vol. 2). The Corning Museum of Glass. Cherian, P. J., Tomber, R., Abraham, S. A., Guimlia-Mair, A., Kelly, G. O., & Nayar, P. (2016). Items of personal adornment from Pattanam, Kerela. Journal of Indian Ocean Archaeology, 12, 34–63.

Glass Bangles in South Asia: Production, Variability …

381

Dikshit, M. G. (1969). History of Indian Glass. University of Bombay. Dobbs, H. R. C. (1895). A monograph on the pottery and glass industries of the North-Western Provinces and Oudh. North-Western Provinces and the Oudh Government Press. Dussubieux, L. (2009). Compositional analysis of ancient glass fragments from North Sumatra, Indonesia, In D. Perret, & H. Surachman (Eds.), Histoire de Barus III: Regards sur une Place Marchande de l’Océan Indien (XIIe - milieu du-XVIIe s) (pp. 384–417). Association Archipel/EFEO. Dussubieux, L., Gratuze, B., & Blet-Lemarquand, M. (2010). Mineral soda alumina glass: Occurrence and meaning. Journal of Archaeological Science, 37(7), 1646–1655. Dussubieux, L., & Bellina, B. (2018). Glass ornament production and trade polities in the Upper-Thai Peninsula during the early Iron age. Archaeological Research in Asia, 13, 25–36. Dussubieux, L., & Bellina, B. (2017). Glass from an early Southeast Asian producing and trading centre. In B. Bellina (Ed.), Khao Sam Kaeo: A late prehistoric early port-city between the Indian Ocean and the South China Sea (pp. 549–585). Ecole française d’Extrême-Orient. Fiorentino, S., Venezia, B., Schibille, N., & Vandini, M. (2019). Streams across the silk roads? The case of Islamic glass from Ghazni. Journal of Archaeological Science: Reports, 25, 153–170. Francis, Jr., P. (1982). Glass beads of India. Lapis Route Books/The World of Beads Monograph Series 7. Francis, Jr., P. (1994. Toward a social history of beadmakers. BEADS: Journal of the Society of Bead Researchers, 6, 61–80. Francis, P., Jr. (2013). The glass bangles and rings. In J. Carswell, S. Deraniyagala, & A. Graham (Eds.), Mantai: City by the sea (pp. 381–385). Linden Soft Verlag. Gaborieau, M. (1977). Bracelets et grosses perles de verre Fabrication Et Vente En Inde Et Au Népal. Objets Et Mondes Paris, 17(3), 111–130. Gaborieau, M. (1993). Ni brahmanes ni ancêtres: colporteurs musulmans du Népal. Société d’ethnologie. Gebhard, R. (1989). Pour une nouvelle typologie des bracelets celtiques en verre. In M. Mergoil (Ed.), Le verre préromain en Europe occidentale (journée d’étude de Lyon, 8 novembre 1986) (pp. 73–83). Guillot, C., Surachman, H., Perret, D., et al. (2003). Histoire de Barus. Le site de Lobu Tua. II: Etude archéologique et Documents (Vol. 30). Cahiers d’Archipe. Haevernick, T. E. (1960). Die Glasarmringe und Ringperlen der Mittel-und Spätlatènezeit auf dem europäischen Festland. R. Habelt. Hallifax, C. J. (1892). Monograph on the pottery and glass industries of the Punjab, 1890–91. Hawkes, J. D. (2014). Finding the “Early Medieval” in South Asian archaeology. Asian Perspectives, 53(1), 53–96. Hawkes, J. D. (2014). Chronological sequences and the problem of early medieval settlement in India. Puratattva, 44, 209–228. Kanungo, A. K. (2004). Glass beads in ancient India and furnace-wound beads at Purdalpur: An ethnoarchaeological approach. Asian Perspectives, 43(1), 123–150. Kanungo, A. K., & Brill, R. H. (2009). Kopia India’s first glassmaking site: Dating and chemical analysis. Journal of Glass Studies, 51, 11–25. Kanungo, A. K. (2013). Glass in ancient India: Excavations at Kopia. Kerala Council for Historical Research. Kanungo, A. K. (forthcoming). Glass bangles in India: Antiquity, functional use and traditional production. Karwowski, M. (2010). Spätantike Glasarmringe aus Niederleis in Niederösterreich. In A. Urbaniak (Ed.), Terra Barbarica, Monumenta Archaeologica Barbarica, Series Gemina, Tomus II (pp. 283– 299). Łód´z-Warszawa. Kenoyer, J. M. (2003). Bangles. In P.J. Claus, S. Diamond, & M.A. Mills (Eds.), South Asian Folklore: An encyclopedia: Afghanistan, Bangladesh, India, Nepal, Pakistan, Sri Lanka (pp. 51– 52). Taylor & Francis.

382

M. Trivedi

Lal, B. B. (1954–55). Excavation at Hastinapura and other explorations in the upper Ganga and Satlej basins 1950–52: New light on the dark age between the end of the Harappa Culture and the early historical period, Ancient India 10–11, 5–151. Lal, B. B. (1952). Examination of some ancient Indian glass specimens. Ancient India, 8, 17–27. Lal, B. B. (1952). Studies in early and mediaeval Indian Ceramics—Some glass and glass-like artefacts from bellary Kolhapur, Maski, Nasik and Maheshwar. Bulletin of the Deccan College Post-Graduate and Research Institute, 14(1), 48–58. Maharashtra Census Office. (1968). Bangles at Tarapur, Census of India 1961, X, part VII-A(7), pp. 27–42. Bombay. Mishra, A., Singh, D., & Sharma, A. (2010). A study of glass bangles from Abhaipur, District Pilibhit, Uttar Pradesh. Man and Environment, 35(1), 103–108. Panchamukhi, R. S. (1941). Karnatak inscriptions: With introductory notes in English (Vol. 2). Kannada Research Office. Pawar, K., Lankton, J., Gratuze, B., & Yongjun, K. (2014). Glass bangles from Hirapur, Chadrapur District, Maharashtra. Man and Environment, 39(2), 64–72. Rolland, J. (2017). L’artisanat du verre dans le monde celtique au second âge du fer: Approches archéométriques, technologiques et sociales. [PhD Thesis, Université Paris 1—PanthéonSorbonne]. Sahi, M. D. N. (1988). Beginning of glass-technology and glass bangles in India. Proceedings of the Indian History Congress, 49, 640–644. Sankalia, H. D. (1947). The antiquity of glass bangles in India. Bulletin of the Deccan College Post-Graduate and Research Institute, 8, 252–259. Sankalia, H. D. (1952). Bangles. In H. D. Sankalia, & M. G. Dikshit (Eds.), Excavations at Brahmapuri (Kolhapur) (pp. 115–121). Deccan College Post-graduate and Research Institute. Sankalia, H. D. (1960). From history to pre-history at Nevasa: Report on the excavations and explorations in and around Nevasa (1954–56). Deccan College Post-Graduate & Research Institute. Selvakumar, V. (2021). History of glass ornaments in Tamil Nadu, South India: Cultural perspectives. In A. K. Kanungo, & L. Dussubieux (Eds.), Ancient glass of South Asia—Archaeology, Ethnography and Global Connection. Springer Nature/IIT Gandhinagar. Spaer, M. (1988). The pre-islamic glass bracelets of Palestine. Journal of Glass Studies, 30, 51–61. Spaer, M. (1992). The Islamic glass bracelets of Palestine: Preliminary findings. Journal of Glass Studies, 34, 44–62. Subbarayalu, Y. (1998). The Tamil merchant-guild inscription at Barus: A rediscovery. In C. Guillot (Ed.), Histoire de Barus (Sumatra): Le site de Lobu Tua (Vol. 1. Etudes et Documents) (pp. 25–34). Archipel. Thapar, B. K. (1957). Maski 1954: A chalcolithic site of the southern Deccan. Ancient India, 13, 4–142. Thapar, B. K. (1967). Prakash 1955: A chalcolithic site in the Tapti valley. Ancient India, 20–21, 5–167. Trivedi, M. (2008). Multiple pathways in the emergence of complex societies in the Ganga Basin: Perspectives from a survey of Kausambi (2500 BCE to 600 CE). [M. Phil Dissertation, Centre for Historical Studies]. Jawaharlal Nehru University. Trivedi, M. (2020). An archaeology of virtue: Tradition, Islam and the materiality of conversion in Medieval North India. [PhD Thesis, University of Chicago]. Trivedi, M., & Rajasthan Department of Archaeology and Museums. (2016). Report on salvage archaeology excavations at Indor, Report on File. Varshneya, A. K., Tong, S. S., & Gogte, V. (1988). Analysis of early glass objects from Nevasa excavations in India. Transactions of the Indian Ceramic Society, 47(5), 149–155.

West Asian Glass in Early Medieval India as Seen from the Excavations of Sanjan Kurush F. Dalal and Rhea Mitra-Dalal

Abstract The site of Sanjan, Dist. Valsad, Gujarat (India), was excavated from 2002 to 2004. While excavating here for three seasons, we realised that we were looking at the remains of an urban early medieval centre engaged with West Asian merchants trading the Maritime Silk Route from the Persian Gulf to China. Among the artefacts recovered during the excavations were encountered thousands of glass fragments. This discovery was unanticipated and also unprecedented, perhaps due to the lack of investigations into this period in this region. Subsequent academic perusal of this database led to the identification of numerous shapes as belonging to the period between the tenth and twelfth century CE which, much to our surprise, were made in workshops in West Asia and Asia Minor. The Sanjan glass thus opened a new chapter in the research into glass in the Indian subcontinent. This paper is a review of our preliminary work done on the glass objects recovered from the Sanjan excavations. It includes not just the foreign glassware but also lists the continuing Indian traditions of beads and bangles manufacture which continued side by side with the imports.

1 Introduction This paper deals with the glass finds and especially the ones of probable West Asian origin excavated at the site of Sanjan, Taluka Umargam, District Valsad, Gujarat (20°11 59.6 N; 72°48 00.2 E). The site of Sanjan (Fig. 1) was excavated for three seasons between 2002 and 2004 (Gupta et al., 2002a, 2002b, 2003, 2004a, 2005; Mitra & Dalal, 2005) as part of a project initiated by the World Zarathusti Cultural Foundation, Mumbai. The main purpose of which was to look for evidence of the landing and occupation at the site by Zoroastrian refugees from Iran in the eighth century CE. After the fall of the Sassanian Empire, there was a systematic proselytization of the populace of Persia and the remaining Zoroastrians faced adverse conditions. A group of Zoroastrian refugees who had fled to the hills of Khorasan (which had remained in the hands of K. F. Dalal (B) · R. Mitra-Dalal INSTUCEN School of Archaeology, 103 Sunshine, Plot 58, Sector 21, Kharghar, Navi Mumbai 410210, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_16

383

384

K. F. Dalal and R. Mitra-Dalal

Fig. 1 Sanjan with its ancient and modern localities

Zoroastrian princes for the next 100 years) came down to Hormuz after Khorasan was subdued by the Islamic forces and then after a stay of 30 years finally took ship for India. This is the origin story of the Parsis of India and is told in a quasihistorical poem, written in December 1600 CE, called the Kisseh-i-Sanjan (Edulji, 1991). According to this text and the oral traditions of the Parsis, their ancestors made landfall on mainland India 149 years after the fall of the Sassanid dynasty (641 CE) around 790 CE at Sanjan on the west coast of India. They prospered here and spread to various parts of India (Kamerkar & Dhunjisha, 2002) before Sanjan was sacked sometime around 1300 CE by Sultan Mahmud’s general Alf Khan. There is much debate which Sultan is being alluded to here as the text is dated to 1600 CE, and there were numerous Sultans named Mahmud (many with generals called Alf Khan, which is a title and not really a name) who marched through Gujarat or ruled over it in the period prior to the writing of this poem. The earliest of these was Sultan Mahmud Alauddin Khilji who marched through this region on his way to the Deccan in 1297 CE (Roy, 1960). Excavations revealed a port site (Figs. 2 and 3) and entrepot that existed between the eighth and fourteenth century CE, with its heyday between the tenth and twelfth century CE, and a huge West Asian connection in the form of ceramics, glassware and other artefacts. Prior to the excavations at the site, very little, if anything, was known about early medieval glass in India and even less about the possible import and influences of glass from West Asia during the intense period of Indian Ocean Littoral trading during the early Islamic Period (eighth to thirteenth century CE).

West Asian Glass in Early Medieval India as Seen …

Fig. 2 Anchorage and freshwater lake (manmade) at Sanjan

Fig. 3 Seagoing vessel at Sanjan

385

386

K. F. Dalal and R. Mitra-Dalal

Fig. 4 Brick built structure

The excavations revealed a bustling port city with brick-built houses (Fig. 4), wells, sanitation and drainage (Fig. 5) and evidences of wealth. While excavating here for three seasons (2002, 2003 and 2004), we realized that we were looking beyond any doubt at the remains of an urban early medieval centre engaged in trade with West Asian traders trading through the Maritime Silk Route from the Persian Gulf to China.

2 The Evidence from Sanjan The excavations yielded a treasure trove of data. Previously unidentified glazed wares were recovered in large numbers (Fig. 6) along with the remains of brick houses and drainage systems using ring wells as soak pits to limit the spread of kitchen and bathroom waste. A very large and diverse range of numismatic findings were very useful for dating the site ranging from Abbasid Dinars, Rashtrakutas, Amirs of Sindh, Gadhaiyya Coins, Yadava Coins and a coin of Md Allauddin Khilji (Fig. 7). Beads of glass, stone and terracotta were also recovered (Fig. 8) as were glass bangles (Fig. 9). Agate beads dominated the terminal phase at the site (Fig. 10). The site was littered with sculptural and structural remains of one or more Hindu temples which could be attributed on art historical grounds to the Rashtrakuta-Shilahara Period, i.e. the ninth to the fourteenth century CE (Figs. 11 and 12).

West Asian Glass in Early Medieval India as Seen …

Fig. 5 Ring wells

Fig. 6 West Asian ceramics from Sanjan

387

388

Fig. 7 Coins from the Sanjan excavations

Fig. 8 Beads from Sanjan

K. F. Dalal and R. Mitra-Dalal

West Asian Glass in Early Medieval India as Seen …

389

Fig. 9 Intact glass bangles from Sanjan

Fig. 10 Agate beads collected at Sanjan

Among the artefacts recovered during the excavations, we encountered huge numbers of glass fragments. This discovery was unanticipated. It was also unprecedented, perhaps due to the lack of investigations into this period in this region. Subsequent academic perusal of this database led to the identification of numerous shapes as belonging to the period between the tenth and twelfth century CE which, much to our surprise, were made in workshops in West Asia and Asia Minor. This opened a new chapter in the research into glass in the Indian subcontinent.

390

Fig. 11 Sculpture

Fig. 12 Temple fragment (Gajathara)

K. F. Dalal and R. Mitra-Dalal

West Asian Glass in Early Medieval India as Seen …

391

3 The Glassware from Sanjan 3.1 Introduction to the Glass When we started the excavations at Sanjan, we had no idea that we would find the quantity of glass that we did. Almost from the first lot in the first quadrant we opened, we started finding small chips and fragments of glass vessels. Though we bagged them all, we were quite convinced that this was modern detritus, as the site, was occupied by the local fishing community, the Maachhi. It was only as we penetrated beneath the humus that we realized quite what we were looking at. Since almost no such parallels were hitherto known from India and none of the excavation team had any idea what we were really looking at, we quietly went into a standard collection mode. The artefacts were hand sorted from the excavated soil and subsequently from the sieved soil. The fragments were then sorted and packed in cloth lined paper envelopes if large and in plastic boxes with cotton wool in case of smaller objects. The packets were labelled with all details of their provenance as is the standard practice. These packets were registered in the antiquity register, and each was given an individual registration number. It was only when we sat to try and analyse the data that we realized that we should have set up a small glass yard along the lines of a ceramic yard and sorted the fragments on site itself. On our return to Mumbai, the data were sorted on the basis of diagnostic versus non diagnostic fragment, shape, colour, decorations, state of preservation, etc. An attempt was then made to identify the specimens.

3.2 Previous History of Research Research on medieval glass in India is virtually non-existent. As a matter of fact, research into the early medieval period in India (eighth to fifteenth century CE) is also extremely rare. This is a field area frequented by art historians and historians but seldom by archaeologists. Where glass is reported, it is rarely described in any detail. The work done on glass in Ancient India comprises a single volume made up of a lecture series delivered in the 1940s by M.G. Dikshit (1969) and subsequent work by Alok Kanungo, here we may also add the work done by Kock and Sode (2002). Sadly, most of this work is either dated (as in the case of Dikshit) or belongs to the very early historic period (Kanungo, 2016) or modern ethnographic work (Kanungo, 2002). The severe paucity of work done on Indian glassware has been a major hurdle in carrying out research on the Sanjan glass, and the authors had to perforce turn to foreign authorship to fill this void.1 The first stroke of luck was a 1

Note must be made here of work on Indo-Pacific Beads done by Peter Francis, Jr. Alok Kumar Kanungo, Sunil Gupta, Vishwas Gogte, Felix Chami and Rahul Oka to name just a few. What must

392

K. F. Dalal and R. Mitra-Dalal

detailed correspondence with the Corning Museum of Glass and its former Director, Dr. Robert Brill via his secretary Ms. Shauna Wilson. A large portion of the initial work was done with his advice and help. The other sources of direct import were Stefano Carboni’s catalogue of the al-Sabah Collection in Kuwait (Carboni, 2001), and the material from the Serçe Limani shipwreck for which Dr George F Bass was kind enough to share his unpublished (since published) drawing of the vessels found on board (Bass et al., 2009). There have been a few other subsequent databases that the authors have been able to use. The shortcoming though is that in the absence of a centre for the study of archaeological glass in India, the database has never been done adequate justice.

3.3 Apart from the Vessels Apart from the vessels which appear to be solely of foreign origin, the site has also yielded a number of glass beads, mostly non-local, and glass bangles/bangle fragments which in turn appear to be very local. The non-local beads are mainly typically West Asian eye beads (Fig. 13), the likes of which are seen at numerous sites in Syria, Arabia and even in Southeast Asia (Chaisuwan & Naiyawat, 2009; Nakai & Shindo, 2013). There are also a number of segmented gold foil beads (Fig. 14), one example of a purple bitruncated hexagonally faceted tubular bead, and this last bead is highly enigmatic and unlike any bead hitherto known from India (see Fig. 8 middle row right). Fig. 13 Glass eye beads

be noted though that this type is studied in the period between the sixth century BCE and the fourth century CE and is of little consequence to this study. This is not to belittle any of this amazing research just to point out that the early medieval is virtually unrepresented.

West Asian Glass in Early Medieval India as Seen …

393

Fig. 14 Segmented (gold foil) beads

Bangle fragments are found from all layers at the site and from inside the Dokhma/Tower of Silence (Parsi mortuary structure) (Fig. 15) which was excavated in Season 3. Historical research with community elders has revealed that Zoroastrian women adopted the bangles as a sign of matrimony and till approximately 250 years ago demanded to be placed in the Dokhma with these symbols of matrimony upon them if they predeceased their husbands. Many complete (and in one case articulated) bangles (Fig. 16) have been recovered intact from inside the Dokhma. Most of them are an opaque black in colour though there are examples in other colours and also bichrome examples.

3.4 The Vessels The Sanjan glassware (Mitra & Dalal, 2005) consists of a variety of daily use objects, tableware, storage vessels, small perfume/unguent jars and miscellaneous artefacts. The majority of the fragments are tiny and measure less than 0.5 cm in length, with many small slivers. The authors have been unable to make much headway with these objects. Despite this problem, the Sanjan glassware collection numbers in the thousands, there are a reasonable number of identifiable fragments both small and large. The fragments vary greatly in colour too. Colours range from light to dark green (predominant), pale blue, dark blue (uncommon), clear, yellow, red, black and a very small number of opaque white examples. Most of them are heavily patinated with typical silica skeletons in rainbow-coloured nacre which comes off to the touch. This

394

K. F. Dalal and R. Mitra-Dalal

Fig. 15 Plan of the Dokhma at Sanjan

is mainly because of the pH of the soil reacting with the buried glass over a long time. The fragments have been categorized into seven groups: rims, bases, body sherds, neck fragments, handles, spouts and complete/clearly identifiable objects. As expected, body sherds form the single largest group, base and rim sherds are almost equal in numbers. Necks, spouts and handles are very rare. There is only one

West Asian Glass in Early Medieval India as Seen …

395

Fig. 16 Articulated bichrome bangle from the Dokhma

complete vessel, one fragmented vessel, one vertically split vessel and a handful of objects that can be specifically identified. A small number of fragments had appliqué decorations and in one case incised decorations. There are also instances of pinching, and small bow tie like decorations are seen on a number of fragments. One of the common features of the bases is a deep kick at the centre and whether this is an aesthetic factor or a result of the manufacturing process is not clear.

4 Important/Significant Discoveries from Sanjan Include An 11 facetted complete unguent/perfume bottle (Fig. 17 right) is made of clear glass. It is flat-based with 11 straight sides and a flared shoulder from which a straight vertical neck ends in a slightly outward flaring rim. Two identical bottles, one with nine and another with 12 facets are found in al-Sabah Collection, dating to the ninth–tenth century CE (Carboni, 2001: 100–101). Whitehouse (1968; Pl VIIb) has reported a similar bottle from Siraf. Sadly, the report is yet awaited. Similar faceted bottles are seen from the excavations of the Cirebon shipwreck (Needel, 2018: Fig. 9, middle row third and fourth from left). A small cuboidal bottle of clear glass with a square base, straight sides and small neck which is unfortunately broken (Fig. 17 left): it is similar to one illustrated from the al-Sabah Collection (Carboni, 2001: 245, Cat. 3.32). There is an identical one reported from the Serçe Limani wreck (Bass et al., 2009). A tubular necked round bottom flask (small) with flaring rim (Fig. 18): this one is in green glass and is very thin and fragile. Similar vessels have been reported

396 Fig. 17 Small bottles from Sanjan

Fig. 18 Small globular bottle in green glass

K. F. Dalal and R. Mitra-Dalal

West Asian Glass in Early Medieval India as Seen …

397

Fig. 19 Basal knob of a polycandelon lamp

at Banbhore (Khan, 1960), Samarra (Kröger, 1995: Catalogue No. 60) and Serçe Limani (Bass et al., 2009). Lamp base/finial: this is an object that we could only identify thanks to the data from the Serçe Limani wreck (Fig. 19). It is an inverted drop/balloon-shaped waisted and elongated knob attached to a horizontal body. Large numbers of these are seen in the Serce Limani wreck (Bass et al., 2009). This is part of a metal polycandelon on which numerous such glass lamps were hung in circular holders attached on top a larger circular frame which was suspended from the ceiling. These were the precursors of modern chandeliers. One of the rarest examples is that of a footed plate. It is a fragment with one foot still attached to it (Fig. 20). There are numerous intact flat-rimmed bottle tops with a small portion of the neck still attached (Fig. 21). Similar ones have been dated to the ninth–tenth century CE (Carboni, 2001: 146–47, 202). Some of these have balloon-shaped necks, one has a ribbed neck, and others have simple tapering necks. Similar examples are seen in the Serçe Limani (Bass et al., 2004: 267, Fig. 15–2 top row objects 1 and 3 from left and bottom row object 2 from left), the Cirebon (Needell, 2018: Fig. 5 bottom row second, Fig. 8 top row first and third, Fig. 9 top row second, Fig. 10 top row fourth Fig. 11 top row third and second row first, Fig. 12 Top row), and at Ramla (Gorin-Rosen, 2010: 234, Pl. 10.6. 17, 18). There is one very unique vessel fragment. It is part of a heavily decorated bowl (Fig. 22). It was decorated with ridged appliqué bands which are thick near the rim

398

K. F. Dalal and R. Mitra-Dalal

Fig. 20 Footed plate

Fig. 21 Flat-rimmed bottle top

and thin towards the base. The ridges are further decorated with thumbnail incisions. The rim is plain, and it has a ledged shoulder. It is too patinated to make out the colour.

West Asian Glass in Early Medieval India as Seen …

399

Fig. 22 Decorated bowl fragment

There is a tiny fragment of pale green glass with a bow tie-shaped pinched decoration. A number of these have been found at Sanjan, and all are on pale green glass. The stem and part of the base of a chalice have been recovered from the site (Fig. 23). It is very similar to one displayed in The Metropolitan Museum of Art, New York, the provenance for which is Siraf. The stem is made up of eight distinct globules stacked one atop the other. It is a clear glass. Two spouts belonging to cupping vessels or alembics (Fig. 24) have also been recovered from the excavations. Both are in pale green glass. Unfortunately, the ‘cup’ is absent in both cases. Similar examples exist in the al-Sabah Collection (Carboni, 2001: 144–145), from excavations to the north of the White Mosque at Ramla (Israel) (Gorin-Rosen, 2010: 227, Pl 10.2.18–21) and on the Serçe Limani wreck (Bass et al., 2009: 377–384). There is also from the collection at Sanjan one thick, marvered disc (Fig. 25). It is heavily weathered. It appears to be a decorative member originally luted to a vessel which has subsequently fallen off. One such jar with numerous marvered discs is illustrated in the al-Sabah Collection (Carboni, 2001: 291–293).

400

K. F. Dalal and R. Mitra-Dalal

Fig. 23 Stem and partial base of a glass chalice

Fig. 24 Spouts of cupping vessels or alembics

5 Conclusion The study of the Sanjan glass is far from over. The scientific analyses of the samples are yet to be carried out. What we can say though the morphological and comparative

West Asian Glass in Early Medieval India as Seen …

401

Fig. 25 Marvered disc

study is that the glass recovered from Sanjan is in quantity unlike any other site of a similar period from Western India. The vessels and many of the beads are most definitely West Asian in origin. Early medieval horizons have been encountered at Kamrej (Gupta et al., 2004b), Chinchani (Gogte et al., 2006a), Chaul (Gogte et al., 2006b), Kelshi (Joglekar et al., 2002), Chandore (Dalal, 2017) and Vizhinjam (Kumar et al., 2013). Glasses have been recovered, but it is usually just a few pieces. The sheer numbers seen at Sanjan are unknown from any other contemporary Indian sites. Almost without exception, the glass appears to be of foreign-West Asian origins. This is quite peculiar considering that the manufacture of glass is known in India from the tenth century BCE in Karnataka and from at least the fifth century BCE onwards in northern India (Dalal, 2001: 94–98). Many of the beads and all of the bangles seen at Sanjan are of local manufacture. Yet the vessels are foreign. This is a dichotomy difficult to explain. It has been proven that there was a large Zoroastrian community resident at Sanjan, and they have no glass using taboos unlike the Hindu upper castes who believe that only metal utensils can be purified and not those made of glass or ceramic. The authors have wondered whether this is one reason for the sheer volume of glass excavated by the team. What we can say with surety is that Sanjan was a flourishing urban port during the period between the eighth and the thirteenth century CE and that it saw its heyday in the period between the ninth and twelfth centuries as evidenced by the ceramic, numismatic, glass data and the huge quantities of West Asian ceramics found there. Acknowledgements We thank ICHR for granting us a project to carry out the work; the ASI for permission for three seasons; The Dorab Tata Trust for helping finance Season 3 (2004); the Indian Archaeological Society, New Delhi, for their expertise, technical staff and all other things; the WZCF for initialising the project; the Alpaiwala Museum for housing the collection and giving us a space to study it; our teachers, family and friends for coming in and helping out onsite and off it; Bob Brill and George Bass for sharing their articles, expertise and their time; the people of Sanjan

402

K. F. Dalal and R. Mitra-Dalal

and the ‘labour gang’ for making this excavation possible; and finally my friend Alok Kanungo and IIT Gandhinagar for inviting us to share our work with you—Thank you all.

References Bass, G. F., Matthews, S. D., Steffy, J. R., & van Doorninck, Jr.., F. H. (2004). Serçe Limani (Vol I). A&M University Press. Bass, G. F., Brill, R. H., Lledó, B., & Matthews, S. D. (2009). Serçe Limani (Vol. II). A&M University Press. Carboni, S. (2001). Glass from Islamic Lands. Thames and Hudson. Chaisuwan, B., & Naiyawat, R. (2009). Thung Tuk: A settlement linking together the Maritime Trade Route. Trio Creations. Dalal, K. F. (2001). Settlement of Southern Rajasthan (Mewar) during the early Iron age: An artefactual approach [PhD Thesis, Deccan College Post-Graduate & Research Institute]. Dalal, K. F. (2017). Excavations at the medieval site of Chandore, Dist Raigad, Maharashtra (20112015)—A preliminary report. In Md. Nazrul Bari, & H. M. Maheshwaraiah (Eds.), Deccan: Culture, heritage and literature (pp. 175-202). Manak Publications Pvt. Ltd. Dikshit, M. G. (1969). History of Indian glass. University of Bombay. Edulji, H. (1991). Kisseh-I-Sanjan. K.R. Cama Oriental Institute. Gogte, V., Pradhan, S., Dandekar, A., Joshi, S., Kadgaonkar, S., & Bomble, S. (2006). Explorations at Dahanu-Chinchani ancient ports on the western coast of India. Journal of Indian Ocean Studies, 3, 137–145. Gogte, V., Pradhan, S., Dandekar, A., Joshi, S., Nanji, R., Kadgaonkar, S., & Marathe, V. (2006). The ancient port at Chaul. Journal of Indian Ocean Studies, 3, 62–80. Gorin-Rosen Y. (2010). The Islamic glass vessels. In O. Gutfeld (Ed.), Ramla: Final report on the excavations north of the White Mosque (Qedem 51, pp. 213–264). Jerusalem. Gupta, S. P., Dalal, K. F., Dandekar, A., Mitra, R., Nanji, R., & Pande, R. (2002). A preliminary report on the excavations at Sanjan (2002). Puratattva, 32, 182–198. Gupta, S. P., Dalal, K. F., et al. (2002). A report on the excavations at Sanjan (2002). History Today, 3, 99–100. Gupta, S. P., Dalal, K. F., Dandekar, A., Nanji, R., Pande, R., & Mitra, R. (2003). Early medieval indian ocean trade: excavations at Sanjan, India. Circle of Inner Asian Art (SOAS Newsletter), 17, 26–33. Gupta, S. P., Dalal, K. F., Dandekar, A., Nanji, R., Aravazhi, P., & Bomble, S. (2004). On the footsteps of zoroastrian Parsis in India: Excavations at Sanjan on the West Coast-2003. The Journal of Indian Ocean Studies, 1, 93–106. Gupta, S. P., Gupta, S., Garge, T., Pandey, R., Geetali, A., & Gupta, S. (2004). On the fast track of the Periplus: Excavations at Kamrej-2003. Journal of Indian Ocean Studies, 1, 9–33. Gupta, S. P., Dalal, K. F., Nanji, R., Dandekar, A., Bomble, S., Kadgaonkar, S., Mushriff-Tripathi, V., Abbas, R., Choudhari, G., & Sharma, P. (2005). A preliminary report on the 3rd season of excavations at Sanjan (2004). The Journal of Indian Ocean Studies, 2, 55–61. Joglekar, P. P., Deo, S., Deshpande-Mukherjee, A., & Ghate, S. (2002). Archaeological onvestigation at Kelshi District Ratnagiri, Maharashtra. Puratattva, 32, 63–73. Kamerkar, M, & Dhunjisha, S. (2002). From the Iranian Plateau to the Shores of Gujarat. K.R. Cama Oriental Institute. Kanungo, A. K. (2002). Bondo Beads: An ethnoarchaeological approach. South Asian Studies, 18, 121–128. Kanungo, A. K. (2016). Mapping the Indo-Pacific Beads vis-à-vis Papanaidupet. Aryan Books International/International Commission on Glass.

West Asian Glass in Early Medieval India as Seen …

403

Khan F. A. (1960). Banbhore: A preliminary report on the recent archaeological excavations at Banbhore. Government of Pakistan: Department of Archaeology and Museums, Ministry of Education and Information. Kock, J., & Sode, T. (2002). Medieval glass mirrors in Southern Scandinavia and their technique as still practiced in India. Journal of Glass Studies, 44, 79–94. Kröger, J. (1995). Nishapur: Glass of the early Islamic period. Metropolitan Museum of Art. Kumar, A., Rajesh, S. V., Abhayan, G. S., Vinod, V., & Stephen, S. (2013). Indian ocean maritime trade: Evidences from Vizhinjam, South Kerala, India. Journal of Indian Ocean Studies, 9, 195– 201. Mitra, R., & Dalal, K. F. (2005). A report on the glass vessels from Sanjan, 2002. The Journal of Indian Ocean Studies, 2, 62–68. Nakai, I., & Shindo, Y. (2013). Glass trade between the Middle East and Asia. In K. Janssens (Ed.), Modern methods for analysing archaeological and historical glass (pp. 445–457). Wiley. Needell, C. S. (2018). Cirebon: Islamic glass from a 10th century shipwreck in the Java Sea. Journal of Glass Studies, 60, 69–113. Roy, S. (1960). The Khalji Dynasty. In R. C. Majumdar (Ed.) The Delhi Sultanate (Vol. 6, pp. 12–48). Bhartiya Vidya Bhavan. Whitehouse, D. (1968). Excavations at S¯ır¯af: First interim report. Iran, 6, 1–22.

Interrelations in Glass and Glazing Technologies in Mughal Tilework Maninder Singh Gill

Abstract Glazed tiles were extensively used as a means of architectural embellishment in the medieval to pre-modern Islamic world. The Mughals employed them in significant numbers on their buildings over the sixteenth and seventeenth centuries. Laboratory investigations of tile glazes, sourced from a wide range of Mughal buildings spread over northern India, indicate the existence of two distinct compositional varieties. A reconstruction of technologies shows that the two glaze varieties were differently made, one using a plant ash flux, and the other a mineral-soda flux. The glazes were coloured using five distinct pigments or colourants. Findings indicate a close relationship in the technologies of Mughal tile glazes and local or traditional glass manufacture. In the broader context, the paper highlights the potential of archaeometric investigations for the study of ceramic glazes, notably to correlate observations on regional craft practices with technologies that can be evidenced through an assessment of the archaeological record.

1 Introduction The antiquity of use of glazed tiles on architecture pre-dates the advent of the Mughals (CE 1526–1857) in the Indian subcontinent. However, it is under their patronage that this art or craft was most widely employed for building embellishment. The period from the second half of the sixteenth century to the third quarter of the seventeenth century, coinciding with the reigns of the third to sixth Mughal emperors, was particularly marked by heightened tiling activity. Many grand structures were erected at this time, embellished with tilework in a range of colours and patterns, the diversity and richness of which can still be witnessed on architecture from that period that survives to the current date. While Mughal tiling has been a subject of scholarly interest for long, most studies conducted are largely focused on identifying buildings that exhibit tiles on their exteriors and detailing their colours and other visible stylistic features (Akhund & M. S. Gill (B) Art Conservation Solutions, Noida, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_17

405

406

M. S. Gill

Askari, 2011; Degeorge & Porter, 2002; Furnival, 1904; Nath, 1989; Parihar, 1985; Vogel, 1920). Comparatively, less is known of the technologies that have gone into the original production of these tiles, and the technological similarities that they share with related craft industries. This paper seeks to highlight the interrelationship in the production technologies of Mughal tile glazes with that of traditional glass manufacture using science as an investigative tool. A brief history of the development of Mughal tilework is first given, followed by the findings of scientific investigations conducted on glazed tile samples attributed to a series of Mughal buildings.

2 Mughal Tiling: Growth and Development The growth and development of tiling and the tilework industry under the Mughals were largely restricted to the northern part of their empire, centred mainly on the cities of Agra, Delhi, and Lahore that served as their capital at different times. A particular initial impetus was provided under the third Emperor Akbar’s rule (CE 1556–1605), in the second half of the sixteenth century, when increasing numbers of buildings adorned with coloured tiles began to be erected at Agra and Delhi. The tilework of this time, as on Khairul Manzil at Delhi (CE 1561–1562) for instance (Fig. 1), is marked by an overall restraint in surface coverage, and more emphasis laid on maintaining

Fig. 1 Detail of some of the tilework on Khairul Manzil at Delhi illustrating the glaze colours and restrained form of application typical of Mughal tiling of the sixteenth century

Interrelations in Glass and Glazing Technologies …

407

a harmonious correlation with architecture (Nath, 1989: 24–26). The range of glaze colours displayed includes two shades of blue (turquoise and a dark-blue), yellow, green and white, with individual tiles being invariably monochrome, and often but not always laid in mosaic compositions. Similar tilework was also installed during Jahangir’s rule (CE 1605–1627), in the first quarter of the seventeenth century, but a diminishing use is noted in the Delhi–Agra region for this time. More use is instead evident in the Punjab region where several remarkable buildings, such as the Tomb of Ustad at Nakodar (CE 1612), having tiles with the same five colours as at DelhiAgra are found (Fig. 2). The technique of laying tiles in the Punjab region, however, differs from that followed at Delhi–Agra (Parihar, 1985) indicating that there were some regional variances in the craft. The zenith of Mughal tiling can be related to Shah Jahan’s reign (CE 1628– 1658), marked by the extensive employment of a multicoloured tile–mosaic (Vogel, 1920; Parihar, 1985; Rehmani 1997–1998). Elaborate compositions, covering larger expanses on architecture, and featuring monochrome tiles of seven different colours are the hallmark of this period. The expanded range of colours includes the use of purple and orange glazes, in addition to the existing repertoire of five. While most buildings with tilework of this kind are concentrated in Punjab, notably at Lahore, they can also be found at other sites on the old Mughal highway connecting Lahore with Delhi and Agra, the Tomb of Shagird at Nakodar (CE 1657) being one such specimen (Fig. 3). A polychrome variety of tiles was also employed on buildings

Fig. 2 A tiled panel on the façade of Tomb of Ustad at Nakodar in Punjab. Note the use of unglazed bricks to separate individual tiles, a peculiarity of early seventeenth-century tiling in this region

408

M. S. Gill

Fig. 3 A highly intricate tile-mosaic panel on the Tomb of Shagird at Nakodar in Punjab. The use of seven distinct glaze colours, including purple and orange, can be evidenced in the composition depicted here

Interrelations in Glass and Glazing Technologies …

409

during Shah Jahan’s rule, but very few specimens of this kind are known to exist. The Shah Jahani style of tilework of seven colours carried on being employed under Aurangzeb (CE 1658–1707), again mainly in Punjab, but only for a while thereafter. Its use rapidly declined in the third quarter of the seventeenth century before reaching a rather abrupt end, bringing with it the chapter of Mughal tiling to a close.

3 Glaze and Glass Technologies With little information on the functioning of tile workshops forthcoming through archaeological activities, and with the industry itself having ceased to exist for a considerable period of time, the options available for reconstructing the original technologies of Mughal tilework are limited. Detailed field surveys and ethnographic studies on related traditional crafts are undoubtedly beneficial in generating further information, but it is mainly through scientific or archaeometric studies of the material that an enhanced understanding can be gained in such circumstances. Archaeometric approaches typically involve detailed laboratory-based investigations of material remains, the findings of which are then suitably interpreted to reconstruct the various stages involved in the original creation or production of an object or article–in this case glazed tiles. A glazed tile, in its simplest form, consists of a ceramic body with an overlying thin coloured glaze layer. The glaze is nothing but a layer of glass that has passed through a primary stage of production and affixed on the tile body through a firing sequence following some further working. The usual method is to first produce raw glass in a furnace, which is then ground down or converted otherwise to obtain glaze frit or powder. The frit is then coloured and applied on the surface of a separately created tile body, and fused thereon through firing in a kiln (Fig. 4). The basic raw materials of a glaze and the glass from which it was produced are therefore the same. As a corollary, variations in glaze compositions that are determined through analysis are mainly on account of differences in technologies employed for the production of the glass from which the glazes were derived. In the northern part of the subcontinent, different technologies are known to have been employed for pre-modern glass manufacture. In the Gangetic belt, raw or primary glass was being produced traditionally till very recently using only a single ingredient called reh (Dobbs, 1895; Gill, 2017). Reh, a soda-rich alkaline deposit, occurs naturally as an efflorescence on saline soils in the region and contains in itself both the soda and silica that is required for the formation of glass–silica being the main glass-former and soda the fluxing agent used to lower its melting temperature. The melting of reh in a furnace would yield glass, which would then be subject to secondary working for the production of finished glass articles of varying types. The use of efflorescent alkaline deposits for glassmaking was apparently a common feature in much of India, as earthen pots or crucibles purportedly used for this kind of glass production have been recovered from various sites all over the country, many

410

M. S. Gill

Fig. 4 Generalized model illustrating the processes involved in the production of glazed tiles ( adapted from Miller, 2009: 131)

of which date well back to the medieval period and even earlier (Dikshit, 1969: 139–144; Chaudhuri, 1983; Kanungo & Brill, 2009). A different method, albeit of less certain antiquity, was apparently being employed in the Punjab region. Available ethnographic accounts from the nineteenth century inform that glass for making bangles was being manufactured at several centres all over the region at that time, the common method being by melting together equal parts of a quartz-rich stone and locally available carbonate of soda (Baden-Powell, 1872: 235; Hallifax, 1892: 23). A more refined version was apparently being produced at least at one such centre, where blown glass articles such as bowls, glasses and chimneys were also being made in addition to the ubiquitous bangles. Here, raw glass is described as being made in a two-stage process. In the first stage, balls made of roughly equal parts of crushed quartz and soda were first prepared and heated in a furnace or kiln to a semi-fused state and then crushed to attain frit powder. In the second stage, the powdered frit so obtained was mixed with other fluxes, including borax and saltpetre, and subject to firing in earthen crucibles to obtain glass of better quality. Similar technologies are documented as prevailing even in more recent times in parts of Punjab–Pakistan, with little variation in the materials and their respective proportions (Rye & Evans, 1976: 95–96). The soda here is identified as being derived by the burning of Haloxylon recurvum (Haloxylon stocksii), a local desert plant and common crude alkali source that grows in arid parts of Pakistan

Interrelations in Glass and Glazing Technologies …

411

and adjoining western India (Rye & Evans, 1976: 180–185; Rathore et al., 2012). The glass produced in this manner differs compositionally from that made through the reh-based technologies, the distinctions being made evident on the chemical analysis of each. It is worth mentioning here that chemical signature of glass is likely to vary for different geographical locations, even in instances of the use of the same or similar materials and technologies. Some differences are bound to exist since the raw materials employed were not chemically pure, and ancient recipes were not meant to be as precise as the analysis being done in present times. Significant variations, involving the use of different raw materials, would however still be quite apparent.

4 Analytical Methods For glazes, as in glass, most analysis is conducted using instruments that work on the principles of electron microscopy and mass spectrometry. The commonly used instruments and their corresponding techniques are scanning electron microscopy–energy dispersive spectrometry (SEM–EDS), electron probe microanalyser–wavelength dispersive spectrometry (EPMA-WDS) and laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS). The SEM–EDS system is considered advantageous when it comes to glazed ceramics, as the combination of imaging and quantitative analysis that it offers allows an examination of both the ceramic body and glaze layer to the required detail. The working principles of the SEM involve bombarding a sample with a focused beam of electrons, by which low-energy secondary electrons (SE) are knocked out from the sample surface and backscattered electrons (BSE) of higher energy bounced back from deeper within the sample. SE images that are generated reveal the surface topography of the sample being examined, while BSE images provide information on the composition of the area examined (Pollard et al., 2007: 109–113). Chemical analysis is made possible through an EDS detector attached to the SEM, which detects and quantifies X-rays that are also generated on the bombardment of the sample with electrons. The EPMA works broadly on the same principles as the SEM, involving the bombardment of a sample with a focused electron beam to generate both derivative electrons and X-rays, but differs in that it is operated using wavelength dispersive spectrometry (WDS) which enables chemical analysis at high sensitivity. Quantitative analysis in this case is facilitated through crystal spectrometers that are fitted to the instrument. The analytical crystals are able to distinguish and isolate X-rays from individual elements sequentially and direct them to a detector for quantification. The EPMA-WDS has much lower detection limits than the SEM–EDS, of the order c. 0.05 wt % or so, as opposed to c. 0.3 wt % for the latter, and is especially useful for measuring elements present in minor or low concentrations. A technique that is being increasingly applied in glass and glaze analysis is LAICP-MS, which offers several benefits over conventional electron microscopy. In this technique, a tiny part of a sample is first ablated using a pulsed laser, creating a vaporized plume of the sample materials that is then transported to a plasma torch

412

M. S. Gill

for ionization. The ions are then transferred to a mass spectrometer where they are separated based on their mass-to-charge ratio and quantified through a detector. An optical microscope attached to the laser allows a specific area on the sample to be selected for analysis (Pollard et al., 2007: 197–199). The main advantage gained through this technique is trace level detection of elements present, down to parts per million (ppm) or even less, with little or no sample preparation needed. The wide elemental coverage, and major, minor, and trace information provided allow an enhanced level of compositional characterization of the examined material and are particularly useful for provenance studies. In all the above cases, irrespective of the technique used, information on the original technologies is attained by meaningfully interpreting the results of the analysis undertaken. Much of this is made possible through an informed reading of the major and minor element oxide values returned on the quantitative analysis of the glaze or glass as the case may be. For the silica–soda glass type, the prevailing typology in the Indian subcontinent: silica (SiO2 ), soda (Na2 O), potash (K2 O), magnesia (MgO), lime (CaO), alumina (Al2 O3 ) and iron oxide (FeO or Fe2 O3 ) is considered to be the major or base glass-forming oxides, associated with primary production. Their concentrations and presence in the glass or glaze bulk compositions assist in identifying the raw materials originally used. Silica, which is the main ingredient and glass network-former, comprises the bulk of the composition. The source for this is typically either quartz-rich sand or rock. Lime and alumina, which act as network stabilizers, and iron oxide, are incorporated along with the silica, being from the same source. Soda and potash, which act as fluxes and network modifiers, together with magnesia, help determine the nature of the flux employed. Other element oxides that are also detected in minor concentrations are mostly related to the pigments that would have been used for coloration.

5 Analysis The data considered for this paper are extracted from a larger research (Gill, 2015) conducted on a series of fifteenth–seventeenth-century tiled buildings located along the old Mughal imperial highway. Here, the results of investigations undertaken on glaze tiles only from Mughal buildings located at Delhi, Agra and some in the Indian Punjab region from among these are being detailed. A total of 14 buildings are included, all dating to between the second half of the sixteenth century to the third quarter of the seventeenth century. The Delhi buildings include Arab-ki Sarai (AS), Khairul Manzil (KM), Tomb of Atgah Khan (AK), Nila Gumbad (NG), and Tomb of Quli Khan (QK), while those from Agra are Kanch Mahal (KMA), Naubat Khana (NK) in Agra Fort, and Chini-ka Rauza (CR). The Punjab buildings comprise Doraha Sarai (DS), Fatehabad Sarai (FS), Tomb of Ustad (TU), Dakhini Sarai (DS), Sheesh Mahal (SM) and Tomb of Shagird (TS).

Interrelations in Glass and Glazing Technologies …

413

Representative samples from these buildings, mounted as cross sections, were first examined using a JEOL SEM–EDS system, operating at 20 kV, to study distinguishing features and for a preliminary analysis of the glazes. The glaze layers could easily be differentiated from the tile bodies in the BSE images generated, appearing as a distinct bright layer on top. Pigment and mineral particles in the glazes were similarly identifiable by their relatively brighter appearance vis-à-vis the glaze layer. These were subject to spot analysis for their individual characterization, where required. Bulk chemical analysis of the glaze layers was conducted using a JEOL EPMAWDS system operating at 15 kV, current 50 nA and count time 20 s per element. The performance of the instruments employed was checked using Corning Glass Standards and found to be well within acceptable limits. High analytical totals, close to 100 wt% were consistently achieved in the analysis, the glazes for the most part being homogeneous and free of signs of weathering or corrosion. Area scans typically covered clean areas of the glazes, avoiding trapped air bubbles and minor impurities, but including pigment or opacifier particles where present. The quantitative results of the bulk chemical analyses are reported as oxides by stoichiometry, the list of elements1 considered for analysis including silicon (silica, SiO2 ), sodium (soda, Na2 O), potassium (potash, K2 O), magnesium (magnesia, MgO), calcium (lime, CaO), aluminium (alumina, Al2 O3 ), iron (iron oxide, Fe2 O3 ), titanium (titania, TiO2 ), copper (copper oxide, CuO), cobalt (cobalt oxide, CoO), manganese (manganese oxide, MnO), lead (lead oxide, PbO), tin (tin oxide, SnO2 ), nickel (nickel oxide, NiO), zinc (zinc oxide, ZnO), arsenic (arsenic oxide, As2 O5 ), phosphorus (as phosphates, P2 O5 ) and sulphur (as sulphates, SO3 ) (Gill, 2015: 108). Chlorine (as chlorides, Cl) was consistently detected in the bulk analysis but was eliminated in the stoichiometric modelling by the software employed.

6 Results and Discussions Through analysis, all the tiles are found to be the stonepaste type, the underlying bodies being essentially made up of coarse to fine quartz grains bonded together at places with a poor to moderately formed glassy phase. The glaze layers, which lie on top (Fig. 5), can be viewed as divided over a lower interaction zone with quartz particles at the glaze–body interface, followed by an upper core zone. The upper zone is usually clear, with only the occasional mineral grain at places, and at times few randomly dispersed spherical/round bubbles of varying sizes. The yellow, green and orange glazes are different in this regard, all of which have numerous small bright particles distributed uniformly across the glaze layer (Fig. 6). Spot analyses conducted on these particles identify them as the pigment and opacifier lead stannate (Pb[Sn,Si]O3 ), or lead–tin yellow, the lead and tin oxide contents that are recorded only for these glazes thus apparently related to their coloration. Undissolved bright 1

The name and chemical formula of their derivative oxides are given in parentheses against each.

414

M. S. Gill

Fig. 5 SEM microphotograph of the cross section of a tile from Chini-ka Rauza, Agra. The relatively brighter glaze layer on top is clearly distinguishable from the underlying body, which is made up mainly of silica grains of varying sizes. Note the presence of fine silica grains from the body in the lower portion of the glaze

particles in the glazes of the other colours are mostly found to be iron-rich minerals, and occasionally a zircon grain. Rarely are pigment particles used for the coloration of this glaze found present, these apparently having dissolved in the glaze melt on its fusing. Bulk chemical analysis of the glazes reveals that they all are of the silica-soda type. Reduced glaze compositions, calculated by considering only the base glassforming oxides and normalizing their totals to 100 wt%, indicate that silica and soda values are mostly comparable for all the samples. Variations are however apparent in the measured values of the other base glass oxides. Alumina, for example, is detected in significantly high concentrations of c. 7–8 wt% in one group of samples, but is present in half or even a third of these values in the others. Potash, magnesia and lime contents likewise are together consistently higher or lower across a given set of samples from a particular building or region. On plotting the magnesia and alumina values of all the samples, the glazes can be distinguished and divided over two distinct groups (Fig. 7). One group (Group I) consists of glazes that are low in alumina and high in magnesia. The second group (Group II) of glazes conversely are relatively higher in alumina and lower in magnesia. All glaze samples (except for one from NG) from the Delhi buildings of

Interrelations in Glass and Glazing Technologies …

415

Fig. 6 SEM microphotograph detail of the glaze layer of a tile from Dakhini Sarai in Punjab. The bright (white) particles spread across its glaze layer are the colourant lead stannate

the sixteenth and seventeenth centuries (AS, KM, AK, NG and QK), as well as all those from the two Agra buildings ascribed to the first quarter of the seventeenth century (KMA and NK), fall in the Group II category. Nearly all the glazes from the seventeenth-century Punjab buildings (DS, FS, TU, DKS, SM and TS), on the other hand, are of the Group I variety. The Agra CR glaze samples, from the second quarter of the seventeenth century, also belong to the Group I variety. Two glazes each from DS and SM and one each from TS and CR differ, in that they belong to the Group II category, as opposed to being of the Group I type like the majority from these buildings. The outliers generally match the glaze compositions of their ascribed type. The sole Delhi Group I glaze, for instance, broadly matches the characteristics of the Punjab group of this variety. The few Group II Punjab glazes that exist likewise are compositionally quite similar to the Delhi–Agra samples of that type. Variations, where existing, are of a minor nature only, and do not indicate any compositional or technological differences of importance. The outliers probably represent some subsequent repair or restoration work on the buildings where they exist, but their presence could also be suggestive of two or more different tile workshops, following different manufacturing methods, being engaged for the same commission. It is interesting to see that the early Punjab seventeenth century tiles (from DS, FS and TU), although stylistically closer to the Delhi–Agra tiles, are compositionally

416

M. S. Gill

Fig. 7 Scatter plot of alumina versus magnesia contents of the Mughal tile glazes illustrating the existence of two distinct glaze groups, and their chronological and regional spread. ‘*’ indicates reduced composition. ‘Q’ signifies ‘Qtr.’, Q1 implying 1st Qtr. and so on

similar to specimens from the same region where they belong. The only anomaly to the regional grouping of glaze types is the CR tiles, from Agra, which are compositionally of the Punjab typology. This is not too surprising, however, considering that the tilework on this building stylistically matches that found in the Punjab region, having the same kind of intricate compositions with seven glaze colours. In this case, it would be logical to assume that the work on this particular building was carried out by a team of skilled artisans sourced from the Punjab region, most likely Lahore, who executed the contract using technologies that they were most familiar with, possibly also carrying the required raw materials along with them to the new locale.

6.1 Plant Ash Glass and Glazes Reduced compositions of the samples compiled building-wise for the two groups (Tables 1 and 2), excluding the outliers, show that soda contents in the Group I glazes typically lie in the range of 16–17 wt%, while silica for them varies mostly over 67–71 wt%. Lime, potash and magnesia values are quite consistent for this group, averaging about 4 wt% in the case of the first two, and around 3 wt% for magnesia.

Doraha Sarai (DS)

Fatehabad Sarai (FS)

Tomb of Ustad (TU)

Dakhini Sarai (DKS)

Sheesh Mahal (SM)

Chini-ka Rauza (CR)

Tomb of Shagird (TS)

1

2

3

4

5

6

7

Punjab

Agra

Punjab

Punjab

Punjab

Punjab

Punjab

Region

1657 CE

c. 1639 CE

c. 1634 CE

17th cent. (Q2)

1612 CE

c. 1606 CE

17th cent. (Q1)

Date/Period

68.9 69.1

Average

71.7

72.7

70.4

66.8

65.7

67.2

SiO2

5

8

8

4

4

2

8

Nos. of samples

All results are in wt% from EPMA-WDS analyses and normalized to 100%. Outliers to the groupings in individual buildings are excluded in the calculations

Building

No.

16.8

17.5

16.0

15.4

17.3

15.9

17.5

17.9

Na2 O

3.9

3.6

3.3

3.3

3.9

4.2

4.4

4.2

CaO

3.8

4.0

3.5

3.7

3.2

4.9

3.9

3.3

K2 O

3.0

2.8

2.9

2.4

2.7

3.2

3.5

3.2

MgO

2.5

2.2

1.8

1.7

1.8

3.5

3.4

3.1

Al2 O3

Table 1 Average chemical compositions of the plant ash (Group I) variety of Mughal glazes, reduced to their base glass-forming oxides

1.0

1.0

0.8

0.8

0.7

1.5

1.4

1.1

FeO

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

Total

Interrelations in Glass and Glazing Technologies … 417

Arab-ki Sarai (AS)

Khairul Manzil (KM)

Tomb of Atgah Khan (AK)

Nila Gumbad (NG)

Tomb of Quli Khan (QK)

Kanch Mahal (KMA)

Naubat Khana (NK)

1

2

3

4

5

6

7

Agra

Agra

Delhi

Delhi

Delhi

Delhi

Delhi

Region

17th cent. (Q1)

17th cent. (Q1)

17th cent. (Q1)

17th cent. (Q1)

1566-1567 CE

1561-1562 CE

c. 1560 CE

Date/Period

66.8 66.2

Average

66.1

64.6

67.5

66.0

64.8

67.6

SiO2

2

3

4

9

5

4

4

Nos. of samples

All results are in wt% from EPMA-WDS analyses and normalized to 100%. Outliers to the groupings in individual buildings are excluded in the calculations

Building

No.

18.4

17.2

18.9

19.3

18.8

18.9

18.8

17.2

Na2 O

2.3

1.7

3.2

2.3

2.2

1.7

2.4

2.6

CaO

2.2

1.8

1.9

2.6

2.2

2.3

2.6

2.0

K2 O

1.2

1.1

1.9

1.1

1.2

0.9

0.9

1.4

MgO

7.7

9.1

6.4

8.2

6.4

8.3

8.2

7.6

Al2 O3

1.9

2.3

1.7

2.0

1.6

1.9

2.2

1.5

FeO

Table 2 Average chemical compositions of the mineral-soda (Group II) variety of Mughal glazes, reduced to their base glass-forming oxides

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

Total

418 M. S. Gill

Interrelations in Glass and Glazing Technologies …

419

Alumina varies more comparatively, being higher in the earlier seventeenth century (Q1) samples, mostly in excess of 3 wt%, but lower, averaging 2 wt% or so, in those of a later date (from Q2 and Q3). Iron oxide concentrations that vary over 0.5 to 1.5 wt% correlate positively with alumina, the recorded values being higher in the seventeenth century Q1 samples, and lower for the later ones that are also low in alumina. The measured potash and magnesia contents of these glazes, typically in excess of 2.5 wt% for each, are characteristic of the use of a plant ash alkali flux (Sayre & Smith, 1961; Tite et al., 2006). Given that this glaze group is associated with tilework from the Punjab region (CR being an exception as stated earlier), this correlates well with the technologies known to be employed for traditional glass manufacture locally. The frit for the Group I glazes is therefore most likely to have been prepared on the same broad lines as described by Hallifax (1892: 23) and Rye and Evans (1976: 95–96), using crushed quartz and plant ash soda to produce glass balls in a furnace, which were then crushed and milled to attain glaze powder (Gill, 2015: 282). The secondary stage of refinement involving the use of borax does not seem to have been followed through, since no trace of this is evident on the analysis. The possibility that the glass or glass frit was originally manufactured in a different manner, as done in medieval workshops in Iran for instance—by melting the same ingredients together in a crucible and pouring the resultant molten glass in water to form frit (Allan, 1973)—cannot however be ruled out completely. The raw materials in any case would have been locally procured, the silica or quartz source being the same as that used for the tile bodies, while the plant ash would have been derived from Haloxylon recurvum. The likely use of Haloxylon recurvum is corroborated to an extent through the analysis, the soda to potash ratios (Na2 O/K2 O) and normalized lime-plus-magnesia (CaO + MgO/Na2 O + K2 O) contents of these glazes matching values accorded to plant ash for this particular plant species (Tite et al., 2006).

6.2 Mineral-Soda Glass and Glazes Soda is marginally higher in the Group II glazes, mostly from 18–19 wt%, and silica correspondingly slightly lower from 65–68 wt%. Lime and potash contents are consistently lower here than in Group I, averaging 2 wt% or so for both. Magnesia is markedly low, often 1 wt% or lower, but is also in slightly higher concentrations up to 2 wt% in a few specimens. Alumina, which is a more reliable discriminator, is consistently present in characteristic high concentrations of 7–8 wt% here. Iron oxide values are also relatively higher in this group, averaging close to 2 wt%, varying parallelly with alumina as for the (Group I) plant ash glazes. These (Group II) glazes are strikingly similar to a wide range of early to premodern glasses of Indian origin in their composition. In particular, the alumina and magnesia values of these glazes correlate with an Indian variety of high-alumina mineral-soda glass, considered to be made using reh-based or similar technologies (Brill, 1987; Dussubieux et al., 2010; Lankton & Dussubieux, 2006). The Group II

420

M. S. Gill

glazes are therefore likely to have been similarly made, by the melting of just crude reh, or a similar alkaline efflorescent, in a furnace. This is supported by the fact that the Delhi–Agra region to which this glaze type is particular lies on the western periphery of the Gangetic plains, where reh deposits are known to be plentiful, and where glass was being manufactured on an industrial scale using the very same ingredient till recent times. The method of production prevailing in the nineteenth and twentieth centuries in this region, of melting reh in large tank furnaces (Dobbs, 1895; Sode & Kock, 2001), may not have been followed in Mughal times though, as no evidence of the existence of such furnaces from that time has yet come to light. It is more likely that the reh was melted in glassmaking crucibles, the likes of which have been discovered in excavations at various medieval/late-medieval sites elsewhere in India as mentioned earlier. The glass in this case is likely to have been produced independently and then supplied to tile-making workshops where it was powdered to frit and put to use. This would certainly have been more economical given that the requirement for raw glass could not have been limited to just the tilework industry, a significant demand for bangles, beads and other glass articles with a larger consumption base coexisting. The existence of a separate production unit for raw glass is less certain in the case of the Punjab workshops, ethnographic studies and historical accounts suggesting that these were probably more like self-contained units (Baden-Powell, 1872: 220–227; Rye & Evans, 1976: 89–107), with all activities, including glassmaking and fritting being conducted in-house.

6.3 Colourants The colourants are similar across the two groups for the common glaze colours, the measured values of each indicating the proportion by which they would have been added to the glaze frit (Table 3). The turquoise glazes in both groups all contain copper oxide (CuO) in concentrations of 2–5 wt%, while all the dark-blue glazes consistently have small amounts of cobalt oxide (CoO), typically around 0.4 wt%, in their compositions. Both these pigments are mostly in a dissolved state in their respective glaze layers, although an occasional cobalt-rich particle is detected in some of the dark-blue glazes at times. The yellow glazes, as stated earlier, are coloured by lead stannate (Pb[Sn,Si]O3 ), the lead and tin oxides contents in these glazes together adding up to between 15–20 wt% in most instances. The green glazes have similar quantities of lead stannate as in the yellow glazes, but also have 1–3 wt% of copper oxide in them—the lead stannate-yellow with copper-blue resulting in green here. No colourant is identifiable in the white glazes, the colour apparently attained from the stonepaste body below, the presence of a basal layer of quartz particles at the glazebody interface rendering the otherwise colourless glaze into an opaque white. The purple glazes, which are particular only to Group I, are coloured by manganese oxide (MnO), present in concentrations of around 1 wt%. The orange glazes, which are also specific to Group I only, are coloured as in the yellow glazes by just lead stannate.

Interrelations in Glass and Glazing Technologies …

421

Table 3 Average values of the metal oxide colourants extracted from the glaze bulk chemical compositions, given against their corresponding glaze colours. Glaze colour

CuO

CoO

MnO

PbO

SnO2

ZnO

Turquoise

3.0

−

−

−

−

−

Dark-Blue

−

0.4

−

−

−

−

Yellow

−

−

−

15.2

2.6

0.3

Green

2.0

−

−

13.2

2.5

0.3

Orange

−

−

−

20.6

3.4

1.9

Purple

−

−

1.2

−

−

−

White

−

−

−

−

−

−

All results are in wt% from EPMA-WDS analyses. ‘-’ indicates ‘below detection limit’ No colourant or opacifier could be detected in the white glazes The purple and orange glazes are specific to the plant ash type of glazes only ZnO is in association with lead stannate, and not a colourant by itself here

The pigment here, however, is an unusual zinc-containing variant (Pb[Sn,Zn]O3 ), given the name lead–tin orange (Gill & Rehren, 2014), which imparts an orange colour to the glaze instead of an anticipated yellow. The glaze pigments in general, with two notable exceptions, conform to those used for the coloration of glass objects. The oxides of manganese, cobalt and copper are known to have been used to colour beads, bangles and miscellaneous glass articles from ancient to pre-modern times (Rehren & Freestone, 2015), including in the Indian context (Choudhury, 1970). Lead stannate similarly, at times in combination with other metal oxide colourants, has been used as a pigment and opacifier in glass for the last two millennia or more, over most of the civilized world (Matin, 2019; Tite et al., 2015). The two colourants, or coloration techniques, that are peculiar to the Mughal glazes are related to the white and orange glaze colours. Opaque white in glass, and also for glazed ceramics or tiles elsewhere in the world, is known to have been typically attained by adding either tin or antimony oxide as an opacifying agent. Exploiting the presence of a silica-rich body to effectively obtain a white coloured glaze, and eliminating the requirement of an added opacifier, as in the case of the Mughal tiles, appears to be a local characteristic and specific to tiles only. No such use elsewhere, other than on tilework at some other locations in the subcontinent has yet come to light. The existence of a zinc-containing variant of lead stannate that imparts an orange colour to glazes is again a peculiarity that seems to be limited to tiles in the subcontinent, as no parallel or prior use in glass has yet been reported anywhere. The possibility of its use beforehand in the glass industry in India cannot however be discounted completely.

422

M. S. Gill

7 Conclusion Two distinct technologies were followed for the manufacture of Mughal tile glazes (in northern India–Pakistan) over the sixteenth and seventeenth centuries, one being particular to the tilework in the Delhi–Agra region, and the other to that in the Punjab region. The compositional makeup of the Delhi–Agra tile glazes closely resembles that of locally manufactured so-called Indian Glass, the characteristic highalumina mineral-soda signature that they both carry indicating they were similarly made, using reh or an equivalent alkaline efflorescent as the principal ingredient. The compositional characteristics of the Punjab type of tile glazes on the other hand, notably their magnesia and potash values, indicate that they were made using a plant ash flux, derived from a specific desert plant species, as followed in local traditional practice. A close relationship in the technologies of Mughal tile glazes and local glass manufacture in both the regions is therefore evident, signifying a fair degree of interdependence, and interrelationship, between the two industries. Similarities are also evidenced in the colourants employed, the oxides of copper, cobalt and manganese, and lead stannate finding use in both Mughal tile glazes and contemporaneous glass objects. The techniques of coloration of the orange and white tile glazes, however, have no reported earlier use in glass, and it remains to be seen if this was a development that was confined to the tilework industry or was prevalent in the glass industry as well. Further research on colourants in Indian glass is anticipated to shed light on this matter. Acknowledgements I am grateful to the authorities of IIT Gandhinagar, and especially Dr. Alok Kumar Kanungo, for providing me the opportunity to present my work at the Conference cum Workshop on History, Science and Technology of Ancient Indian Glass held at IIT-Gandhinagar from 21 to 25 Jan 2019.

References Akhund, A. H. & Askari, N. (2011). In Tale of the tile: The ceramic traditions of Pakistan. Karachi: Mohatta Palace Museum. Allan, J. W. (1973). Abu’l Qasim’s treatise on ceramics. Iran, 11, 111–120. Baden-Powell, B. H. (1872). In Hand-book of the manufactures and arts of the Punjab Vol II. Lahore: Punjab Printing Company. Brill, R. H. (1987). Chemical analyses of some early Indian glasses. In H. C. Bhardwaj (Ed.), Archaeometry of glass: Proceedings of the archaeometry session of the XIV International Congress on Glass 1986 New Delhi India (pp. 1–25). Calcutta: Indian Ceramic Society. Chaudhuri, M. (1983). The technique of glass making in India (1400–1800 A.D.). Indian Journal of History of Science, 18(2), 206–219. Choudhury, M. (1970). The techniques of colouring glass and ceramic materials in ancient and mediaeval India. Indian Journal of History of Science, 5(2), 271–280. Degeorge, G., & Porter, Y. (2002). The art of the islamic tile. Flammarion. Dikshit, M. G. (1969). History of Indian glass. University of Bombay.

Interrelations in Glass and Glazing Technologies …

423

Dobbs, H. R. C. (1895). A monograph on the pottery and glass industries of the North-Western Provinces and Oudh. North-Western Provinces and Oudh Government Press. Dussubieux, L., Gratuze, B., & Blet-Lemarquand, M. (2010). Mineral soda alumina Glass: Occurrence and meaning. Journal of Archaeological Science, 37, 1646–1655. Furnival, W. J. (1904). In Leadless decorative tiles, faience, and mosaic. Staffordshire. Gill, M. S. (2015). Glazed tiles from Lodhi and Mughal Northern India: A technological appraisal. unpublished doctoral thesis, University College London (UCL). Gill, M. S. (2017). A single ingredient for primary glass production: Reassessing traditional glass manufacture in Northern India. Journal of Glass Studies, 59, 249–259. Gill, M. S., & Rehren, Th. (2014). The intentional use of lead-tin orange in Indian Islamic glazes and its preliminary characterization. Archaeometry, 56(6), 1009–1023. Hallifax, C. J. (1892). Monograph on the pottery and glass industries of the Punjab 1890–91. The Civil and Military Gazette Press. Kanungo, A. K., & Brill, R. H. (2009). Kopia, India’s first glassmaking site: Dating and chemical analyses. Journal of Glass Studies, 51, 11—25 Lankton, J. W., & Dussubieux, L. (2006). Early glass in Asian maritime trade: A review and an interpretation of compositional analyses. Journal of Glass Studies, 48, 121–144. Matin, M. (2019). Tin-based opacifiers in archaeological glass and ceramic glazes: A review and new perspectives. Archaeological and Anthropological Sciences, 11(4), 1155–1167. Miller, H. M. (2009). Archaeological approaches to technology. Left Coast Press Inc. Nath, R. (1989). In Colour decoration in Mughal Architecture (India and Pakistan). Jaipur: The Historical Research Documentation Programme. Parihar, S. (1985). Mughal monuments in the Punjab and Haryana. Inter-India Publications. Pollard, M., Batt, C., Stern, B., & Young, S. M. M. (2007). Analytical chemistry in archaeology. Cambridge University Press. Rathore, V. S., Singh, J. P., & Roy, M. M. (2012). Haloxylon Stocksii (Boiss.) Benth. et Hook. f., A promising Halophyte: distribution, cultivation and utilization. Genetic Resources and Crop Evolution, 59, 1213–1221 Rehmani, A. (1997–1998). The Persian glazed tile revetment of mughal buildings in Lahore. Lahore Museum Bulletin, 10–11, 74–98. Rehren, Th., & Freestone, I. C. (2015). Ancient glass: From Kaleidoscope to crystal ball. Journal of Archaeological Science, 56, 233–241. Rye, O. S., & Evans, C. (1976). In Traditional pottery techniques of Pakistan: Field and laboratory studies. Washington D.C.: Smithsonian Institution Press. Sayre, E. V., & Smith, R. W. (1961). Compositional categories of ancient glass. Science, 133, 1824–1826. Sode, T., & Kock, J. (2001). Traditional raw glass production in Northern India: The final stage of an ancient technology. Journal of Glass Studies, 43, 155–169. Tite, M., Watson, O., Pradell, T., Matin, M., Molina, G., Domoney, K., & Bouquillon, A. (2015). Revisiting the beginnings of tin-opacified Islamic glazes. Journal of Archaeological Science, 57, 80–91. Tite, M. S., Shortland, A., Maniatis, Y., Kavoussanaki, D., & Harris, S. A. (2006). The composition of the soda-rich and mixed Alkali Plant Ashes used in the production of glass. Journal of Archaeological Science, 33, 1284–1292. Vogel, J. P. (1920). Tile-mosaics of the Lahore Fort (Vol. 41). Calcutta: Archaeological Survey of India. New Imperial Series.

The Diffusion of South Asian Glass

Indian Glass Beads in Western and North Europe in Early Middle Age Bernard Gratuze, Constantin Pion, and Torben Sode

Abstract In recent years, chemical analyses of glass beads excavated from late Antique and Early Middle Age sites in western and north-western Europe (France, Belgium, Switzerland, Denmark, Germany and Sweden) have revealed for the first time the presence of two groups of glass beads with unexpected compositions for these periods and geographic areas. The first group is composed of tiny (less than 1.5 mm in diameter) glass beads recovered from Merovingian (mid-fifth/sixth century CE) graves located in Western Europe. Although their chromatic spectrum is varied: green (more than 6300 examples), orange with a red core (233 examples), black (184 examples), yellow (18 examples), slightly translucent ‘milky’ white (5 examples) and red (3 examples), we observe a predominance of green beads (93%) which may express a preference for this colour. Up to now, only beads recovered in France, Belgium and Switzerland have been analysed. But according to their particular typology, these beads have also been identified in The Netherlands, Germany and Spain. Their glass has a high-alumina soda composition which was proven to be identical with the one identified by Dussubieux for small drawn ‘Indo-Pacific’ beads produced in southern India and Sri Lanka from the fourth century BCE through the eleventh century CE (composition referenced as m-Na-Al 1, Dussubieux et al. in J Archaeol Sci 37:1646–1655, 2010). The second group is composed of large opaque red or orange barrel-shaped beads (c. 10–12 mm in diameter) found in early seventh century graves in north-western Europe. The beads studied here originate mainly from Denmark (Ribe and Sandegård—Bornholm island), Sweden (Helgö and

B. Gratuze (B) IRAMAT/CEB, National Centre for Scientific Research, Orléans, France e-mail: [email protected] C. Pion Belgian Royal Institute for Cultural Heritage, Brussels, Belgium Free University, Jette, Belgium e-mail: [email protected] T. Sode Glass Bead Trading, Copenhagen, Denmark e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_18

427

428

B. Gratuze et al.

several sites on Gotland island) and Germany (Frankfurt Harheim). Most of the analysed beads from this period were made from recycled Roman glass. However, some of them, only made with red and orange glasses, show a composition which differs from all known Western European glasses in both main components composition and colouring recipe. Their trace element pattern, although not identical, shows several similarities with those of South Asian glasses.

1 Introduction 1.1 Glass Compositions Encountered in Europe During the Early Middle Ages In Western Europe, most glass objects that circulated during the Antiquity and the Early Middle Ages (first century BCE–ninth century CE) were made from a natron soda-lime glass produced in the eastern part of the Mediterranean basin. This glass was manufactured from a calcareous sand with little aluminium content, typically from 2 to 3%, and rarely more than 4% (Rehren et al., 2015; Schibille, 2011) and natron, a relatively pure hydrated sodium carbonate (Shortland et al., 2006) that was probably collected in dry Egyptian lakes of the Wadi Natrum region. Several centres of production of raw glass (“called primary workshops”) located between Lebanon and Egypt were probably active during this period (Freestone et al., 2000, 2008; Rehren et al., 2010). In Western Europe, the production centres of vessels and beads (“secondary workshops”) used blocks of raw glass from these primary workshops or recycled glass (cullet, tesserae). The various archaeometric studies carried out on ancient and mediaeval glass objects have enabled the identification of several compositional subgroups (Brill, 1999; Foy et al., 2003; Freestone et al., 2000; Rehren & Freestone, 2015; Velde, 2013; Velde & Motteau, 2013). These subgroups are probably related to the geographic origin of the primary workshops or to particular combinations of elements. Overall, the material is nonetheless characterized by a relative homogeneity in its composition (Foy & Picon, 2005; Foy et al., 2003; Picon & Vichy, 2003). Within this period, a second type of soda-lime glass, fabricated from a soda extracted from the ashes of halophyte plants such as Salicornia sp. or Salsola sp., was also available. This glass, not as common as the previous one, has higher contents of potash (K2 O), magnesia (MgO) and phosphoric oxide (P2 O5 ). Two families of plant ash soda glass were identified. The first one, found mainly among mosaic glasses, is characterized by similar contents of magnesia and potash. It is frequently found among red and green-emerald glasses of the Antiquity (Jackson & Cottam, 2015; Jackson et al., 2015; Nenna & Gratuze, 2009). The second type is characterized by magnesia levels that exceed those of potash. During the classical period and the Early Middle Ages, these glasses were probably produced inland in Mesopotamia. The studies carried out on Sassanian soda-lime plant ash glasses from Veh Ardasir (Iraq)

Indian Glass Beads in Western and North Europe …

429

provide evidence for the existence of subgroups that are distinguished by different magnesia/potash ratios (Ganio et al., 2013; Mirti et al., 2009). Starting from the end of the eighth century, soda plant ash glasses progressively replaced natron glasses in the Mediterranean zone. During the same period, potash-lime glasses made using the ashes of forest plants appeared in Western Europe (Gratuze et al., 2021; Velde, 2013; Velde & Motteau, 2013; Wedepohl & Simon, 2010). Most Early Middle Ages glass objects were typically produced from different natron soda-lime glasses that originated from the Levantine and Egyptian coastal zones. The late antique and Early Middle Ages glasses differ however from those of the earliest periods in terms of their colouring and opacifying techniques. Starting in the fourth and fifth centuries and spreading from the east towards the western Mediterranean, tin became increasingly important as the principal colouring and opacifying agent for yellow glass, in the form of lead stannate that replaced lead antimonate, and for white glass, in the form of tin oxide that replaced calcium antimonate (Tite et al., 2008). More recent work by Marco Verità on the yellow tesserae of the Gorga collection (Italy) shows however that opacification with tin yellow may have started as early as the second century CE (Verità et al., 2013). This development did not affect the colour of white glass, but in yellow glass, the dominant yellow–orange created by lead antimonate changed to a ‘lemon’ yellow hue.

1.2 Early Middle Ages Glass Beads in Western Europe From a technological and typological point of view, four main types of glass beads are encountered in Western Europe during the Early Middle Ages. One of them consists of wound beads and the three others of drawn beads. Unlike wound beads, fabricated by wrapping a mass or gather of molten glass around a rotating metal rod (called a mandrel), drawn beads are obtained by segmentation of thin glass tubes. As no trace of fabrication or use of drawn tubes have ever been found in Western Europe, whether at bead-making or at glassmaking workshops, drawn beads are generally considered as imports. The fabrication of drawn beads can be carried out in several ways that are difficult to identify from the finished object. One method consists of gathering a mass of molten glass on the end of a hollow metal rod (a pontil or blowpipe) and to imprison an air bubble within it, either by blowing or by inserting a metal rod. The method of introducing a metal rod is characteristic of beads of the ‘Indo-Pacific’ type (Francis, 1990, 2002, 2004; Kanungo, 2004; Stern, 1987). Using a tool, so-called lada, the glass is drawn into a tube of the desired diameter and size. These glass tubes are then cut into beads. Judging from the shape of the edges of the beads found in Western Europe during Early Middle Ages, there are two principal methods: hot and cold cutting. The four main types of glass beads can be defined as follows: • Type 1, wound glass beads with various sizes, shapes and decorations (Figs. 1 and 2). They are found with the highest frequency and have usually a large size

430

B. Gratuze et al.

Fig. 1 Examples of simple wound Merovingian glass beads found in Europe, type1. Photo C. Pion

Fig. 2 Examples of types of decorated wound Merovingian glass beads found in Europe, type 1. Photo C. Pion

(diameter from a few mm up to several cm). The large opaque red or orange barrelshaped beads (c. 10–12 mm diameter) found in early seventh-century graves in north-western Europe belong to that type. • Type 2, drawn beads obtained by pre-individualization of the beads on the hot tube, which is then cut after cooling (in one or several segments). This type constitutes the second most common type of beads among European Early Middle Ages assemblages (Fig. 3). In practice, a fine metal rod is pushed into the tube to enable its manipulation and avoid deformation or closing of the hole once softened. Such an iron rod within a tube of drawn glass was discovered at the Egyptian site of Kôm el-Dikka (fifth–sixth centuries CE) at Alexandria (Arveiller-Dulong

Indian Glass Beads in Western and North Europe …

431

Fig. 3 Principal shapes of drawn glass beads with umbilical-shaped edges indicative of segmentation. a cylindrical ‘perfect’, b cylindrical ‘rounded’, c tubular undulating or not, d fusiform, e baluster, f ninepin, type 2. Photo C. Pion

& Nenna, 2011). The segmentation points can be made individually with a blunt instrument such as a blade, or collectively by rolling the tube on a mould with a crenellated surface. This second procedure is known to commence in the Roman period and allows the uniform segmentation of the tubes at regular intervals with one action. These beads have a smaller diameter than wound beads (diameter below 3.5 mm), and their edges present an umbilical shape, rarely rounded unless they have received a cold finishing (polishing) or a heat treatment (refiring). These beads are generally considered to be imports from the eastern part of the Mediterranean Sea (Callmer, 1977; Spear, 2001; Francis, 2002 and 2004; Greiff & Nallbani, 2008; Arveiller-Dulong & Nenna, 2011). • Type 3, drawn beads obtained by chopping the tubes directly after cooling, without reheating for rounding the edges (Fig. 4). They are characterized by sharp edges and larger size than the preceding type (diameter above 3.5 mm). Although they are less common, they may be found in high numbers at a given site. These beads are also considered to be imports. • Type 4, drawn beads obtained by chopping the tubes directly after cooling. They are characterized by very small size (diameter lower than 2.5 mm) and soft edges (Figs. 4 and 5). They are the less common type, but they are sometimes found in high numbers at a given site. Cold cutting, followed by heat treatment, characterizes beads of the ‘Indo-Pacific’ type. On the Indian subcontinent, the craftsman places a series of tubes (about a dozen) side by side on the cutting edge of a blade fixed on the ground and then cuts these tubes into small cylindrical segments by means of a second blade. This method enables a rapid and quasi-industrial production of beads with edges that display a more or less regular clean cut. These segments are generally not used as it is. The majority of beads show a more or less pronounced degree of roundness as a result from heat treatment. This treatment can be carried out by placing thousands of beads in a ceramic container,

432

B. Gratuze et al.

Fig. 4 Examples of drawn glass beads with edges cut when cold, with or without heat finishing (types 3 and 4). Photo C. Pion

mixed with dung and ashes on a fire or in an oven for several minutes and the content stirred regularly. The mixing of cow dung and ash is done intentionally to avoid sticking of beads or collapsing of holes in the process. Once cooled, the beads are cleaned and sometimes polished in a mortar. Refiring gives the beads a more rounded shape, makes them shinier and causes the striations on the surface to partly disappear. The work of many scholars has revealed that southern Asia was the most important centre of production of ‘Indo-Pacific’ beads for nearly

Indian Glass Beads in Western and North Europe …

433

Fig. 5 Glass beads of ‘Indo-Pacific’ type from India or Sri Lanka. Blanzac-Porcheresse ‘le Molle’ (Charente), grave 931, type 4. Photo C. Pion

two millennia (Basa, 1991; Bellina, 2003; Bellina & Glover, 2004; Dussubieux, 2001; Dussubieux & Gratuze, 2004, 2013; Dussubieux et al., 2010, 2012; Francis, 2002; Glover & Henderson, 1995; Gratuze et al., 2000; Lankton & Dussubieux, 2006; Lankton & Lee, 2006). In addition to the main type described above (wound and drawn glass beads), we also found mosaic glass beads, as well as folded glass and perforated glass beads (Adams, 2005; Pion, 2014).

2 Method of Analysis More than a thousand glass beads from about forty necropolises and sites situated in Belgium, France, Switzerland, Denmark, Sweden and Germany are studied. All the objects were analysed by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) at the Ernest-Babelon Centre (IRAMAT, UMR 5060 CNRS/Université d’Orléans). No preparation of the sample is needed, and this method is particularly well adapted to composite objects and to very small objects like beads (Gratuze, 2016). For analysis, the beads were placed inside an ablation cell. A microsample, invisible to the naked eye, was taken by a laser beam. The material sampled (a few micrograms) was carried to a plasma torch by a gaseous flow of argon or argon and helium. The high temperature of the plasma (8000 °C) dissociates and ionizes the material, of which

434

B. Gratuze et al.

the different constituents are identified according to their mass-on-charge ratio. An electronic detector enables their quantification. The instrumentation utilized consists of a Thermo Fisher Scientific Element XR mass spectrometer, combined with a VG elemental Nd:YAG pulsed laser beam operating at 266 nm (for analyses carried out before October 2014) and a Resonetics RESOlution M50e laser probe ablation device operating at 193 nm (for analyses carried out after October 2014).

3 Small Drawn Merovingian Glass Beads Exhibiting Unusual High-Alumina Composition 3.1 Studied Corpus and Archaeological Contexts The term ‘Merovingian’ designates a dynasty of kings, which reigned from the fifth to the middle of the eighth century CE on a territory corresponding more or less to Roman Gaul, located in the Western Empire. In its greatest extension, this kingdom covered the current European territories of France, Belgium, Luxembourg, the southern part of the Netherlands, south-west Germany as well as Switzerland. The first Merovingian king was Clovis I, who reached power around CE 482. The last Merovingian king was Childeric III, who was superseded in 751 by Pippin the Short, the founder of the Carolingian dynasty. During the Merovingian period, people were buried with more or less abundant and varied grave goods. Among them, glass beads are one of the most original and emblematic artisanal products, found in the hundreds and thousands at funerary sites in Early Mediaeval Gaul. In spite of their apparent profusion in graves, few production sites of glass beads have so far been found within the boundaries of the Merovingian territory. Hence, there is a distinct gap in our understanding of the technology and the mechanisms of supply of this type of material. Our recent work allowed the identification of a large group of small drawn beads within fifth to sixth-century funerary sites (Pion, 2014; Pion & Gratuze, 2016). These small beads appear to be morphologically and typologically identical to Indo-Pacific beads produced on the Indian subcontinent (type 4). If we exclude their size, the morphologies of drawn glass beads with edges indicative of segmentation (type 2) and those of Indo-Pacific type (type 4) look fairly similar. It was thus interesting to explore whether the thousands of beads found in the Merovingian graves (mid-fifth/sixth century CE) came from southern Asia as suggested by their fabrication technique. This hypothesis was first put forward following an archaeometric study of fourteen glass beads from a grave in the French necropolis of Saint-Laurent-des-Hommes ‘Belou Nord’ (Dordogne) (Poulain et al., 2013).

Indian Glass Beads in Western and North Europe …

435

3.2 Analytical Results According to the principal constituents of the sand source (CaO, Al2 O3 ) and the modifiers (MgO, K2 O, P2 O5 ), the results obtained (Table 1) enable the classification of the beads into four main groups (Fig. 6). Three of the groups (groups 1 to 3) are soda-lime glasses, elaborated either with natron or soda plant ashes, commonly encountered in the western world, as described above. Group 4 represents an unusual composition for European glass, characterized by high contents in alumina, potash and several trace elements. Group 1: Natron glass and related subgroups. Most of the beads analysed here are made with natron soda-lime glass, as are most of the glass objects of the Merovingian period. This glass is characterized by low contents of potash and magnesia (4%), low lime contents ( Aufstieg und Niedergang der Roemischen Welt, 2(9.2), 604–1076. Rehren, T. (2021). The origin of glass and the first glass industries. In A. K. Kanungo & L. Dussubieux (Eds.), Ancient glass of south asia—archaeology, ethnography and global connection. Springer Nature/IIT Gandhinagar. Salles, J.-F. (1993). The Periplus of the Erythraean Sea and the Arab-Persian Gulf. Topoi, 3(2), 493–523. Sankalia, H. D., & Dikshit, M. G. (1952). In Excavations at Brahmagiri (Kolhapur) 1945–46. Deccan College Monograph Series No. 5. Sankalia, H. D., & Deo, S. B. (1955). In Report on excavations at Nasik and Jorwe 1950–51. Deccan College Monograph Series No. 13. Sankalia, H. D., Deo, S. B., Ansari, Z. D., & Ehrhardt, S. (1960). In From history to pre-history at Nevasa 1954–56. Deccan College Post-Graduate & Research Institute. Schenk, H., & H-J Weisshaar. (2016). The Citadel of Tissamaharama: Urban Habitat and Commercial Interrelations. In M.-F. Boussac, J.-F. Salles, & J.-B. Yon (Eds.), in ports of the ancient Indian Ocean (pp. 459–479). Primus Books. Schoff, W. H. (Ed.). (1912). The Periplus of the Erythraean Sea: Travel and trade in the Indian Ocean by a merchant of the first century. Longmans, Green, & Co. Sen, S. N., & Chaudhuri, M. (1985). In Ancient glass and India. Indian National Science Academy. Sharma, A. K. (1994). Manipur, the glorious past. Aryan Books International. Shortland, A., Kirk, S., Eremin, K., Degryse, P., & Walton, M. (2018). The analysis of late bronze age glass from Nuzi and the question of the origin of glass-making. Archaeometry, 60(4), 764–783.

488

S. Gupta

Sidebotham, S. E. (1986). In Roman economic policy in the Erythra Thalassa (31 B.C.—217 A.D.). Leiden: E.J. Brill. Simpson, S. J. (2014). Sasanian glass: An overview, In D. Keller, J. Price & C. Jackson (Eds.), Neighbours and successors of rome: Traditions of glass production and use in Europe and the Middle East in the later 1st millennium AD (pp. 200-232). Oxford: Oxbow Books. Stern, M. E. (1992). Early Export Glass in India. In V. Begley & R. D. De Puma (Eds.), Tamil Nadu, in Rome and India—The Ancient Sea Trade (pp. 113–124). Oxford University Press. Taddei, M. (1994). In R. M. Cimino (Ed.), Foreword in Ancient Rome and India (pp. 167–173). Rome: IsMEO and New Delhi: Italian Embassy Cultural Centre. Then-Obłuska, J. (2015). Cross-cultural bead encounters at the Red Sea port site of Berenike, Egypt. Preliminary assessment (seasons 2009–2012). Polish Archaeology in the Mediterranean, 24(1), 735–777. Then-Obłuska, J. (2019). Bead trade in roman ports: A view from the red sea port of MarsaNakari. In A. Manzo, C. Zazzaro & D. Joyce de Falco (Eds.), Stories of globalisation: The red sea and the persian gulf from late prehistory to early modernity (pp. 264–280). Leiden: Brill. Torres-Rodriguez, Jorge de., Gonzales–Ruibal, A., Kleinitz, C., Rodriguez, M. A. F., Barrio, C. M., & Jama, A. D. (2019, June). Excavation of a first century AD tomb in Heis (Somaliland): Evincing long distance trade contacts. Nyame Akuma, 91, 30–35. Wang, K.-W., & Jackson, C. (2014). A review of glass compositions around the South China Sea region (The Late 1st Millennium BC to the 1st Millennium AD): Placing iron age glass beads from Taiwan in context. Journal of Indo-Pacific Archaeology, 34, 51–60. Warmington, E. H. (1928). (reprint 1995). In The commerce between the roman empire and India. Munshiram Manoharlal. Wheeler, R. E. M., Ghosh, A., & Deva, K. (1946). Arikamedu: An indo-roman trading station on the East Coast of India. Ancient India, 2, 17–124. Whitehouse, D. (1989). Begram, the Periplus and Gandharan Art. Journal of Roman Archaeology, 2, 93–100. Wood, M. (2016). Eastern Africa and the Indian Ocean World in the first millennium CE: The glass bead evidence. In G. Campbell (Ed.), Early exchange between Africa and the wider Indian Ocean World (pp. 173–194). Palgrave MacMillan and Springer. Wood, M., Dussubieux, L., & Robertshaw, P. (2012). The glass of Chibuene, Mozambique: New insights into early Indian Ocean trade. South African Archaeological Bulletin, 67(195), 59–74. Wood, M., Panighello, S., Orsega, E. F., Robertshaw, P., van Elteran, J. T., Crowther, A., Horton, M., & Boivin, N. (2017). Zanzibar and Indian Ocean trade in the first millennium CE: The glass bead evidence. Archaeological and Anthropological Sciences, 9, 879–901. Zuchowska, M. (2013). Palmyra and the far eastern trade. Studia Palmyrenskie, 12, 381–387.

Indian Glass in Southeast Asia Laure Dussubieux

Abstract For the last decade, Southeast Asia has been an area where glass research has been extremely active. Although many questions remain unanswered, a clearer picture of the organization of the glass industry and trade/exchange through time has emerged. At a very early period (fourth–second century BCE), glass ornaments were manufactured in modern Thailand using techniques and raw materials imported from northern India, as well as other regions, possibly within Southeast Asia. Production from these very early workshops was distributed around the South China Sea. At the same early period, production centres located in other regions of Southeast Asia might have operated too, as evidenced by the very specific type of beads found in modern Myanmar with familiar compositions but with very local typologies. Around the second-first century BCE, a shift occurs with the loss of the northern Indian connection and instead glass beads from southern India and/or Sri Lanka appears at certain Southeast Asian sites. Other sites present a very different glass pattern suggesting that different exchange networks co-existed for a while. Around the middle of the 1st millennium CE, available data become scarce. Even if evidence indicates that ties with India seem to remain strong, more research would be necessary to define the nature and intensity of exchange between the two areas. Another point that needs investigation is the disappearance of certain primary glass production around that time. This might have stimulated imports from India, as well as the Middle East and later from China. Around the tenth century CE and later, connections with South India/Sri Lanka are still visible in Southeast Asia suggesting continuity over more than a millennium. In parallel, the presence of beads possibly from north-eastern India would indicate that a connection lost around the second century BCE was reestablished.

L. Dussubieux (B) Field Museum, Chicago, USA e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_20

489

490

L. Dussubieux

1 Introduction Southeast Asia is a vast and complex region bordered along its southern part by a complicated coastline and including a multitude of islands, creating difficulties for the circulation of people and goods from one point to the other of this region. The Bay of Bengal and the South China Sea create spheres of exchanges within Southeast Asia and between Southeast Asia and its immediate neighbours (India and China). As part of the Indian Ocean, Southeast Asia is connected to regions, albeit indirectly, as distant as the Middle East or Africa. In general, the earliest glass found in Southeast Asia appears around the middle of the fifth century BCE in the forms of beads or small personal ornaments (Dussubieux & Gratuze, 2010). Glass vessels are extremely rare in early periods and it is only at the end of the 1st millennium– beginning of the 2nd millennium CE that these types of artefacts start appearing in significant quantities (e.g. Guillot & Wibisono, 1998; Guillot et al., 2003; McKinnon, 1984; Perret & Jaafar, 2014). Scientific research on glass in Southeast Asia has been quite active in the last few decades with quantities of new studies relying on the elemental composition of glass (e.g. Carter, 2013; Dussubieux, 2001; Lankton et al., 2008). For decades, the work of Peter Francis Jr has dominated the research of South and Southeast Asian glass beads (e.g. Francis, 1990); however, his work predates the wide use of elemental analysis. According to Francis’ research, monochrome drawn beads found in Southeast Asia at early periods were manufactured in India (Francis, 1990). A site, dating from the second century BCE and onward, especially caught his attention (Francis 1991c). Arikamedu on the southeast coast of India in Tamil Nadu is a site that first drew attention for the Roman material found there (Wheeler et al., 1946). It also yielded large amounts of glass waste in the forms of tubes and glass chunks, suggesting that drawn beads were manufactured there. Francis (1990, 1991c) identified similar debris at a number of other sites in South and Southeast Asia (Fig. 1): Mantai (Sri Lanka), Oc-Eo (Vietnam), Khlong Thom (also called Khuan Lukpad), Sating Pra and Takua Pa (Thailand) and Kuala Selinsing and Sungai Mas (Malaysia). The model he proposed assumed that drawn bead making was developed in South India first, and craftsmen would have moved to different places where they would have perpetuated the drawn glass bead tradition. Based on the dating of the sites and the type of material evidence found there, he hypothesized that they might have first settled in Mantai, Khlong Thom and Oc-Eo around the second century CE that was, from earlier research about Arikamedu, the time believed to be when the site was abandoned. From there, the bead makers of Khlong Thom would have moved to Kuala Selinsing and Sungai Mas while the craftsmen in Oc-Eo would have transferred to Sating Pra and Takua Pa in Thailand. Around the tenth century CE, following the Chola invasion, the craftsmen in Mantai would have returned to southern India. Francis (2002) recognized that this model was not final and he was already making adjustments based on more recent archaeological findings. He included the possibility of other contemporaneous glass bead-making sites in South Asia such as Karaikadu. Also, the most recent archaeological excavations at Arikamedu indicated that the

Indian Glass in Southeast Asia

491

Fig. 1 Maps from South and Southeast Asia with some of the sites mentioned in the article

site was occupied much later than previously thought (Begley, 1996), making a movement of the glass bead makers around the second century CE unnecessary. He also pointed out that the movement of craftsmen would not have been due to their personal initiative but would have required a more powerful intervention, either governmental or from the trading entities or guild they were dependent from to dispose of their production. Even if more than twenty years later, Francis’ model is considered obsolete (Carter, 2016), it has been an essential starting point for the research presented below. Beyond new archaeological findings, one of the most important elements that greatly advanced the glass bead research is the inclusion of analytical techniques such as laser ablation—inductively coupled plasma—mass spectrometry in research designs. Such a technique, providing the major, minor and trace element concentrations for ancient glass was key to identifying new glass types and refined previously recognized glass types (Dussubieux et al., 2008, 2010; Dussubieux, 2001; Dussubieux & Gratuze, 2010).

2 Glass Analysis in South and Southeast Asia The glass situation in South and Southeast Asia is quite complex with many glass types and their distribution varying greatly at the regional scale. In Southeast Asia, glass types belong to different categories: • the mineral soda–high alumina glass; • the potash glass;

492

L. Dussubieux

• the mineral soda–lime–alumina glass; • the mineral soda-lime and soda plant ash glasses. A few glass types (such as the mixed-alkali glass) are voluntarily omitted from this paper, as their roles in our conclusions are minor.

2.1 Mineral Soda–High Alumina Glass The mineral soda–high alumina or m-Na-Al glass is one of the most abundant glass types found in Southeast Asia. This is a soda glass with high alumina concentrations (>5%). Its high concentration in India convinced Brill (1987) that this glass had been manufactured there. High trace element levels and more especially relatively high uranium concentrations in the m-Na-Al glass set it apart from most other ancient glasses. Its composition suggests the use of an immature sand, rich in feldspar, with a composition very close to that of the granite it came from. Ethnographic data as well as scientific experiments (Brill, 2003; Gill, 2017; Kock & Sode, 1995) indicate that in several parts of India, glass makers were using a sand naturally mixed to sodic efflorescence called reh containing large amounts of sodium salts (carbonate, bicarbonate and sulphate) and varying proportions of calcium and magnesium salts. (Wadia, 1975: 489, 501, 502). Five different m-Na–Al glasses named m-Na–Al 1, 2, 3, 4 and 6 were identified based on the concentrations of the following constituents Mg, Ca, Sr, Zr, Cs, Ba and U (Dussubieux & Wood, 2021; Dussubieux et al., 2010). In South India and Sri Lanka, the most abundant glass available is the m-Na–Al 1 glass type. It is associated in a few instances with evidence of glass working. The third century BCE–second century CE settlement of Giribawa, located in the northwestern part of Sri Lanka, yielded glass beads and furnaces lined with vitrified mNa–Al 1 materials and blocks of raw glass with similar compositions (Bopearachchi, 1999, 2002; Dussubieux, 2001; Gratuze et al., 2000). High alumina sand sources were identified in the area (Dussubieux, 2001). Another place in Sri Lanka where glass working is suspected is Mantai (Francis, 2013). At this site m-Na–Al 1 glass dominates the studied glass assemblage (Dussubieux, unpublished report). In South India, there are a few sites where drawn beads were possibly manufactured from mNa–Al 1 glass. Unfinished beads and raw glass were found at the site of Karaikadu (Deshpande, 1975: 21; Selvakumar, 2021), a site mentioned by Francis (2002) dating from the first centuries CE. The analysis of a few glass samples from this site revealed an m-Na–Al 1 glass composition (Dussubieux, 2001). More recently, Manikollai was identified as a possible location of drawn bead production. The analysis of surface finds revealed the use of m-Na–Al 1 glass at this site (Lankton et al., 2014). Initially, the m-Na–Al 2 glass was identified at sites dating from the ninth to the nineteenth century CE located on the west coast of India and the east coast of Africa (Dussubieux et al., 2008). Recently, additional research on beads from the east coast of Africa refined the chronology for this glass and suggested an occurrence from

Indian Glass in Southeast Asia

493

the fourteenth century CE and onward (Dussubieux & Wood, 2021). M-Na–Al 2 beads were found at the port of Chaul in Maharashtra, suggesting that it could have been an entry point for these beads into the Indian Ocean trade and possibly a place of manufacture. Gogte et al. (2006) present Chaul as a bead making location but the article does not offer any description of the artefacts and features leading to this conclusion. Glass beads made in ‘Chiawle’ (interpreted as Chaul) that the Portuguese would bring to ports along the east coast of Africa are mentioned in the travel account of Caesar Frederick (Federeci and Hickock, online), a Venetian merchant that visited the western part of India in the sixteenth century CE. The m-Na–Al 3 glass was identified at the site of Kopia located in the eastern part of Uttar Pradesh. This site yielded from a context dating to the first century CE (Kanungo et al., 2010), glass objects but also crucible sherds lined with greenish glass and raw glass fragments of the same colour with a m-Na–Al 3 composition (Dussubieux & Kanungo, 2013). Glass artefacts with a m-Na–Al 3 compositions were found in Southeast Asia at sites dating from the fourth–second century BCE showing that this glass was manufactured at very earlier periods and was exchanged as raw material and finished goods in Southeast Asia (Dussubieux & Bellina, 2017, 2018; Dussubieux & Pryce, 2016; Lankton et al., 2008). The m-Na–Al 4 glass type was found in Indonesia at sites dating from the fifteenth to sixteenth century CE in the forms of glass fragments (Dussubieux, 2009). It also appeared at the eleventh–fourteenth century CE site of Pulau Kampai, Malaysia (Dussubieux & Soedewo, 2018) and at the twelfth–fourteenth century CE site of Angkor Thom in Cambodia (Carter et al., 2019) suggesting an occurrence dating from the twelfth to the sixteenth century CE. This glass might have been produced in Uttar Pradesh like the m-Na–Al 3 glass as both glasses have very similar compositions characterized by higher cesium and uranium concentrations compared to the other m-Na–Al glasses. The m-Na–Al 6 glass has been recently identified among glass beads excavated on the east coast of Africa. It is present at sites dating from the ninth to the thirteenth century CE (Dussubieux & Wood, 2021). More details will be given in this article about its distribution in Southeast Asia. Its origin is extremely obscure at this point as such a glass type has not been identified in India yet. Recent Sr isotope analysis conducted on m-Na–Al 6 glass samples revealed a fairly radiogenic signature that is compatible with an origin from the Indo-Ganges region (Seman et al., 2021) where this glass might have been produced.

2.2 Potash Glass This glass will be described here briefly as it is extremely abundant in Southeast Asia but its origin, although quite uncertain at this point, does not seem to be Indian. In this glass, potash is the most abundant constituent after silica and its concentration is around 15%. Soda concentrations are generally low as are magnesia concentrations. Lime and alumina concentrations vary in a wide range, and based on compositions

494

L. Dussubieux

identified in China, South and Southeast Asia, three groups were distinguished: low lime–high alumina–potash glass, low alumina–high lime–potash glass and a moderate lime and alumina–potash glass (Dussubieux & Gratuze, 2010; Lankton & Dussubieux, 2013; Lankton et al., 2008). From a general point of view, it seems that most of the potash glass from South Asia belong to the group with moderate lime and alumina concentrations whereas, the potash glass from China seems to have higher alumina concentrations. Some samples from Southeast Asia seem to have low alumina and high lime concentrations, with others belonging to the moderate lime and alumina group or the high alumina group. Potash glass is widely distributed in China (Li, 1999; Qishan, 1996; Zhang et al., 2004) and in South and Southeast Asia (Dussubieux, 2001; Dussubieux & Gratuze, 2010; Lankton & Dussubieux, 2006; Lankton et al., 2008). Its presence is also attested in Central Asia (Hall & Yablonsky, 1998) and as far as North Korea and Japan (Koesuka & Yamasaki, 1996; Lee et al., 1993). The existence of several subgroups supports the hypothesis that potash glass was produced at different locations. The provenance of the potash glass was explored using isotope analysis (Brill & Fullagar, 2009; Li, 1999). At least three different strontium isotope signatures were identified suggesting that potash glass could have been manufactured in at least three locations. Li (1999) tested high alumina–low lime potash glass and moderate alumina and lime potash glass from the southern Chinese province of Guangxi and concluded based on their lead isotope signatures that they were manufactured locally. Orange glass beads from Myanmar with a potash composition and containing several per cent of copper exhibited a lead isotope signature that pointed towards a copper source located in northern Vietnam (Dussubieux & Pryce, 2016). The presence of potash-based faience artefacts from different locations in China also suggests a long tradition of potash glassmaking in China (Henderson et al., 2018; Lin et al., 2019).

2.3 Mineral Soda–Lime–Alumina Glass This glass group, also named m-Na-Ca-Al, is characterized by a very exceptional heterogeneity, but it is hoped that additional data will help disentangle some subgroups. Its composition varies widely but in general, it is a soda glass with soda concentrations ranging from 10 to 15%. Lime and alumina concentrations vary in a wide range and can be as high as approximately 7%. Magnesia and potash concentrations are generally low, below 1.5%, but there are exceptions with much higher concentrations of these constituents for specific colours such as red or black. This glass often has relatively high uranium concentrations (Dussubieux & Gratuze, 2003; Dussubieux et al., 2012). The colours in this glass group include opaque red, green and yellow; dark glass is often purple (coloured with manganese) or dark blue (coloured with cobalt). The cobalt added to the m-Na–Ca–Al glass contains a number of impurities including barium and is generally associated with manganese. The provenance of this glass is very uncertain but it is important to notice that a very similar cobalt colourant with Mn and Ba is used in the potash glass described earlier. This kind of

Indian Glass in Southeast Asia

495

cobalt ore is different from the ones used in the m-Na–Al glasses (Dussubieux et al., 2008). Colourant and more especially cobalt are often traded long distances but the fact that the m-Na-Ca-Al and potash glasses used the same type of cobalt suggests that the production centres for both glasses were geographically close to each other and were perhaps using a local cobalt source. A certain number of sites such as Phu Khao Thong, Thailand, second century BCE-fourth century CE, Arikamedu, India, second century BCE and onward and Aw Gyi, Myanmar (the dating of this site is uncertain but it is quite likely contemporaneous of Phu Khao Thong and Arikamedu) yielded particularly high proportions (~50%) of m-Na–Ca–Al glass (Dussubieux et al. 2020; Dussubieux et al., 2012). Other sites all over South and Southeast Asia included m-Na–Ca–Al glass beads but in comparatively smaller proportions. M-Na–Ca–Al glass beads were found at the site of Kish in Iraq in a context dating from the second– first century BCE which constitutes its unique known occurrence in western regions (Dussubieux, 2021).

2.4 Mineral Soda-Lime and Soda Plant Ash Glasses Those two types of glass are both imports from the west. The soda plant ash glass is the earliest glass manufactured in the Middle East and Egypt during the Late Bronze Age (e.g. Shortland et al., 2018). Around the ninth century BCE the plant ash flux is replaced by natron in Egypt and the Syro-Palestinian region (Shortland et al., 2006) but the plant ash-based recipe continued to be used in the Sasanian Empire (Mirti et al., 2008, 2009). Finally, natron glass disappeared at the end of the 1st millennium CE and it is replaced by soda plant ash glass. Started around the eighth century CE, the diffusion of soda plant ash follows the expansion of the Arab trade, especially around the Indian Ocean. In Southeast Asia, natron glass appears mostly in the form of fragments of glass vessels although there is a major exception with the site of Ban Ro in Thailand, dating from the first half of the 1st millennium CE that yielded significant quantities of natron glass waste and drawn beads raising the possibility of the recycling of imported glass (Dussubieux, unpublished data). Glass vessel fragments with a soda plant ash composition in Southeast Asia are generally dated from the end of the first millennium CE (Carter et al., 2019; Dussubieux, 2009, 2014; Dussubieux & Allen, 2014); however, this glass type, exclusively in the forms of beads, is available at many sites at an earlier period, around the beginning of the 1st millennium CE (Dussubieux, 2001). The mechanisms of the exchanges bringing these beads to Southeast Asia are not known at this point.

496

L. Dussubieux

3 Current State of the Research on Glass in Southeast Asia 3.1 Ornament Making at Khao Sam Kaeo and Khao Sek Between 2005 and 2009, the French-Thai mission lead by Bérénice Bellina, excavated the site of Khao Sam Kaeo, located at one of the narrowest points of the Kra Isthmus in Thailand (Bellina, 2017). This site was a trading and industrial centre dating from the fourth to the second century BCE where more than 2500 glass artefacts were recovered (Dussubieux & Bellina, 2017). The glass findings at Khao Sam Kaeo are special in many ways. The types of artefacts recognized at the site were beads, bangles but more importantly waste: waste obtained by knapping and wastes exhibiting signs of hot working processes. In addition to drawn monochrome beads so common at many South and Southeast Asian sites, Khao Sam Kaeo yielded beads shaped with a lapidary technique such as the one used by Indian stone bead makers. Those lapidary beads were manufactured from a transparent green glass (maybe to imitate beryl) that was also used for bangles (Fig. 2a). Most of the glass waste were also from this translucent greenish glass. No structure related to a glass-working area was found; however, the highest concentration of debris was found near the Tha Tapao river that borders the site on its south and west sides, and it is quite possible that a change in the riverbed location destroyed the glass-working area. Also, heavy

Fig. 2 a Lapidary beads from Khao Sam Kaeo, b and c glass material from Khao Sek (no scale as the pictures were taken with the objects in a display case), d glass material from the Samon Valley, Myanmar

Indian Glass in Southeast Asia

497

looting affected this site and might have obliterated any trace of a glass workshop. The chemical analysis of glass samples from the site revealed high proportions of potash and m-Na–Al 3 glasses. An important observation that one needs to keep in mind as far as Khao Sam Kaeo is concerned is that this site yielded a mature glass industry that quite likely was set up by experienced craftsmen. It was therefore established that lapidary beads and bangles were manufactured at Khao Sam Kaeo quite likely by Indian craftsmen present at the site that were familiar with glass and lapidary work, with glass imported from north-eastern India and the southern part of China. The Khao Sam Kaeo glass production was likely distributed around the South China Sea as similar products with a potash or m-Na–Al 3 compositions were found in Thailand, Vietnam and the Philippines (Lankton et al., 2008). The production of glass ornaments at Khao Sam Kaeo during the fourth to second century BCE was fairly surprising based on our knowledge available at the time of the discovery; therefore, finding a second site, contemporaneous and quite similar 50 km south of Khao Sam Kaeo was quite puzzling. Alerted by the discovery of a villager, the French-Thai mission moved its attention to the site of Khao Sek that they excavated in 2013 and 2014 (Bellina and Sinopoli, 2018). Like Khao Sam Kaeo, Khao Sek was a trading and a crafting centre, although of a slightly lesser importance. Large quantities of glass objects, including lapidary beads, bangles and glass wastes were found at the site although not in situ (Fig. 2b, c). Indeed, all of the glass material was removed by the owner of the land, destroying the context in the process, and any chance to look for structure indicative of a workshop. The similarity of the glass artefacts from Khao Sam Kaeo and Khao Sek was obvious whether it was in terms of shapes, colours, techniques and compositions (Dussubieux & Bellina, 2018). The data yielded by the study of the glass found at Khao Sek and Khao Sam Kaeo provide unprecedented evidence in Southeast Asia for a regional craft standardization at a very early period (fourth–second century BCE) with a hybrid industrial model involving an exogenous technology. Further research in the central region of Myanmar revealed there the presence of mNa–Al 3 and potash glasses, which is a marker of earlier sites (fourth- second century BCE), but the typology of the beads was specific to the area, with unique disc-shaped translucent turquoise and dark blue beads and tire-shaped translucent turquoise beads (Fig. 2d). The beads in question were recovered at cemeteries scattered along the Samon Valley and no traces of a bead making workshop was identified in the vicinity (Dussubieux & Pryce, 2016). The uniqueness of the bead typology suggests a local production and, contrasting with Khao Sek and Khao Sam Kaeo, a limited and very regional diffusion of finished products. Here again, the presence of a mature glass technology in a region where no precursor for it was identified suggests exogenous involvement. Small drawn red beads with a m-Na–Al 3 composition were found at Khao Sam Kaeo, Khao Sek, and in the Samon Valley, as well as Ban Don Ta Phet, Thailand (fourth century BCE). No evidence of drawn bead manufacturing was found in Thailand or elsewhere in Southeast Asia at such an early period and it is quite likely that these beads were imported from India.

498

L. Dussubieux

3.2 The Bay of Bengal Bead Trade Network Moving forward in time, we will focus now our attention on Arikamedu. This site, yielded as mentioned previously, quantities of glass wastes suggesting that drawn beads were manufactured there (Fig. 3a). Some glass objects from this site were chemically analysed along with objects from a range of contemporaneous sites in South India and Sri Lanka. They include Alagankulam, Kodumanal (both dating from the third century BCE–third century CE), Karaikadu (first–fifth century CE) in India and Giribawa, Kelanyia (both dating from the third century BCE–second century CE) and Ridiyagama (fourth century BCE-first century CE) in Sri Lanka. The results from Arikamedu contrast with the results from all the other sites. The dominant glass type at other sites from this time period is the m-Na–Al 1 glass type, which is extremely rare at Arikamedu. Instead, Arikamedu yielded equivalent proportions of potash and m-Na–Ca–Al glasses (Dussubieux, 2001; Dussubieux et al., 2008, 2012). On the other side of the Bay of Bengal, on the west coast of the Kra Isthmus, the sites of Aw Gyi at the extreme south of Myanmar and Phu Khao Thong in Thailand yielded glass artefacts including finished beads but also tubes, glass chunks and melted bead clusters (Fig. 3b, c) (Bellina et al., 2018). Phu Khao Thong is part of a trading complex including several other sites located in the Sukhsamran district, Ranong province, on the Andaman coast. Phu Khao Thong was dated, based on the material collected at the site, between the second century BCE and the fourth century CE. A sherd with a Tamil-Brahmi inscription, dated from the second century CE (Bellina et al., 2014: 81) and rouletted ware fragments indicates contacts with South Asia (Bouvet, 2017), whereas the presence of Roman intaglios and a granulated gold bead from Iran shows that Phu Khao Thong was part of a trading networks extending over long distances (Chaisuwan, 2007; Chaisuwan and Naiyawat, 2009).

Fig. 3 a Glass material from Arikamedu, b glass material from Phu Khao Thong, c glass material from Aw Gyi

Indian Glass in Southeast Asia

499

Aw Gyi, in southern Myanmar, is a site located north of Phu Khao Thong on the west coast of the Kra Isthmus. It consists of three hills bordered by two watercourses linking it to the sea. The site was disturbed by heavy looting limiting the area that could be excavated. No absolute dating is available but material recovered at the site seems to indicate an occupation during the first centuries CE. Evidence of semiprecious stone production, as well as a large quantity of pottery fragments, some of which parallels with those found east in southern Vietnam at Oc Eo, in China (the Han pottery) and in the Philippines (the Kalanay-related pottery), and west along the eastern coast of India (the Fine Grey pottery) were found at Aw Gyi. Long distance connections are attested by the presence of glass artefacts with a Syro-Palestinian composition (Dussubieux et al., 2020). Elemental analysis conducted on glass material found at these two sites exhibits high proportions of potash and m-Na–Ca–Al glass, similar to what was found at Arikamedu (Dussubieux et al., 2012, 2020). A connection between Arikamedu, Aw Gyi and Phu Khao Thong seems obvious, through a network operating around the Bay of Bengal. What about a transfer of technology from Arikamedu to Phu Khao Thong and Aw Gyi as far as drawn bead making is concerned? Aw Gyi and Phu Khao Thong were only partially investigated and heavily affected by looting. No glass-working areas could be identified. At Arikamedu, craft areas remained elusive despite extensive excavations. Therefore, a full comparison between the three sites cannot be carried out; however, transfer of technology between Arikamedu and Phu Khao Thong and Aw Gyi should not be totally dismissed and would offer a more likely (at this point in time) alternative to the hypothesis proposed by Francis. The site of Khlong Thom in Thailand (second–seventh century CE) yielded evidence of bead making and possibly glassmaking (Bronson, 1990; Veraprasert, 1987). Unpublished glass compositions at this site include m-Na-Ca-Al glass with an extremely small proportion of potash glass. This smaller amount of potash glass could be tied to the later dating of the site but the presence of m-Na–Ca–Al glass could indicate a continuity for glass working between the sites of Arikamedu, Phu Khao Thong and Aw Gyi and the later site of Khlong Thom.

3.3 The South India/Sri Lanka-Southeast Asia Connection In Southeast Asia, two regions were particularly well studied as far as glass is concerned: Thailand and Cambodia. Carter (2015) studied 7 Cambodian sites and 3 sites located in central and northeast Thailand (Table 1). To that, we can add sites located on the Kra Isthmus (Bellina et al., 2019) (Table 2). Sites can be separated in two groups whether the dominant glass type is potash or m-Na–Al 1. Potash glass as indicated earlier might originate from China and/or northern Vietnam. The m-Na–Al 1 glass was manufactured in South India and Sri Lanka. Carter (2015) interpreted the dominating presence of potash glass at certain sites as a continuity of the South China Sea trade network from earlier periods (fourth– second century CE) into the next period (300 BCE to 300 CE). In the first centuries

500

L. Dussubieux

Table 1 Sites appearing in Carter (2015) Country

Site

Dating

Results

Cambodia

Angkor Borei

200BC–200AD

m-Na–Al 1, m-Na-Ca-Al

Cambodia

Bit Meas

200BC–100AD

Potash

Cambodia

Phnom Borei

200BC–200AD

m-Na–Al 1, potash

Cambodia

Phum Snay

350BC–200AD

m-Na–Al 1, potash

Cambodia

Prei Khmeng

1–6 c. AD

m-Na–Al 1

Cambodia

Prohear

200BC–100AD

Potash, m-Na–Ca–Al

Cambodia

Village 10.8

400BC–50AD

Potash

Thailand

Ban Non Wat

450BC–600AD

m-Na–Al 1, potash, m-Na–Ca–Al

Thailand

Noen U-Loke

100BC–600AD

m-Na–Al 1, mixed-alkali

Thailand

Promtin Tai

500BC–500AD

m-Na–Al 1, mixed-alkali

Table 2 Sites appearing in Bellina et al. (2019) Site

Material

Glass types

Khao Krim

Drawn beads

Potash (100%)

Pang Wan

Drawn beads

Potash (100%)

Ban Nai Hyan

Drawn and lapidary beads, cluster of melted beads, wastes

Potash (100%)

Tham Chaeng

Drawn beads

Potash (100%)

Tham Nam Lot Yai

Drawn bead (1 specimen)

m-Na–Al 1 (100%)

CE, more m-Na–Al 1 glass beads indicate a higher intensity of trade with South India/Sri Lanka that would signal the emergence of a new network with ramifications in peninsular Thailand and in the east part of Cambodia.

3.4 From the End of the 1st Millennium—to the Beginning of the 2nd Millennium CE The first part of this article shows a picture of glass bead production and trade from 500 BCE to 500 CE. From the mid-first century CE and onward, the situation is less clear. This is due mostly to the fact that fewer sites belonging to this later period have been studied. A few results for the elemental analysis of glass material from three later sites are nevertheless available and will be discussed in this section: Sungai Mas in Malaysia, dating from the ninth–eleventh century CE, Pulau Kampai in Indonesia, dating from the eleventh to the fourteenth century CE and the burial sites of the Cardamom Mountains, dating from the fifteenth to the seventeenth century CE. Sungai Mas, one of the sites included in Francis’ model, is located in Malaysia, on the west part of the Thai-Malay peninsula. This site was surveyed by Jane Allen

Indian Glass in Southeast Asia

501

in the 1980s for her PhD dissertation (Allen, 1988). It is a trading complex dating from the ninth to the eleventh century CE. According to Francis, evidence of drawn bead making was found there (Francis, 1991c). Beads were collected at the site along with glass vessel fragments although, glass vessel will not be discussed here (Fig. 4). The beads belong either to the m-Na–Al 1 glass or the soda plant ash glass type (Dussubieux & Allen, 2014). The m-Na–Al 1 glass was mentioned earlier as it was a glass type that was particularly common in South India and Sri Lanka and found in abundance in Southeast Asia starting at the beginning of the 1st millennium CE. Obviously, this glass type was produced over a long period and was still being distributed in Southeast Asia during the ninth–eleventh century CE period. The other glass made from soda plant ash was manufactured from a recipe used in the Middle East. Some polychrome beads (e.g. eye beads) had a typology well known in that region suggesting that these beads were imported as finished products. The manufacturing location is less clear for some of the soda plant ash monochrome drawn beads that were indistinguishable from other drawn beads with a m-Na–Al 1 composition. The soda plant ash composition indicates it was an import, but the manufacturing technique could have been local suggesting the possibility of glass recycling using glass vessel fragments available at the site (Dussubieux & Allen, 2014). The second of the later sites discussed here is called Pulau Kampai (Dussubieux & Soedewo, 2018). It is an island in the north part of Sumatra in Indonesia facing the Strait of Malacca. It seems that the island might have been a port of call for the ships using the maritime road connecting India to China. The site was dated from the eleventh to the fourteenth century CE. It was excavated recently by Ery Soedewo, an Indonesian archaeologist (Soedewo, 2013). One of the glass types identified at this site is the soda plant ash glass also found in Sungai Mas. Two other glass types were identified: mineral soda–high alumina 4 and 6. Those two types of glass are not available at earlier periods. They appear later and their origin is unclear. Because of certain similarities between the m-Na–Al 3 and m-Na–Al 4 glass, it was assumed that the m-Na–Al 4 glass might have been manufactured in eastern Uttar Pradesh Fig. 4 Glass material from Sungai Mas

502 Table 3 Recapitulation of periods and glass types based on the study of the following three sites: Sungai Mas, Pulau Kampai and sites in the Cardamom Mountains

L. Dussubieux Dating

Indian glass

Other glass types

9–11 c. CE

m-Na–Al 1

v-Na–Ca (Middle East)

11–14 c. CE

m-Na–Al 4 and 6

v-Na–Ca (Middle East)

15–17 c. CE

m-Na–Al 2

Si–Pb–K (China) and other

and could have been a much later resurgence of the m-Na–Al 3 glass. So far, no m-Na–Al 6 glass was identified in India; isotope analysis suggests a north Indian origin (Seman et al., 2021). The Cardamom Mountain burial sites were examined by Nancy Beavan and colleagues and the glass was studied by Alison Carter (Carter et al., 2016). The material was dated from the fifteenth to the seventeenth century CE. The two most abundant glass types identified at this site included mineral soda—high alumina 2 glass and high lead-high potash silica glass. A couple of beads belonged to the mNa–Al 4 glass group and a high Mg-m-Na–Al glass, previously unknown and of uncertain origin, was also present. The lead–potash–silica glass corresponds to some Chinese imports that start appearing after the thirteenth century CE. The m-Na–Al 2 glass is from India and was likely exported through the port of Chaul, in Maharashtra. Even if the study of these three sites constitutes only a very small window in what could have happened in the glass trade in Southeast Asia at the end of the 1st millennium-first half of the 2nd millennium CE, some trends appear that will need to be confirmed and refined (Table 3). It seems that there was a constant shift in the kind of glasses that were traded. The reason for the shift is unclear at this point and there is a need to understand these changes at the regional level. This demonstrates the usefulness of glass as a proxy to understand better the interactions within Southeast Asia and between Southeast Asia and neighbouring regions.

4 Summary and Discussion 4.1 Summary The study of glass in Southeast Asia gives precious information about Indian glass and more especially about the m-Na–Al glasses. Data from Southeast Asia are contributing significantly to the chronology of the different subtypes of this glass type. The m-Na–Al 3 is present at the earliest sites dating from the fourth to the second century BCE. A little later, high volumes of m-Na–Al 1 glass reach some regions of Southeast Asia (Carter, 2015). This glass type dominates at a number of sites until the eleventh century CE. The m-Na–Al 6 glass type replaces this glass (eleventh–fourteenth century CE). The m-Na–Al 4 glass type, although less abundant than the m-Na–Al 6 glass, is present during the same period. Around the fifteenth century CE, the m-Na–Al 2 glass appears at the latest sites.

Indian Glass in Southeast Asia

503

Although these different glass productions are present during given periods in Southeast Asia, it does not mean that their production lifespan coincides exactly with these periods; however, results of the glass research in Southeast Asia points towards the existence of major glass-working centres in India with shifting connections to long distance trade networks. The earliest of these glass-working centres are located in eastern Uttar Pradesh. These centres exported raw glass (used at Khao Sam Kaeo and Khao Sek) and finished objects such as the small m-Na–Al 3 drawn red beads found at different sites of Southeast Asia. Next is the emergence of the m-Na–Al 1 glass making centres in South India and Sri Lanka. They manufactured drawn beads. It would be interesting to explore the possibility of a transfer of technology for the glassmaking and the drawn bead technique from eastern Uttar Pradesh to the south. Research at Khao Sam Kaeo and Khao Sek in Thailand clearly indicates that craftsmen were circulating from one region to another as evidence at those sites show a transfer of technologies used in northern India to the east coast of the Kra Isthmus around the fourth–second century BCE. The m-Na–Al 1 sites in South India and Sri Lanka are dated very broadly and seem contemporaneous. It is not possible to determine whether the m-Na–Al 1 glass bead production started in South India and then expanded to Sri Lanka or whether a different scenario should be considered. The production of m-Na–Al 1 glass beads lasted more than a millennium but finally vanished around the eleventh century CE. At this point, the presence of two new m-Na–Al glasses is detected: the m-Na–Al 6, which has an unknown provenance and the m-Na–Al 4 that bears similarities with the m-Na–Al 3 glass and therefore could have been produced in the same region. Here again, should we consider that the craftsmen mastering the production of m-Na–Al 1 drawn glass beads transferred their knowledge to different regions? The presence of the m-Na–Al 4 and 6 glasses is relatively short lived compared to that of the m-Na–Al 1 glass as by the fifteenth century CE, it is replaced by the m-Na–Al 2 glass. However, considering the m-Na–Al glasses only (and ignoring at this point the other glasses) takes into account only one piece of the ‘glass puzzle’. The combination of the potash/m-Na–Ca–Al glasses at Arikamedu, contrasting with the m-Na–Al 1 composition found at other South Indian contemporaneous sites, could indicate that different groups of people were involved in the manufacture of glass and glass beads in the south part of India. The Arikamedu production seems to be tied earlier than the m-Na–Al 1 production to Southeast Asia, as early as the second century BCE when the m-Na–Al 1 export would be visible in the Southeast Asian record only in the first century CE. The m-Na–Ca–Al production vanishes after the seventh century CE based on the dating from Khlong Thom but possibly earlier. The lack of certainty about the provenance of the m-Na-Ca-Al glass type found at Arikamedu is a limitation in our interpretation.

504

L. Dussubieux

4.2 Rewriting Francis’ Model Circling back to Francis’ model, it is obvious that new data available since Francis formulated his hypothesis about the circulation of drawn bead technology are revealing a reality that is way more complicated than what was thought 20 years ago. New data suggest that the technology for drawn beads in South India/Sri Lanka might have been imported from elsewhere and not created in this region as earlier drawn material with a non-South Indian origin was spotted in Southeast Asia. This technology might have come from the east part of modern Uttar Pradesh. This hypothesis is not without challenge as this area is known today for wound rather than drawn beads (Kanungo, 2004). A number of sites were studied in South India/Sri Lanka but data are scarcer in the northern part of India as only compositions from Kopia were recently published. Independently from where this technology come from, it is possible that the drawn bead technology might have been transmitted to different groups of people in South India: a group using the m-Na–Al 1 glass for the manufacture of their beads and another group using the m-Na–Ca–Al/potash glass in Arikamedu. Glass material from the three sites (Mantai, Klong Thom and Oc-Eo) proposed as landing places for Arikamedu’s drawn bead craftsmen have been analysed since Francis’ publications. We already discussed Mantai, which yielded m-Na–Al 1 glass. Therefore, a connection with other South Indian/Sri Lankan sites with m-Na–Al 1 is possible (Karaikadu, Manikollai, Giribawa) but a connection with Arikamedu is unlikely. The material from Khlong Thom includes m-Na–Ca–Al glass with very small proportions of potash glass. This is compatible with a connection to Arikamedu. This connection might be indirect and occurred through the extension of the Bay of Bengal network beyond the second century CE. All the other sites mentioned by Francis, with recently analysed material, have a majority of m-Na–Al 1 glass (Oc-Eo, Kuala Selinsing and Sungai Mas). A comment comes to mind when considering the presence of m-Na–Al 1 glass at different sites of Southeast Asia where bead making might have occurred. If we consider that the beads were manufactured at those sites, only the technology to manufacture beads was transferred, the technology to produce the raw glass seems to have remained within India as there is no evidence of a new recipe or variation of well-established recipes developed at Southeast Asian sites. The bead makers would have kept their connection with South India/Sri Lanka to obtain the glass. They might have obtained raw glass from elsewhere too: it is possible that at a site such as Sungai Mas, where soda plant ash glass fragments and beads were found, recycling occurred. Imported glass fragments from the Middle East might have been used as raw material to manufacture drawn beads at the site. The later part of Francis’ model is difficult to revisit as we lack data for some of the sites (Sating Pra, Taku Pa and Kuala Selinsing). The m-Na–Al 1 glass that originates from Sri Lanka/South India disappears from Southeast Asia around the eleventh century CE. The glassmakers might have moved to different regions of India, either

Indian Glass in Southeast Asia

505

creating new glassmaking centres or joining existing ones (and bringing their trade connections with them) causing a shift in glass procurement from the southern part of the country to the north (m-Na-Al 4 and 6 glasses) and possibly other regions in India. Based on our observations, glass bead makers would be more mobile than the craftsmen that produce the raw glass. The possibility of the bead makers moving around Southeast Asia is supported by the presence of bead making wastes at different sites but evidence for glassmaking outside India is lacking at this point. The bead makers in Southeast Asia would have procured glass from India or from other regions (e.g. glass fragments from the Middle East).

5 Conclusion Research in Southeast Asia is vital to understand Indian glass bead production and diffusion and has provided new information to improve Francis’ model describing the evolution of drawn bead production in India and elsewhere. The most striking results of this synthesis are that lapidary beads were manufactured in Thailand during the fourth–second century BCE period and that drawn bead making might originate from the northern part of India and might have moved south sometime during the second half of the 1st millennium BCE. Two distinct groups of people manufactured drawn beads in South India/Sri Lanka: one group used m-Na–Al 1 glass and the other group, limited to the site of Arikamedu used m-Na–Ca–Al and potash glasses suggesting that each group was part of different exchange networks. The presence of m-Na–Ca–Al glass at Khlong Thom could indicate a connection between the craftsmen at this site and the craftsmen from Arikamedu. Other sites such as Mantai, Sungai Mas and Oc-Eo, mentioned by Francis, yielded mostly m-Na-Al 1 glass. If a transfer of technology occurred between South India and those sites, the craftsmen in Southeast Asia kept using glass imported from India. The research about glass in Southeast Asia conducted over the last twenty years is improving our understanding of the interactions between South and Southeast Asia through time but raises many questions that will be addressed only if more research is conducted in both regions. First, there is a need for a better mapping of the occurrence of the different glass types that were recognized so as to draw a better picture indicating where they are concentrated and where they are distributed. In India, South India was relatively well documented, but information is extremely scarce for the rest of the country. In Southeast Asia, Thailand and Cambodia constitute most of the research. Large areas have been understudied. They include Laos, Vietnam, Malaysia, Indonesia and the Philippines. Later periods thus far have been neglected and glass compositions from sites dated from the mid-1st millennium and onward would need to be added to the datasets to understand the evolution of the glass production and glass distribution beyond the Iron Age in India and Southeast Asia.

506

L. Dussubieux

References Allen, J. (1988). Trade, transportation, and tributaries: Exchange, agriculture, and settlement distribution in early historic-period Kedah, Malaysia. PhD Thesis. University of Hawai‘i at M¯anoa. Begley, V. (Ed.). (1996). The ancient port of Arikamedu: New excavations and researches 1989–1992 (Vol. 1). Paris / Bangkok: Centre d’histoire et d’archéologie / École Française d’Extrême-Orient. Bellina, B., Silapanth, P., Chaisuwan, B., Thongcharoenchaikit, C., Bernard, J. V., Borell, B., Bouvet, P., Castillo, C., Dussubieux, L., Malakie LaClair, J., Srikanlaya, S., Peronnet, S., & Pryce, T. O. (2014). The development of coastal polities in the Upper Thai-Malay Peninsula in the late first millennium BCE. In N. Revire & S. Murphy (Eds.), Before Siam was born: New insights on the art and archaeology of pre-modern Thailand and its neighbouring regions (pp. 68–89). River Books. Bellina, B. (Ed.). (2017). Khao Sam Kaeo: An early port-city between the Indian Ocean and the South China Sea. Mémoires Archéologiques 28. Écoles Française d’Êxtrême-Orient. Bellina, B., & Sinopoli, C.M. (Eds.). (2018). Special issue: The late prehistoric port-of-trade of Khao Sek (Chumpon province, Thailand) and its implications for the understanding of early trade polities in Maritime Southeast Asia (Vol. 13, pp 1–78). Archaeological Research in Asia. Bellina, B., Win, M. S., Htwe, K. M. M., Thu, H. M., Castillo, C., Colonna, C., Dussubieux, L., Favereau, A., Miyama, E., Pradier, B., Pryce, O., Srikanlaya, S., & Trivière, E. (2018). Myanmar’s earliest Maritime Silk Road port-settlements revealed, Antiquity, 92(366) e6, 1–5. Bellina, B., Favereau, A., & Dussubieux, L. (2019). Southeast Asian early Maritime Silk Road trading polities’ hinterland and the sea-nomads of the Isthmus of Kra. Journal of Anthropological Archaeology, 54, 102–120. Bopearachchi, O. (1999). Sites portuaires et emporia de l’ancien Sri Lanka, nouvelles données archéologiques. Arts Asiatiques, 54, 5–23. Bopearachchi, O. (2002). Les relations commerciales et culturelles entre Sri Lanka et Inde du sud: Nouvelles données archéologiques et épigraphiques. Cahier Du Cercle D’études Et De Recherches Sri Lankaises, 4, 1–16. Bouvet, P. (2017). Indian Fine Ware. In B. Bellina (Ed.), Khao Sam Kaeo: A late prehistoric early city between the Indian Ocean and the South China Sea Mémoires Archéologiques 28 (pp. 281–308). Paris: Ecole française d’Extrême-Orient. Brill, R. H. (1987). Chemical analyses of some early Indian glasses. In H. C. Bhardwaj (Ed.), Archaeometry of glass—Proceedings of the archaeometry session of the XIV international congress on glass—1986, part I (pp. 1–25). New Delhi, India: Calcutta, Indian Ceramic Society. Brill, R. H. (2003). The glassmakers of Firozabad and the glassmakers of Kapadwanj: Two pilot video projects. In Annales du 15e congrès de l’Association internationale pour l’Histoire du Verre, Corning New York, 2001 (pp. 267–268). Nottingham, UK: AIHV. Brill, R. H., & Fullagar, P. D. (2009). Strontium-Isotope studies of historical glasses and related materials: A progress report. In Annales AIHV, 17, Antwerp, 2006 (pp. 552–557). Bronson, B. (1990). Glass and beads at Khuan Lukpad, Southern Thailand. In I. Glover, & E. Glover (Eds.), Southeast Asia Archaeology 1986, proceedings of the first conference of the association of Southeast Asian archaeologist in western Europe, Institute of Archaeology, University College, London, 8–10th, September 1986 (pp. 213–229). Oxford: BAR International Series 56. Carter, A. K. (2013). Trade, exchange, and, socio-political development in Iron Age (500 BC– AD 500) mainland Southeast Asia: An examination of stone and glass beads from Cambodia and Thailand. PhD Thesis. Madison: Department of Anthropology, University of Wisconsin-Madison. Carter, A. K. (2015). Beads, exchange networks and emerging complexity: A case study from Cambodia and Thailand (500 BCE–CE 500). Cambridge Archaeological Journal, 25(04), 733– 757. Carter, A. K. (2016). The production and exchange of glass and stone beads in Southeast Asia from 500BCE to the early second millennium CE: An assessment of the work of Peter Francis in light of recent research. Archaeological Research in Asia, 6, 16–29.

Indian Glass in Southeast Asia

507

Carter, A. K., Dussubieux, L., & Beavan, N. (2016). Glass beads from 15th–17th century CE Jar Burial Sites in Cambodia’s Cardamom Mountains. Archaeometry, 58(3), 401–412. Carter, A. K., Dussubieux, L., Polkinghorne, M., & Pottier, C. (2019). Glass artifacts at Angkor: Evidence for exchange. Archaeological and Anthropological Sciences, 11(3), 1013–1027. Chaisuwan, B. 2007. Early contacts between India and the Andaman Coast in Thailand from the 2nd c. B.C. to 11th c. A.D. Singapore: Institute of Southeast Asian Studies. Chaisuwan, B., & Naiyawat, R. (2009). Thung Tuk: A settlement linking together the Maritime Silk Route. Trio Creation. Deshpande, M. N. (Ed.). (1975). Indian Archaeology 1966–67—A review. Archaeological Survey of India. New Delhi: Government of India. Dussubieux, L. (2021). Elemental compositions and glass recipes. In A. K. Kanungo, & L. Dussubieux (Eds.), Ancient glass of South Asia—Archaeology, ethnography and gobal connection. Singapore/Gandhinagar: Springer Nature/IIT Gandhinagar. Dussubieux, L., Bellina, B., Oo, W. H., Win, U. M. S., Tut, H. M., Htwe, K. M. M., & Kyaw, K. (2020). First elemental analysis of glass from Southern Myanmar: Replacing the region in the early Maritime Silk Road. Archaeological and Anthropological Sciences, 12(139). Dussubieux, L., & B. Bellina. (2018). Glass ornament production and trade polities in the UpperThai Peninsula during the Early Iron Age. In B. Bellina, & C. M. Sinopoli (Eds.), The Late Prehistoric port-of-trade of Khao Sek (Chumphon province, Thailand) and its implications for the understanding of the early trade polities in Maritime Southeast Asia, Archaeological Research in Asia (Vol. 13, pp. 25–36). Dussubieux, L., & Soedewo, E. (2018). The glass beads of Kampai Island, Sumatra. Archaeological and Anthropological Sciences, 10(5), 1129–1139. Dussubieux, L., & Bellina, B. (2017). Glass from an early Southeast Asian producing and trading centre. In B. Bellina (Ed.), Khao Sam Kaeo: A late prehistoric early port-city between the Indian Ocean and the South China Sea (pp. 549–585). Mémoires Archéologiques 28. Paris: Écoles Française d’Êxtrême-Orient. Dussubieux, L., & Pryce, T. O. (2016). Myanmar’s role in Iron Age interaction networks linking Southeast Asia and India: Recent glass and copper-base metal exchange research from the Mission Archéologique Française au Myanmar. Journal of Archaeological Science—Reports, 5, 598–614. Dussubieux, L. (2014). Compositional analysis of ancient glass fragments from Si Pamutung, North Sumatra. In D. Perret, & H. Surachman (Eds.), History of Padang Lawas, North Sumatra. I: The site of Si Pamutung (9th c.–13th c.) (pp. 379–400). Paris: Cahiers d’Archipel. Dussubieux, L., & Allen, J. (2014). Compositions of glass artifacts from Malaysia: New data from the sites of Pengkalan Bujang and Sungai Mas. In D. Perret, & Z. B. Jaafar. (Eds.), Ancient glassware in Malaysia: The Pengkalan Bujang collection (pp. 119–161). Kuala Lumpur: Department of Museums Malaysia. Dussubieux, L., & Kanungo, A. K. (2013). Trace element analysis of glass from Kopia. In A. K. Kanungo (Ed.), Glass in ancient India: Excavations at Kopia (pp. 360–366). Thiruvananthapuram: KCHR. Dussubieux, L., Lankton, J., Bellina-Pryce, B., & Chaisuwan, B. (2012). Early Glass Trade in South and Southeast Asia: New insights from two coastal sites, Phu Khao Thong in Thailand and Arikamedu in South India. In Crossing Borders in Southeast Asian Archaeology. Selected papers from the 13th International Conference of the European Association of Southeast Asian Archaeologists, Berlin, 2010 (Vol. 2, pp. 307–328). Singapore: NUS Press. Dussubieux, L., & Gratuze, B. (2010). Glass in Southeast Asia. In B. Bellina, E. A. Bacus, T. O. Pryce, & C. J. Wisseman (Eds.), 50 years of archaeology in Southeast Asia: Essays in honour of Ian Glover (pp. 247–259). River Books. Dussubieux, L., Gratuze, B., & Blet-Lemarquand, M. (2010). Mineral soda alumina glass: Occurrence and meaning. Journal of Archaeological Science, 37, 1645–1655. Dussubieux, L. (2009). Compositional analysis of ancient glass fragments from North Sumatra, Indonesia. In D. Perret, & H. Surachman (Eds.), Histoire de Barus III: Regards sur une

508

L. Dussubieux

place marchande de l’océan Indien (XIIe-milieu du XVIIe s.) (pp. 385–417). Paris: Association Archipel/EFEO. Dussubieux, L., Kusimba, C. M., Gogte, V. D., Kusimba, S. B., Gratuze, B., & Oka, R. (2008). The trading of ancient glass beads: New analytical data from South Asian and East African soda-alumina glass beads. Archaeometry, 50(5), 797–821. Dussubieux, L., & Gratuze, B. (2003). Origine et diffusion du verre dans le monde indien et en Asie du Sud-est : l’importance du dosage des éléments-traces. Revue D’archéométrie, 27, 67–73. Dussubieux, L. (2001). L’Apport de l’ablation laser couplée à l’ICP-MS à l’étude du verre archéologique de l’Océan Indien. PhD Thesis. Orléans: Université d’Orléans. Dussubieux, L., & Wood, M. (2021). Indian glass: Chronology and distribution in Eastern Africa. In A. K. Kanungo, & L. Dussubieux (Eds.), Ancient glass of South Asia—Archaeology, ethnography and global connection. Singapore/Gandhinagar: Springer Nature / IIT Gandhinagar. Dussubieux, L. (Unpublished report). LA-ICP-MS analysis of glass beads from the Indian Ocean, Sealinks Project. Field Museum. Federeci, C. (available online). The Voyage and Travaile: Of M. Caesar Frederick, Merchant of Venice, Into the East India, the Indies, and Beyond, Wherein are contained very pleasant and rare matters, with the customes and rites of those countries. Also, Herein are Discovered the Merchandises and Commodities of those Countreyes, aswell the Aboundance of Goulde and Silver, as Spices, Drugges, Pearles, and Other Jewelles. London: Richard Jones and and Edward White, 18 June 1588, https://quod.lib.umich.edu/e/eebo/A00611.0001.001?rgn=main;view=ful ltext Francis, P. Jr. (2013). The beads. In J. Carswell, S. Deraniyagala, & A. Graham (Eds.), Mantai by the sea (pp. 349–369). Linden Soft Verlag. Francis, P., Jr. (2002). Asia’s maritime bead trade: 300 BC to the present. University of Hawai’i Press. Francis, P., Jr. (1991). Beadmaking at Arikamedu and beyond. World Archaeology, Craft Production and Specialization, 23(1), 28–43. Francis, P. Jr. (1990). Glass beads in Asia, part two. Indo-Pacific beads Asian Perspectives, 29(1), 1–23. Gill, M. S. (2017). A single Ingredient for primary glass production: Reassessing traditional glass manufacture in Northern India. Journal of Glass Studies, 59, 249–259. Gogte, V., Pradhan, S., Dandekar, A., Joshi, S., Nanji, R., Kadgaonkar, S., & Marathe, V. (2006). The Ancient port of Chaul. Journal of Indian Ocean Archaeology, 3, 62–80. Gratuze, B., Dussubieux, L., & Bopearachchi, O. (2000). Etude de perles de verre trouvées au Sri Lanka, IIIe s. av. - IIe s. apr. JC. In Annales du 14e congrès de l’AIHV, Italia\Venezia-Milano 199 (pp. 46–50). Amsterdam: Association Internationale pour l’Histoire du Verre. Guillot, C., & S.Ch. Wibisono 1998. Le verre à Lobu Tua, Etude préliminaire. In C. Guillot (Ed.), Histoire de Barus, Sumatra. Le site de Lobu Tua. I : Etudes et Documents (pp. 189–206). Paris: Cahiers d’Archipel 30. Guillot, C., Surachman, H., Perret, D., et al. (2003). Histoire de Barus. Le site de Lobu Tua. II : Etude archéologique et Documents. Paris : Cahiers D’archipel, 30, 223–274. Hall, M., & Yablonsky, L. (1998). Chemical analyses of Sarmatian glass beads from Pokrovka, Russia. Journal of Archaeological Science, 25, 1239–1245. Henderson, J., An, J., & Ma, H. (2018). The archaeometry and archaeology of ancient Chinese glass: A review. Archaeometry, 60(1), 88–104. Huang Qishan 1996. Discovery and research of ancient glass utensils in Guangxi. In Proceedings of XVII international congress on glass 1995, glass technology, glass archeometry (Vol. 6, pp. 481– 485). Beijing: Chinese Ceramic Society, 1995. Kanungo, A. K., Misra, V. N., Dutta, K., Ravi Prasad, G. V., Yadava, M. G., & Hodgins, G. W. L. (2010). The radiocarbon chronology of Kopia, an early glass manufacturing centre in India. Archaeometry, 52(5), 899–918. Kanungo, A. K. (2004). Glass beads in ancient India and furnace-wound beads at Purdalpur: An ethnoarchaeological approach. Asian Perspectives, 43(1), 123–150.

Indian Glass in Southeast Asia

509

Koesuka, T., & Yamasaki, K. (1996). Chemical compositions of ancient glasses found in Japan—A historical survey. In Proceedings of XVII international congress on glass 1995, glass technology, glass archeometry (Vol. 6, pp. 469–474). Beijing: Chinese Ceramic Society. Kock, J., & Sode, T. (1995). Glass, glass beads and glassmakers in Northern India. Vanlose, Danmark: THOT Print. Lankton, J. W., Rajan, K., & Gratuze, B. (2014). Glass from South India: New analyses January 2012, Appendix E. In K. Rajan, & V. P. Yathees Kumar (Eds.), Archaeology of Amar¯avathi River Valley: Porunthal excavations (Vol. II). Delhi: Sharada Publishing. Lankton, J. W, & Dussubieux, L. (2013). Ancient glass from Southeast Asia. In K. Janssens (Ed.), Modern methods for analysing archaeological and historical glass (pp. 431–441). Wiley. Lankton, J. W., Dussubieux, L., & Gratuze, B. (2008). Glass from Khao Sam Kaeo: Transferred technology for an early Southeast Asian exchange network. In Recent advances in the archaeology of exchange in the upper Thai-Malay Peninsula. Bulletin de l’Ecole Française d’Extrême-Orient (Vol. 2006, No. 93, pp. 317–351.) Lankton, J. W., & Dussubieux, L. (2006). Early glass in Asian maritime trade: A review and an interpretation of compositional analyses. Journal of Glass Studies, 48, 121–144. Lee, I.-S., Brill, R. H., & Fenn, P. (1993). Chemical analyses of some ancient glasses from Korea. In Annales du 12th Congres de l’Association Internationale pour l’Histoire du Verre (pp. 163–174). Amsterdam: Association Internationale pour l’Histoire du Verre. Li, F. (1999). Elemental and isotopic analysis of ancient Chinese glass. Master Thesis. Hong Kong: University of Hong Kong. Lin, Y.-X., Rerhen, T., Wang, H., Ren, X.-Y., & Ma, J. (2019). The beginning of faience in China: A review and new evidence. Journal of Archaeological Science, 105, 97–115. McKinnon, E. E. (1984.) Kota Cina: Its context and meaning in the trade of Southeast Asia in the twelfth to fourteen centuries. PhD Thesis. Ithaca, NY: Cornell University. Mirti, P., Pace, M., Negro Ponzi, M. M., & Aceto, M. (2008). ICP-MS analysis of glass fragments of Parthian and Sasanian epoch from Seleucia and Veh Ardaš¯ır (Central Iraq). Archaeometry, 50(3), 429–450. Mirti, P., Pace, M., Malandrino, M., & Negro Ponzi, M. M. (2009). Sasanian glass from Veh Ardaš¯ır. Journal of Archaeological Science, 36(4), 1061–1069. Perret, D., & Jaafar, Z. (Eds.). (2014). Ancient glassware in Malaysia, The Pengkalan Bujang Collection. Jabatan Muzium Malaysia. Selvakumar, V. (2021). History of glass ornaments in Tamil Nadu, South India: Cultural perspectives. In A. K. Kanungo, & L. Dussubieux (Eds.), Ancient glass of South Asia—Archaeology, ethnography and global connection. Singapore/Gandhinagar: Springer Nature / IIT Gandhinagar. Seman, S., Dussubieux, L., Cloquet, C., & Pryce, T. O. (2021). Strontium isotope analysis in ancient glass from South Asia using portable laser ablation sampling. Archaeometry, 63(11), 88–104. Shortland, A. J., Schachner, L., Freestone, I. C., & Tite, M. (2006). Natron as a flux in the early vitreous materials industry: Sources, beginnings and reasons for decline. Journal of Archaeological Science, 33, 521–530. Shortland, A. J., Kirk, S., Eremin, K., Degryse, P., & Walton, M. (2018). The analysis of Late Bronze Age glass from Nuzi and the question of the origin of glass-making. Archaeometry, 60(4), 764–783. Soedewo, E. (2013). Perkembangan Penelitian Kepurbakalaan di Pulau Kampai Sumatera Utara. Archipel, 86, 131–154. Veraprasert, M. (1987). Khlong Thom: An ancient bead and manufacturing location and ancient entrepot. In Seminar in Prehistory of Southeast Asia (SPAFA Final Report) (pp. 323–31). Bangkok: SEAMEO Project in Archaeology in Fine Arts. Wadia, D. N. (1975). Geology of India (4th ed.). Graw Hill Publishing Co. Wheeler, R. E. M., Ghosh, A., & Deva, K. (1946). Arikamedu: An Indo-Roman Trading station on the east coast of India. Ancient India, 2, 17–124.

510

L. Dussubieux

Zhang, B., Li, Y. H., Ma, B., Gan, F. X., Zhang, Z. Q., Cheng, H. S., & Yang, F. J. (2004). Nondestructive analysis of early glass unearthed in south China by external-beam PIXE. Journal of Radioanalytical and Nuclear Chemistry, 261(2), 387–392.

Indian Glass: Chronology and Distribution in Eastern Africa Laure Dussubieux and Marilee Wood

Abstract The elemental analysis via laser ablation–inductively coupled plasma– mass spectrometry (LA-ICP-MS) of ancient glass beads from archaeological sites located in Kenya, Tanzania and the island of Mayotte revealed the presence of two different types of mineral soda–high alumina (m-Na-Al) glasses. One of them is the already known m-Na-Al 2 glass type that likely originated from the west coast of India and that is associated with later sites dating from the end of the fourteenth century CE and onward. A new type of glass was identified: m-Na-Al 6. It appears at earlier sites dating from the ninth to the thirteenth century CE. Although the composition points towards an Indian origin for the glass, a specific region of provenance cannot be proposed at this time. The m-Na-Al 2 glass was almost exclusively used to manufacture drawn beads. Drawn beads were also manufactured from m-Na-Al 6 glass, but a significant proportion of them (~25%) is wound. Based on morphology, compositions and chronology, it seems that the Khami bead series from southern Africa and the m-Na-Al 2 beads from the east coast of Africa are a same type of beads. The m-Na-Al 6 glass beads are similar with the contemporaneous K2 and East Coast Indo-Pacific beads series found in southern Africa but more wound beads in this glass group are found on the east coast of Africa compared to southern Africa.

1 Introduction Efforts to study glass beads found in Sub-Saharan Africa remain unevenly distributed with western and southern Africa getting most of the attention. Recent discoveries of an indigenous glassmaking industry in the Ile-Ife region in Nigeria have revealed the development of a local recipe taking advantage of the availability in the area of a pegmatite sand with high alumina concentrations that was mixed with high lime L. Dussubieux (B) Field Museum, Chicago, USA e-mail: [email protected] M. Wood University of the Witwatersrand, Johnnesburg, South Africa e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_21

511

512

L. Dussubieux and M. Wood

snail shells. This created a high lime–high-alumina (HLHA) glass that was traded all over western Africa during the first half of the second millennium CE (Babalola et al., 2018a, b; McIntosh et al., 2017; Ogundiran & Ige, 2015; Lankton et al., 2006). Southern Africa is another region that is of great importance for glass bead studies in Africa as thanks to the work of Wood (2011a) and Robertshaw et al. (2010), the glass beads of this region have been thoroughly described, and their elemental compositions characterized, providing chronological and distribution information extremely useful to recreate ancient exchanges within this region and beyond. Among the glass beads recovered in southern Africa, some have a mineral soda– high-alumina (m-Na-Al) composition that most likely originated from South Asia (Robertshaw et al., 2010). They were all manufactured using the drawn technique. Wood (2011b), following Francis’s naming system (Francis, 1990), called them ‘Indo-Pacific’ (IP) and separated them into three main series: ‘K2 Indo-Pacific’ (K2 IP), ‘East Coast Indo-Pacific’ (EC IP) and ‘Khami Indo-Pacific’ (Khami IP). The K2 IP series beads are fairly small and exclusively transparent to translucent blue green or green. They correspond to a period ranging from about 980–1200 CE. The EC IP beads are roughly contemporaneous, being present from about 1000–1250 CE, but they are typologically different. They are slightly larger in diameter but shorter in length, and they exhibit a larger range of mostly opaque colours, including green, blue green, black, brownish-red, orange and yellow. The Khami IP series appears much later (1430–1650 CE), and the beads are larger than in the previous drawn series. Colours, which are opaque, include black, a larger range of blues including cobalt blue, brownish-red, green, orange, yellow and white. Differences in composition between the K2 IP series and the EC IP series are fairly insignificant, but these two groups can be distinguished chemically from the Khami IP series due to higher concentrations of soda, magnesia, lime and uranium, while the Khami IP series has lower alumina concentrations (Robertshaw et al., 2010). The goal of this article is to build on previous research conducted in southern Africa and to increase our knowledge of the composition, consumption and exchange patterns, chronological contexts and provenance of m-Na-Al glass beads found on the eastern coast of Africa. A large sample of glass beads from different sites in Tanzania, Kenya and Mayotte was analysed using laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS). The study found that 238 beads had mNa-Al compositions that fell into two subgroups. These samples were compared to already published data recovered previously from southern Africa. These new results improve the glass chronology on the eastern coast of Africa and provide a better understanding of glass trade in this region. They also add to our knowledge of glass production in South Asia.

2 Background In antiquity, India was one of the major glass bead producers in the world (perhaps even the most important one). Industrial quantities of beads were manufactured using

Indian Glass: Chronology and Distribution in Eastern Africa

513

a very efficient technique, known as the lada process, which created drawn tubes that were broken or cut into bead lengths once the glass was cool. The resultant segments were reheated to obtain rounded monochrome beads (see Kanungo, 2016; Kock & Sode, 1995; Francis, 1991; Stern, 1987 for an ethnographic account of the production of lada drawn beads in India). The significance of the production is apparent in the large quantities of Indian beads found within South and Southeast Asia (e.g. Carter, 2015; Dussubieux & Bellina, 2018; Dussubieux & Pryce, 2016) as well as all around the India Ocean, in China (e.g. Jiayao, 2000), Eastern Asia (e.g. Nakai & Shirataki, 2016) and as far west as Northern Africa (e.g. Then-Obłuska & Dussubieux, 2016) and Europe (e.g. Pion and Gratuze, 2016). The glass used for the production of these beads was locally made and had a very specific soda-based composition, with high-alumina concentrations. It was named mineral soda–high alumina or m-Na-Al glass. The high soda–high-alumina composition of this Indian glass can be explained by the use of rather immature sands with compositions very close to that of the granite from which it derives and which contain a relatively high proportion of feldspar. These sands also contain high concentrations of a range of trace elements, including titanium, zirconium, the rare earth elements and uranium. High concentration of sodium in a glass is generally due to the addition of soda, found either in mineral form or produced from the ashes of burnt halophytic plants (which grow in soils containing saline water). Ethnographic data as well as scientific experiments (Brill, 2003; Gill, 2017; Kock & Sode, 1995) indicate that in several parts of India, glassmakers were using a sand naturally mixed to sodic efflorescence called reh containing large amounts of sodium salts (carbonate, bicarbonate and sulphate) and varying proportion of calcium and magnesium salts. It occurs in areas where river-draining mountains contain dissolved salts that percolate through the subsoil until saturation. Rains dissolve these salts which travel by capillary action upward through the soil during the dry season and form white efflorescence on the surface (Wadia, 1975: 489, 501, 502). They occur in arid or semi-arid regions. Although so far, four subgroups of m-Na-Al glasses have been identified (Dussubieux et al., 2010), mostly two of them, m-Na-Al 1 and 2, have been found in Africa. A fifth m-Na-Al glass will be described below for the first time. The m-Na-Al 1 glass, likely from Sri Lanka and/or South India, was used to manufacture small drawn beads starting at least in the third century BCE. In Africa, this glass type is fairly rare. Current research has recorded the presence of a significant quantity of m-Na-Al 1 beads at only one site, Unguja Ukuu, seventh to early eleventh century, located on the Island of Zanzibar (Wood et al., 2017). All other known sites with m-Na-Al 1 glass yielded only one bead each, Ungwana, ninth–sixteenth century, on the Kenyan coast (Dussubieux et al., 2008: 814), and Mahilaka, ninth–sixteenth century in north-west Madagascar (Robertshaw et al., 2006). The m-Na-Al 2 glass that was found at Chaul in Maharashtra (ninth–nineteenth century CE) (Dussubieux et al., 2008, 2021) has also been identified in southern Africa (e.g. Robertshaw et al., 2010; Wood et al., 2009), on the East Coast of Africa at the Kenyan site of Mtwapa, ninth–nineteenth century CE (Dussubieux et al., 2008), and at Songo Mnara (mainly fifteenth century) in Tanzania (Wood, 2016, 2019).

514

L. Dussubieux and M. Wood

Mahilaka in Madagascar also yielded m-Na-Al 2 glass beads (Robertshaw et al., 2006). To this list, we will now add a new m-Na-Al glass type (m-Na-Al 6) which has been identified among the glass beads from a range of sites along the eastern coast of Africa and Mayotte.

3 Samples The samples were obtained from recent excavations in Kenya, Tanzania and Mayotte and from the Fort Jesus Museum that housed material that has been excavated in Kenya for decades. The sites appear on Fig. 1. A brief description of each site is given below. The Kenyan site of Mtwapa was a coastal urban settlement located 15 km north of Mombasa. The site was inhabited by diverse types of populations including traders, fishermen and craftspeople (Kusimba et al., 2018a). The glass beads belong to the tenth–eighteenth century period (Table 1). The compositions for the beads from this site were described in Dussubieux et al. (2008). Gede is located 15 km south of Malindi on the Kenyan coast. It was a walled town founded in the twelfth century, including two great mosques and six smaller mosques, Fig. 1 Map indicated the location of the sites appearing in Table 1

Indian Glass: Chronology and Distribution in Eastern Africa

515

Table 1 List of sites included in this study with dating and number of beads taken into consideration Country or region

Site

Chronology (all CE)

Number of m-Na-Al 2 or 6 beads

Kenya

Mtwapa

10th–18th c

42

Ungwana

9th–16th c

5

Takwa

16th–17th c

6

Manda 1

7th–16th c

49

Manda 2

mid-11th–late 13th c

4

Gede

12th–17th c

34

Tanzania

Mayotte

Fort Jesus

16th c. and onward

3

Songo Mnara

Late 14th–early 16th c

12

Juani Primary School, Mafia Island

880–1200

22

Antsiraka Boira

12th–13th c

61

a number of tombs, a palace and close to 300 private houses (Pawlowicz, 2018; Pradines, 2010). Gede was temporarily abandoned in the early sixteenth century and then briefly re-occupied at the end of the sixteenth century up to the beginning of the seventeenth century, before being permanently abandoned. The beads included in this study were excavated by Kirkman (1963). Manda was a settlement in Kenya’s Lamu Archipelago that was active from 600 to 1500 CE. Its location is close to the mainland with an easy access to the open ocean. Archaeological evidence shows that the site was engaged in regional and international trade with material from the Persian Gulf, the Indian Ocean and the South China Sea (Kusimba et al., 2018b). Most of the glass beads included in this study (Manda 1) were recently excavated (Kusimba et al., 2018b). A small number of additional beads from Manda come from the work of Chittick (1984) who assigned most of the glass beads (83.2%) to the mid-eleventh to the late thirteenth century (Manda 2). Takwa in Kenya is a walled Swahili stone settlement on Manda Island in the Lamu Archipelago that was occupied during the sixteenth and seventeenth century (Wilson, 1976, 1980). Ungwana, located 150 miles north of Mombasa, at the mouth of the Tana River, was a market place where ivory brought from inland and imported goods from India and the Middle East were exchanged. The beads used in this study came from Kirkman’s excavations (Kirkman, 1966). Fort Jesus is a Portuguese era fortress located on Mombasa Island. It was built in 1505 by the Portuguese who occupied it for two centuries until they were defeated by Omani Arabs. The Omanis used it as a base for conducting commerce until the early twentieth century (Kirkman, 1974). Songo Mnara is located in the Kilwa archipelago on the southern coast of Tanzania. It was occupied for a short period of time from the late fourteenth century to the

516

L. Dussubieux and M. Wood

early sixteenth. The site includes private houses, five mosques and numerous tombs (Wynne-Jones & Fleisher, 2010, 2011). The glass beads from the Juani Primary School site come from recent excavations (Crowther et al., 2016). This site also known as Kisimba Jumbe is located on a small island that is part of the Mafia Archipelago, off the coast of Tanzania. Occupation of the site includes two phases with an early period ranging from 385 to 540 CE (Early Iron Age or EIA) and a later one from 880 to 1200 (Middle to Late Iron Age or MIA–LIA). Most of the glass beads were found in the MIA–LIA layers along with small quantities of glass fragments and imported ceramics. The site of Antsiraka Boira is located on the island of Mayotte which is situated on the north side of the Mozambique Channel between the coast of Mozambique and Madagascar. The beads were excavated from a necropolis dating between the twelfth and thirteenth century. The funerary architecture of this necropolis bears a visible Muslim influence. More than 10,000 beads were found, most of them being made of glass. Drawn beads are the most common, but wound beads, mostly black in colour, are also present (Pauly & Jacquot, 2014). In addition to our LA-ICP-MS analysis of some beads from this site, additional beads were analysed recently with X-ray fluorescence and Raman spectroscopy (Fischbach et al., 2016).

4 Analytical Results Analyses were conducted at the Elemental Analysis Facility at the Field Museum between 2005 and 2017. Before 2016, an Analytik Jena ICP-MS was used. In 2016, a new Thermo ICAP–Q ICP–MS was installed. The mass spectrometers were connected to a New Wave UP213 laser for direct introduction of solid samples. More details about the analytical technique and the performance of our instrumentation can be found in Dussubieux (2021) and Dussubieux et al. (2009).

4.1 The M-Na-Al 6 Glass Group Previously, a strong correspondence between the m-Na-Al 2 composition of some glass samples from Chaul in India and from Mtwapa was identified showing that some glass beads found on the eastern coast of Africa were quite likely imported from the western coast of India, possibly through the port of Chaul (Dussubieux et al., 2008). Further analysis of glass beads coming from different sites located on the eastern coast of Africa showed that another type of m-Na-Al glass, as yet unidentified, was present in the region. The first site where this new composition was identified is Juani Primary School on Mafia Island (Crowther et al., 2016). Working with the same elements that were used to characterize four of the m-Na-Al glass types that were present around the Indian Ocean: MgO, CaO, Sr, Zr, Cs, Ba and U (Dussubieux et al., 2010), principal component analysis (PCA) was carried out using the Gauss 8.0

Indian Glass: Chronology and Distribution in Eastern Africa

517

routine. Samples belonging to glass groups m-Na-Al 1 to 4 and samples from Juani Primary School were included. Before PCA, the composition of each sample was re-calculated taking into account only silica, magnesia, soda, lime, potash and iron to eliminate the dilution effect introduced by the presence of high concentrations of colouring elements (e.g. copper, tin and lead), as suggested in Brill (1999). Figure 2 shows that the samples from Juani Primary School are clustering away from all the other glass samples due to a slightly different elemental signature (Table 2). This new variation of the m-Na-Al composition was named m-Na-Al 6. The m-Na-Al 6 glass is characterized by uranium concentrations in the range of 60 ppm much higher than in the m-Na-Al 1 (9 ppm in average) glass but lower than in the m-Na-Al 2 and 4 glasses (more than 100 ppm). They are fairly similar to that of the m-Na-Al 3 glass, but the two glasses have different Cs concentrations. For this element, concentrations are around 3 ppm in the m-Na-Al 3 glass and ~1.5 ppm in the m-Na-Al 6 glass. The concentrations of Cs in the m-Na-Al 2 glass are lower ( 4%) were detected, others contained extra manganese, while no specifically high concentrations of any elements were found in the third set (Fig. 4). Black beads with high iron and those with no specifically high colouring element concentrations could be found in both types of glass, but those with high manganese

3 0.9 ± 0.2 2.5 ± 0.7 186 ± 40 172 ± 80 0.6 ± 0.2 314 ± 68 158 ± 59

0.9 ± 0.3

3.2 ± 1.0

191 ± 18

157 ± 28

1.1 ± 0.4

326 ± 49

322 ± 27

Number of samples

MgO %

CaO %

Sr ppm

Zr ppm

Cs ppm

Ba ppm

U ppm

Fort Jesus

Antsiraka Boira

3

m-Na-Al 2

108 ± 52

324 ± 84

0.6 ± 0.3

146 ± 46

207 ± 47

3.5 ± 0.9

1.3 ± 0.5

22

Gede

118 ± 65

276 ± 128

0.8 ± 0.6

159 ± 58

224 ± 65

3.8 ± 1.3

1.1 ± 0.3

11

Manda 1

112 ± 82

369 ± 129

0.8 ± 0.3

155 ± 50

221 ± 84

3.5 ± 0.8

1.3 ± 0.4

42

Mtwapa

124 ± 79

226 ± 64

0.8 ± 0.2

152 ± 38

202 ± 75

3.3 ± 1.1

1.0 ± 0.3

12

Songo Mnara

Table 3 Average concentrations with standard deviations for some key elements for each site with m-Na-Al 2 glass

82 ± 34

377 ± 151

0.5 ± 0.2

147 ± 32

230 ± 41

3.7 ± 1.0

1.2 ± 0.4

6

Takwa

168 ± 85

316 ± 133

0.9 ± 0.7

154 ± 47

134 ± 17

2.5 ± 1.0

0.6 ± 0.2

5

Ungwana

Indian Glass: Chronology and Distribution in Eastern Africa 519

520

L. Dussubieux and M. Wood

Table 4 Average concentrations with standard deviations for some key elements for each site with m-Na-Al 6 glass m-Na-Al 6

Antsiraka Boira

Gede

Juani Primary School

Manda 1

Manda 2

Number of samples

58

12

22

38

4

MgO %

0.7 ± 0.2

0.7 ± 0.2

0.8 ± 0.2

0.6 ± 0.2

0.9 ± 0.2

CaO %

3.2 ± 1.0

2.7 ± 0.6

2.2 ± 0.6

2.9 ± 0.8

3.8 ± 1.7

Sr ppm

225 ± 65

237 ± 49

235 ± 88

275 ± 87

390 ± 163

Zr ppm

316 ± 138

285 ± 57

216 ± 35

305 ± 111

278 ± 43

Cs ppm

1.1 ± 0.5

1.0 ± 0.3

1.5 ± 0.4

1.1 ± 0.3

1.4 ± 0.9

Ba ppm

440 ± 112

473 ± 167

401 ± 166

456 ± 145

442 ± 136

U ppm

67 ± 41

78 ± 53

57 ± 22

96 ± 49

59 ± 27

Fig. 4 Concentrations of MnO and Fe2 O3 in black glass beads in the m-Na-Al 2 and 6 groups

concentrations are found only in the m-Na-Al 6 glass. In this group, there is no correlation between the technology used (wound or drawn) and the colouring agent. Amber beads also contain higher concentrations of iron (from 3.6 to 7%) but were found only among the m-Na-Al 6 glass type. Another colour only present in m-Na-Al 6 glass is colourless. These colourless glass beads do not contain any decolouring elements, whereas manganese or antimony were used to create colourless glass in antiquity. Iron can naturally impart a range of colours to glass, and thus, the level of this element is sometimes lower in colourless glass. In the m-Na-Al 6 colourless glass, the concentration of iron is fairly similar to what it is in the other samples. This suggests that a careful control of the atmosphere of the furnace rather than a specific recipe was used to obtain the colourless glass. Dark blue beads can be found in both glass groups. One way to produce dark blue is to add small quantities of cobalt to the glass. Three glass beads with an m-Na-Al 2 composition contain significant concentrations of cobalt (~500 ppm and more). [For more information on elements associated with cobalt in m-Na-Al 2 glass, see

Indian Glass: Chronology and Distribution in Eastern Africa

521

Dussubieux et al., (2008).] Other glass samples belonging to both groups quite likely owe their dark blue colour to the presence of copper or a combination of copper and iron. The two dark blue beads in the m-Na-Al 6 group with copper contain elevated concentrations of iron (3.3 and 3.8%). For the other blue beads in this group and for one bead in the m-Na-Al 2 glass group, no colouring agent seems to have been added: iron concentrations are in their usual range, and there are no significantly high concentrations of cobalt or copper in these beads (Fig. 5). Copper is also involved in the coloration of the turquoise blue and red glass samples from both groups. As Fig. 6 demonstrates, we notice more iron in most of the red glass beads from both groups, an observation that has previously been reported for m-Na-Al 2 red glass beads (Dussubieux et al., 2008). From a general point of view, there are more beads with slightly higher copper concentrations in the m-Na-Al 6 group compared to the m-Na-Al 2 glass group. In comparing the concentrations of lead in the red glasses from the two groups, we found that this element is often present in slightly higher concentrations in the m-Na-Al 6 red glass than the m-Na-Al 2 red glass (Fig. 7). It is often difficult to assess the degree of opacity of the green and yellow beads, which can be fully opaque or translucent. This opacity is quite likely due to the presence of lead stannate (PbSnO3 ) which creates yellow crystals (Rooksby, 1964). In a turquoise blue matrix (due to the presence of copper in the glass), the yellow crystals give an opaque green appearance to the glass (Fig. 8). White beads are not very common in m-Na-Al glasses, and only one bead with such a colour is part of the m-Na-Al 2 glass group, while three white beads are part of the m-Na-Al 6 glass group. No element in specifically high concentrations is present in these beads. Iron also has a fairly typical concentration when compared with other beads. Perhaps an observation of the structure of the glass could explain its white colour. Orange beads are totally absent from the m-Na-Al 6 group. The m-Na-Al 2 group contains several beads with an orange-pinkish colour with a waxy appearance. The colour of these beads looks rather like coral. These beads contain 1.2–1.5% of Fig. 5 Concentrations of CuO and Co in dark blue glass beads in the m-Na-Al 2 and 6 groups

522

L. Dussubieux and M. Wood

Fig. 6 Concentrations of Fe2 O3 and CuO in the red (r) and turquoise blue (TB.) m-Na-Al 2 and 6 beads

Fig. 7 Concentrations of CuO and PbO in the red m-Na-Al 2 and 6 beads

SnO2 , 0.3% of ZnO and close to 6% of PbO. They do not contain any significant concentrations of copper, and their iron concentrations are 1 to 1.7% (as Fe2 O3 ).

Indian Glass: Chronology and Distribution in Eastern Africa

523

Fig. 8 Concentrations of SnO2 and PbO in the yellow and green m-Na-Al 2 and 6 beads

5 Discussion 5.1 Typology and Technology Figures 9 and 10 illustrate some beads belonging to the m-N-Al 2 and the m-Na-Al 6 glass groups. Examining glass colours has made it possible to show that different colouring techniques were used in the two glass groups. For example, the absence of colourless, amber and manganese-coloured black beads in the m-Na-Al 2 glass group has been noted, while cobalt-blue and coral-looking beads are unknown in the m-Na-Al 6 group. However, it is difficult to discuss the proportions of the different colours in the two groups as the sample studied here is not representative of what was found at the sites. This lack of representativeness is also acknowledged for the bead size and manufacturing techniques. Such details would need to be confirmed in the future with comprehensive studies. Another difference between the two glass groups is evident when considering the technique used to create the beads. Figure 11 illustrates percentages of beads that were clearly either drawn or wound. Only a few beads were not made using either of these techniques. For example, a bead found in Ungwana with a flower shape was probably moulded. In Fig. 11, it appears that wound beads are extremely rare in the m-Na-Al 2 glass group. In contrast, a little more than 25% of the m-Na-Al 6 beads were shaped using this technique. It is important to emphasize that wound beads require more time to make because the glass worker must create each bead individually. Kanungo (2004a) reports the presence of both lamp and furnace wound bead traditions in Varanasi and Purdalpur (Uttar Pradesh); however, it is not known when the tradition started in that region. This technique usually produces larger beads. In contrast, the process of drawing beads, as developed in India, is a form of mass production that can create thousands of beads in a short period of time. It typically produces beads less than 6 mm in diameter, sometimes even as small as

524

L. Dussubieux and M. Wood

Fig. 9 Glass beads belonging to the m-Na-Al 2 glass group. a Fort Jesus, b Gede, c Manda1, d Takwa, e Mtwapa, f Ungwana

1 mm. Of interest in this discussion (and perhaps relevant to searching for the beads’ origins) is Kanungo’s (2004b) observation that wound beads are found mainly in north India, while Indo-Pacific beads are found mostly in the south and the Deccan.

5.2 Chronological Implications Despite the long time periods of the contexts in which most of the beads occur, trends are evident. Beads made of m-Na-Al 2 glass were the only Indo-Pacific type found at Mtwapa (tenth–eighteenth century), Ungwana (ninth–sixteenth century), Takwa (sixteenth–seventeenth century), Fort Jesus (sixteenth century onward) and Songo Mnara (late fourteenth century—early sixteenth century). Those sites are fairly

Indian Glass: Chronology and Distribution in Eastern Africa

525

Fig. 10 Glass beads belonging to the m-Na-Al 6 glass group. a Manda2, b Gede, c Manda 1

recent. Two sites, both older, Juani Primary School, (820–1200 CE) and Manda (mid11th to late thirteenth century, based on the Manda 2 sample group), produced beads made of m-Na-Al 6 glass. Thus, this study indicates that the two types of beads were traded sequentially. More data would be necessary to fully understand the transition between the m-Na-Al 6 to the m-Na-Al 2 beads. Did the trade of the two glass types overlap? Did one glass type replace the other one immediately? Did a gap occur between the trades of the two different glass types? It appears highly unlikely that the m-Na-Al 2 glass beads at Mtwapa and Ungwana arrived in the ninth–tenth century since it would be unusually early compared to the other sites. Thus, this study provides an opportunity to assess the occurrence of the m-Na-Al 2 glass beads at these sites under a new light. Based on the dating of the other sites with m-Na-Al 2 beads, it seems that they appeared around the late fourteenth century at the earliest.

526

L. Dussubieux and M. Wood

Fig. 11 Distribution of the techniques used to manufacture each bead for the m-Na-Al 2 and 6 glass groups

These observations suggest that Mtwapa and Ungwana were not strongly connected to the Indian Ocean trading network when m-Na-Al 6 beads were being traded to the eastern coast of Africa. Sometime around the end of the fourteenth century or later, they became part of a network that included Takwa, Fort Jesus and Songo Mnara. For sites with the two types of glass (Antsikara Boira, Gede and Manda 1), it seems probable that the m-Na-Al 6 beads are associated with the earliest part of the site and the m-Na-Al 2 beads with the later one. For these sites, the connection with the Indian Ocean bead network apparently occurred relatively early, and the sites remained connected through time despite changes in glass sources. It is important to note that at Antsikara Boira 95% of the assemblage is made up of m-Na-Al 6 glass and at Manda 1 the proportion of m-Na-Al 6 beads is 79%. This contrasts with Gede where m-Na-Al 6 beads account for only 35%. An explanation for these differences likely lies in the dating of the three sites, with Antsikara Boira and Manda 1 having been occupied somewhat earlier than Gede. Other evidence will have to be considered to refine these observations. For a more accurate and complete interpretation of the trade connections of these sites that is beyond the scope of this article, it would be necessary to take into account the entire bead assemblage from each site, which would often include more than just m-Na-Al 2 and 6 beads. As far as South Asian glass in Africa is concerned, there is now a chronology starting around the seventh century that extends until possibly the nineteenth century. It begins with the m-Na-Al 1 glass found mostly at Unguja Ukuu, a site on the Island of Zanzibar dating from the seventh to the eleventh century. This glass seems to have a limited diffusion in Eastern Africa. Next, beads of m-Na-Al 6 glass appear for periods dating from the ninth to the thirteenth century. By lack of a precise chronology of the bead distribution within each site, it is not yet clear whether the two types overlapped in time and whether the m-Na-Al 6 beads belong only to part of the ninth to thirteenth century time period. This glass type was found at a number of sites, suggesting a wider diffusion due, perhaps, to more intense exchanges with

Indian Glass: Chronology and Distribution in Eastern Africa

527

India. Finally, a shift to the m-Na-Al 2 glass occurred around the fourteenth century. Like the m-Na-Al 6, the m-Na-Al 2 glass appears at many sites. As more data are obtained from well-dated contexts, the chronology might have to be adjusted slightly in the future.

5.3 Comparison with Southern Africa In Southern Africa, several types of m-Na-Al glass groups have been identified: the K2 Indo-Pacific (K2 IP), East Coast Indo-Pacific (EC IP) and Khami Indo-Pacific (Khami IP) series (Wood, 2011b). In Table 5, the average reduced compositions (see Brill, 1999) of these three glass types are reported for comparison with the average reduced compositions of the m-Na-Al 2 and 6 glasses (Table 6), taking into account SiO2 , Na2 O, MgO, Al2 O3 , K2 O, CaO and Fe2 O3 . According to Wood (2011b) and Robertshaw et al. (2010), the K2 IP and EC IP are roughly contemporaneous and were separated largely on typology. The glass compositions of the two are fairly similar. Thus, we will consider the K2 IP and the EC IP bead series as one group for the purpose of our comparison and refer to it as K2/EC IP. Chronologically, it seems that the m-Na-Al 6 group would overlap with Table 5 Reduced composition of m-Na-Al glass beads found in southern Africa according to Robertshaw et al. (2010)

Table 6 Reduced composition of m-Na-Al glass beads found on the eastern coast of Africa

K2 IP c. 980–1200

East Coast IP c. 1000–1250

Khami IP c. 1430–1650

SiO2

64.51 ± 4.01

63.08 ± 4.73

61.40 ± 4.78

Na2 O

16.2 ± 2.8

14.75 ± 2.29

18.66 ± 3.98

MgO

0.43 ± 0.18

0.59 ± 0.32

1.21 ± 0.55

Al2 O3

11.85 ± 3.09

13.00 ± 3.93

9.81 ± 2.01

K2 O

3.34 ± 0.73

3.46 ± 0.97

2.82 ± 1.13

CaO

2.34 ± 0.49

2.85 ± 0.82

3.39 ± 0.92

Fe2 O3

1.30 ± 0.67

2.27 ± 1.33

2.70 ± 1.37

m-Na-Al 2 late 14th c. and onward

m-Na-Al 6 9th–13th c

SiO2

63.8 ± 3.9

63.2 ± 3.4

Na2 O

18.3 ± 2.8

17.8 ± 2.6

MgO

1.15 ± 0.48

0.80 ± 0.47

Al2 O3

7.91 ± 1.67

9.12 ± 1.47

K2 O

2.55 ± 3.40

3.69 ± 0.94

CaO

3.40 ± 0.93

3.13 ± 1.04

Fe2 O3

2.96 ± 1.66

2.32 ± 1.62

528

L. Dussubieux and M. Wood

the K2/EC IP series whereas the m-Na-Al 2 glass group should overlap with the Khami IP series. The major and minor element concentrations in Tables 4 and 5 suggest a correspondence between the K2/EC IP and m-Na-Al 6 glasses on the one hand and between the Khami IP and m-Na-Al 2 glasses on the other hand. Additional comparisons involving trace elements will be necessary in the future to confirm this initial observation but cannot be undertaken immediately as some key trace elements such as cesium are not available in Robertshaw et al. (2010). It is important to note that although the K2/EC IP and the m-Na–Al 6 beads might have identical compositions, wound beads are common in m-Na-Al 6 beads but do not exist in K2/EC IP series beads. The source of the m-Na-Al 6 glass beads is uncertain at this point as no beads with a similar composition have been identified in India. The presence of a tradition of wound bead manufacturing in Uttar Pradesh could hint at an area of production for the m-Na-Al 6 glass in the northern part of India but would not account for the drawn beads. M-Na-Al 2 glass beads were found at Chaul in Maharashtra (Dussubieux et al., 2008). The travel account of Caesar Frederick, a Venetian merchant, who visited the western part of India in the sixteenth century, mentions glass beads made in ‘Chiawle’ that the Portuguese would bring to ports along the East Coast of Africa (Federeci and Hickock, online). Chiawle has been interpreted as Chaul (Gogte et al., 2006). In addition to this account, the presence of m-Na-Al 2 glass beads at Chaul supports a hypothesis of trade of m-Na-Al 2 beads from the port of Chaul and the possibility of local glass bead production at the same site. Early Portuguese records, on the other hand, note that they procured beads for trade to eastern Africa at the port of Negapattinam in south-east India (Theal, 1898, II:303; Wood et al., 2009), so further research is needed to resolve these questions.

6 Conclusion This study provides new information about glass production in India and glass trade around the Indian Ocean. Previous work established that at least four glass production areas, corresponding to the four m-Na-Al glasses identified in Dussubieux et al. (2010), were active in India in the past. Although this study does not focus on glass found at archaeological sites in India, it provides evidence that a fifth south Asian glass production centre, characterized by an m-Na-Al 6 composition, would have been active between the ninth and thirteenth centuries CE, although this time period might be adjusted in the future as more tightly dated material will be added to the current model. It is possible that production of m-Na-Al 6 glass could have extended beyond these temporal parameters and that trade to eastern Africa involving beads made of this glass might have had a shorter time span than the glass itself. This study also suggests that m-Na-Al 2 glass was produced after the thirteenth century although,

Indian Glass: Chronology and Distribution in Eastern Africa

529

here again, trade and production periods might not have coincided completely. Identifying this additional glass type has added another layer of complexity to our understanding of the Indian glass industry which seems to have relied on different production centres that were active in different regions and periods. They also appear to have had access to Indian Ocean trade to Africa at different periods. This study provides insights into the dynamics of Indian Ocean trade demonstrating shifts in bead procurement patterns from India. Earlier sites (seventh– eleventh century) have m-Na-Al 1 beads. These are quite rare in Africa. More widely spread are m-Na-Al 6 beads that are found during a period ranging from the ninth to the thirteenth century, followed by m-Na-Al 2 beads (fourteenth century and onward). The beads that were traded into southern Africa and East Africa seem to have similar chemical compositions but different typologies in earlier periods, which is intriguing and would require more research to understand this discrepancy. The study of a larger number of African sites with well-dated beads is also required to develop a better understating of the dating of the transitions between the different glass types. Investigations focusing particularly on glass production in India would help us understand the reasons for shifts in glass bead trade. Acknowledgements We would like to thank the different individuals and institutions that provided the materials that were analysed. Some of the beads come from the Fort Jesus Museum in Mombasa. Most of the beads from Manda were excavated by Chapurukha Kusimba. For the beads from Juani Primary School, we need to thank Alison Crowther, Mark Horton and Nicole Boivin. The beads from Songo Mnara were excavated by Jeffrey Fleisher and Stephanie Wynne-Jones. The beads from Antsiraka Boira were provided by Martial Pauly. The Thermo ICAP-Q ICP-MS was funded by an NSF MRI Grant (#1531394).

References Babalola, A. B., Dussubieux, L., McIntosh, S. K., & Rehren, T. (2018a). Chemical analysis of glass beads from Igbo Olokun, Ile-Ife (SW Nigeria): New light on raw materials, production, and interregional interactions. Journal of Archaeological Science, 90, 92–105. Babalola, A. B., Rehren, T., Ige, A., & McIntosh, S. K. (2018b). The glass making crucibles from Ile-Ife SW Nigeria. Journal of African Archaeology, 16, 1–29. Brill, R. H. (1999). Chemical analyses of early glasses (Vol. 2). New York: The Corning Museum of Glass. Brill, R. H. (2003). The glassmakers of Firozabad and the glassmakers of Kapadwanj: Two pilot video projects. In Annales du 15e Congrès de l’Association Internationale pour l’Histoire du Verre, Corning New York, 2001 (pp. 267–268). Nottingham, UK: AIHV. Carter, A. K. (2015). Beads, exchange networks and emerging complexity: A case study from Cambodia and Thailand (500 BCE-CE 500). Cambridge Journal of Archaeology, 25(4), 733–757. Chittick, N. (1984). Manda, Excavations at an Island Port on the Kenyan Coast. Nairobi: The British Institute in Eastern Africa, Memoir Number 9. Crowther, A., Faulkner, P., Prendergast, M. E., Quintana Morales, E. M., Horton, M., Wilmsen, E., Kotarba-Morley, A. M., Christie, A., Petek, N., Tibesasa, R., Douka, K., Picornell-Gelabert, L., Carah, X., & Boivin, N. (2016). Coastal subsistence, maritime trade, and the colonization of small

530

L. Dussubieux and M. Wood

offshore Islands in Eastern African prehistory. The Journal of Island and Coastal Archaeology, 11, 211–237. Dussubieux, L. (2009). Compositional analysis of ancient glass fragments from North Sumatra, Indonesia. In D. Perret, & H. Surachman (Eds.), Histoire de Barus III: Regards sur une place marchande de l’océan Indien (XIIe-milieu du XVIIe s.) (pp. 385–417). Paris: Association Archipel/EFEO. Dussubieux, L. (2021). Elemental compositions and glass recipes. In A. K. Kanungo, & L. Dussubieux (Eds.), Ancient glass of South Asia—Archaeology, ethnography and global connection. Singapore/Gandhinagar: Springer Nature / IIT Gandhinagar. Dussubieux, L., & Bellina, B. (2018). Glass ornament production and trade polities in the Upper-Thai Peninsula during the Early Iron Age. In B. Bellina, & C. Sinopoli (Eds.), The Late Prehistoric portof-trade of Khao Sek (Chumphon province, Thailand) and its implications for the understanding of the early trade polities in Maritime Southeast Asia (Vol. 13, pp. 25–36). Archaeological Research in Asia. Dussubieux, L., Cloquet, C., & Pryce, T.O. (2021). Isotope analysis and its applications to the study of ancient Indian glass. In A.K. Kanungo, & L. Dussubieux (Eds.), Ancient glass of South Asia— Archaeology, ethnography and global connection. Singapore/Gandhinagar: Springer Nature /IIT Gandhinagar. Dussubieux, L., Gratuze, B., & Blet-Lemarquand, M. (2010). Mineral soda alumina glass: Occurrence and meaning. Journal of Archaeological Science, 37, 1645–1655. Dussubieux, L., & Kanungo, A. K. (2013). Trace element analysis of glass from Kopia. In A. K. Kanungo (Ed.), Glass in ancient India: Excavations at Kopia (pp. 360–366). Thiruvananthapuram: KCHR. Dussubieux, L., Kusimba, C. M., Gogte, V., Kusimba, S. B., Gratuze, B., & Oka, R. (2008). The trading of ancient glass beads: New analytical data from South Asian and East African sodaalumina glass beads. Archaeometry, 50(5), 797–821. Dussubieux, L., & Pryce, T. O. (2016). Myanmar’s role in Iron Age interaction networks linking Southeast Asia and India: Recent glass and copper-base metal exchange research from the Mission Archéologique Française au Myanmar. Journal of Archaeological—Reports, 5, 598–614. Dussubieux, L., Robertshaw, P., & Glascock, M. D. (2009). LA-ICP-MS analysis of African glass beads: Laboratory inter-comparison with an emphasis on the impact of corrosion on data interpretation. International Journal of Mass Spectrometry, 284, 152–161. Federeci, C. available online.The Voyage and Travaile: Of M. Caesar Frederick, Merchant of Venice, Into the East India, the Indies, and Beyond, Wherein are Contained Very Pleasant and Rare Matters, With the Customes and Rites of Those Countries. Also, Herein are Discovered the Merchandises and Commodities of those Countreyes, aswell the Aboundance of Goulde and Silver, as Spices, Drugges, Pearles, and Other Jewelles. London: Richard Jones and Edward White, 18 June 1588, https://quod.lib.umich.edu/e/eebo/A00611.0001.001?rgn=main;view=ful ltext Fischbach, N., Ngo, A.-T., Colomban, P., & Pauly, M. (2016). Beads excavated from Antsiraka Boira Necropolis (Mayotte Island, 12th–13th centuries) Colouring agents and glass matrix composition comparison with contemporary Southern Africa sites. ArcheoSciences, 40, 83–102. Francis, P., Jr. (1990). Glass beads in Asia, Part two Indo-Pacific Beads. Asian Perspectives, 29(1), 1–23. Francis, P., Jr. (1991). Beadmaking at Arikamedu and beyond. World Archaeology, 23(1), 28–43. Gill, M. S. (2017). A single ingredient for primary glass production: Reassessing traditional glass manufacture in Northern India. Journal of Glass Studies, 59, 249–259. Gogte, V. D., Pradhan, S., Dandekar, A., Joshi, S., Nanji, R., Kadgaonkar, S., & Marathe, V. (2006). The ancient port of Chaul. Journal of Indian Ocean Archaeology, 3, 62–80. Jiayao, A. (2000). Glass beads found at the Yongningsi temple. Journal of Glass Studies, 42, 81–84. Kanungo, A. K. (2004a). Glass beads in ancient India and furnace-wound beads at Purdalpur: An ethnoarchaeological approach. Asian Perspectives, 43(1), 123–150.

Indian Glass: Chronology and Distribution in Eastern Africa

531

Kanungo, A. K. (2004b). A database of glass and glass beads in India. Man and Environment, 29(1), 42–102. Kanungo, A. K. (2016). Mapping the Indo-Pacific Beads vis-à-vis Papanaidupet. Aryan Books International and International Commission on Glass. Kirkman, J. (1963). Gedi the palace. Mouton & Co. Kirkman, J. (1966). Ungwana on the Tana. Mouton & Co. Kirkman, J. (1974). Fort Jesus: A Portuguese Fortress on the East African Coast. Clarendon Press. Oxford. Kock, J., & Sode, T. (1995). Glass, glassbeads and glassmakers in Northern India. Vanlose, Danmark: THOT Print. Kusimba, C., Williams, S., Monge, J. M., Mchulla Mohamed, M., Oka, R., Oteyo, G., Kusimba, S. B., & Dussubieux, L. (2018a). Mtwapa. In S. Wynne-Jones & A. LaViolette (Eds.), The Swahili World (pp. 226–230). Routledge. Kusimba, C., Williams, S., Oteyo, G., Monge, J. M., Cheong, K., & Mchulla Mohamed, M. (2018b). Manda. In S. Wynne-Jones & A. LaViolette (Eds.), The Swahili World (pp. 156–162). Routledge. Lankton, J. W., Ige, O. A., & Rehren, T. (2006). Early primary glass production in Southern Nigeria. Journal of African Archaeology, 4, 111–138. McIntosh, S. K., Babalola, A. B., Dussubieux, L., & Rerhen, T. (2017). Ife and Igbo Olokun in the history of glass in West Africa. Antiquity, 91(357), 732–750. Nakai, I., & Shirataki, J. (2016). Chemical composition of glass beads excavated from Kofun (ca AD 2nd to 7th c.) in Western Japan by portable XRF showing glass trade among Asian countries. In F. Gan, L. Qinghui, & J. Henderson (Eds.), Recent advances in the scientific research on ancient glass and glazes, Series on Archaeology and History of Science in China (Vol. 2, pp. 73–94). World Scientific/World Century Publishing Corporation. Ogundiran, A., & Ige, O. A. (2015). “Our ancestors were material scientists”: Archaeological and geochemical evidence for indigenous Yoruba glass technology. Journal of Black Studies, 46, 751–772. Pauly, M., & Jacquot, E. (2014). Acoua, la nécropole aux perles. Préfecture de Mayotte: Direction des affaires culturelles. Pawlowicz, M. (2018). Gede. In S. Wynne-Jones & A. LaViolette (Eds.), The Swahili World (pp. 220–225). Routledge. Pion, C., & Gratuze, B. (2016). Indo-Pacific glass beads from the Indian subcontinent in Early Merovingian graves (5th -6th century AD). In A. K. Carter, S. A. Abraham, & G. O. Kelly (Eds.), Asia’s Maritime Bead Trade (Vol. 6, pp. 51–64). Archaeological Research in Asia. Pradines, S. (2010). Gedi, une cité portuaire swahilie, Islam médiéval en Afrique orientale. Le Caire : Institut Français d’Archéologie Orientale. Robertshaw, P., Wood, M., Glascock, M., & Rasoarifetra, B. (2006). Chemical analysis of glass beads from Madagascar. Journal of African Archaeology, 4(1), 91–109. Robertshaw, P., Wood, M., Melchiorre, E., Popelka-Filcoff, R., & Glascock, M. (2010). Southern african glass beads: Chemistry, glass sources and patterns of trade. Journal of Archaeological Science, 37(8), 1898–1912. Rooksby, H. P. (1964). A yellow cubic lead tin oxide opacifier in ancient glasses. Physics and Chemistry of Glasses, 5(1), 20–25. Stern, E. M. (1987). Secret of Papanaidupet. Glastechnische Berichte, 60(10), 346–351. Theal, G. M. (1898). Records of south-eastern Africa (Vols 1 & 2). Cape Town: Government of the Cape Colony. Then-Obłuska, J., & Dussubieux, L. (2016). Glass bead trade in the Early Roman and Mamluk Quseir ports—A view from the Oriental Institute Museum assemblage. In A. K. Carter, S. A. Abraham, & G. O. Kelly (Eds.), Asia’s Maritime Bead Trade (Vol. 6, pp. 81–103). Archaeological Research in Asia. Wadia, D. N. (1975). Geology of India (4th ed.). New Delhi. Wilson, T. H. (1976). Takwa: An ancient Swahili settlement of the Lamu Archipelago. Kenya past and Present, 10, 6–16.

532

L. Dussubieux and M. Wood

Wilson, T. H. (1980). The monumental architecture and archaeology of the Northern Kenya Coast. Nairobi: National Museums of Kenya. Wood, M., (2011a). Interconnections, glass beads and trade in Southern and Eastern Africa and the Indian Ocean, 7th to 16th centuries AD. Uppsala: Department of Archaeology and Ancient History, Studies in Global Archaeology 17. Wood, M. (2011b). A glass bead sequence for Southern Africa from the 8th to the 16th century AD. Journal of African Archaeology, 9(1), 67–84. Wood, M. (2016). Glass beads from pre-European contact sub-Saharan Africa: Peter Francis’s work revisited and updated. In A. K. Carter, S. A. Abraham, & G. O. Kelly (Eds.), Asia’s maritime bead trade (Vol. 6, pp. 65–80). Archaeological Research in Asia. Wood, M. (2019). Glass beads and trade in the Western Indian Ocean. Oxford Research Encyclopedia of Asian History. https://doi.org/10.1093/acrefore/9780190277727.013.334 Wood, M., Dussubieux, L., & Wadley, L. (2009). A cache of ~5000 glass beads from the Sibudu Cave Iron Age occupation. Southern African Humanities, 21, 239–261. Wood, M., Panighello, S., Orsega, E. F., Robertshaw, P., van Elteren, J. T., Crowther, A., Horton, M., & Boivin, N. (2017). Zanzibar and Indian Ocean trade in the first millennium CE: The glass bead evidence. Archaeological and Anthropological Sciences, 9(5), 879–901. Wynne-Jones, S., & Fleisher, J. (2010). Archaeological investigations at Songo Mnara, Tanzania, 2009. Nyame Akuma, 73, 2–9. Wynne-Jones, S., & Fleisher, J. (2011). Archaeological investigations at Songo Mnara, Tanzania, 2011. Nyame Akuma, 76, 3–8.

Indian Glass Beads in Northeast Africa Between the First and Sixth Centuries CE Joanna Then-Obłuska

Abstract Beads, next to pottery, are the most abundant archaeological material in Northeast Africa, often constituting the only evidence for direct and indirect trade contacts in archaeological records. While the connection of Northeast Africa to the Mediterranean world is well recognised, its link with Asian cultures is less known. This paper presents the chronological and spatial distribution of Indian glass beads in the territories of ancient Egypt, Nubia and Aksum during a time of intensive Indian Ocean trade. Chemical compositional analysis of selected samples confirms the provenance of monochrome and bichrome drawn and rounded beads to be of South Indian/Sri Lankan origin. Looking for a more comprehensive picture of South Indian/Sri Lankan glass bead imports to Northeast Africa, many museum and site bead collections were macroscopically studied. As a result of this research, some Indian beads are traced to Egyptian Red Sea ports in the Early Roman period that is between the first century BCE and the third century CE. However, large-scale glass bead imports are evident at sites on the Red Sea Coast and in the interior of Northeast Africa in Late Antiquity, i.e. between the fourth and the sixth century CE.

1 Introduction While Northeast Africa’s link with the Mediterranean world is well recognized, its link with Asian cultures, however, is less known. Beads, next to pottery, are the most abundant archaeological material in Roman and Late Antique Northeast Africa, often constituting the only evidence for trade contacts in archaeological records. Finding the archaeological evidence supporting the existence of extensive trade between Northeast Africa and the Indian world is the aim of this interdisciplinary work. The results of combined chemical and/or morphological studies of beads found in Northeast Africa are presented for Egypt, Nubia or Aksum (Fig. 1). Since the loan of samples from Egypt and Ethiopia is restricted, only beads in some European and J. Then-Obłuska (B) Atiquity of Southeastern Europe Research Centre, University of Warsaw, Krakowskie Przedmie´scie 32, 00-927 Warsaw, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 A. K. Kanungo and L. Dussubieux (eds.), Ancient Glass of South Asia, https://doi.org/10.1007/978-981-16-3656-1_22

533

534

J. Then-Obłuska

Fig. 1 Map of the locations with Indo-Pacific beads in Northeast Africa (drawing by Szymon Ma´slak)

Indian Glass Beads in Northeast Africa …

535

Sudanese collections from Nubia as well as the ones housed in the Oriental Institute Museum (OIM) in Chicago, from Nubia and from the Egyptian Red Sea port of Quseir, were analysed for elemental compositions. In general, the Indian beads found in the north-east part of Africa were parts of drawn glass tubes, some of which were heat-rounded in a pot. This is a bead-making technique attributed to South Asia, and the beads so made are called Indo-Pacific (IP) beads (Francis, 2002: 20–26, 48, Fig. 3.3) presents a sketch of the ‘roundness factor’). The beads are usually monochrome, and some are compound bichrome, with occasional bichrome and polychrome striped ones. Peter Francis (2000, 2002, 2007) studied the Indo-Pacific beads from the Egyptian Red Sea port sites of Berenike (Seasons 1994–2000) and Quseir as well as from the Eastern Desert site of Shenshef, and he considered the presence of such beads in Northeast Africa restricted to the port sites and their desert surroundings. Furthermore, Francis stated that dark blue, red, black and violet beads found in Early Roman Berenike mirrored the beads that comprised the bulk of the beads from Arikamedu, Southern India, while blue-green, green and some orange ones from Late Roman Berenike were similar to the ones recorded at Mantai in Sri Lanka (Francis, 2002: 228, Ref. 21, 2004, 2013). This paper summarizes new evidence coming from macroscopic and elemental analyses of beads found at coastal and inland sites of Egypt, Nubia and Aksum. The beads come from the first- to third-century CE assemblages excavated at Late Ptolemaic and Early Roman sites in Egypt as well as from the fourth- to sixth-century CE assemblages excavated at Late Roman sites in Egypt, post-Meroitic (Blemmyan, Nobadian, Early Makurian) sites in Nubia, and Classical and Late Axumite sites in Aksum.

2 Egypt While some IP glass beads can be traced in the literature to the Egyptian Nile Valley (where they are mistakenly described as green stone or glass of Egyptian production), their occurrence there appears to be scant, though it must be stressed that this kind of material is understudied.1 The large variety of glass beads found at Roman Red Sea ports and Eastern Desert sites in Egypt provides ample evidence of overseas contacts. While chemical composition analysis of beads from the port of Quseir has confirmed South Indian/Sri Lankan origin, the provenance of the Berenike, Marsa Nakari, Shenshef and Sikait objects remains uncertain as these beads have not been chemically analysed. All finds from the Egyptian ports and desert sites are presented below.

1

In recent study, green IP beads have been identified in four child graves (the end of fourth–sixth century CE) at the Middle Egyptian Nile Valley cemetery in Matmar (Then-Obłuska & Ple¸sa, 2019: Fig. 4: 11.2, 14.1, 5:18.3, 6:23.3).

536

J. Then-Obłuska

2.1 Quseir Extensive excavations at Quseir by American and British archaeologists between 1978 and 2003 brought to light substantial remains of a thriving harbour catering to international trade between the Mediterranean world and the littorals of the Southern Red Sea and Indian Ocean (Peacock & Blue, 2001, 2006; Whitcomb, 1979). The site was occupied during two periods: the Early Roman period (c. first–mid-third century CE), when the site was known as Myos Hormos, and the Late Ayyubid to Mamluk period, when the site became Quseir al-Qadim (thirteenth–fourteenth century CE). Recently, eleven glass beads were found at Quseir al-Qadim and two of them are similar to IP beads (Peacock, 2011: 78). These beads add to more than sixty glass beads found earlier during the excavations of the University of Chicago’s Oriental Institute at Quseir, directed by Donald Whitcomb and Janet Johnson in 1978, 1980 and 1982. The beads appear in the excavation reports (Whitcomb, 1979: 196–198; Meyer, 1982: 226) and also in a separate volume on Quseir glass (Meyer, 1992: 41–2, 94–5, 180). Additionally, the beads have been analysed with laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) (ThenObłuska & Dussubieux, 2016). This method was chosen as it provides results for trace elements, which are useful in establishing the sourcing of the glass and in distinguishing compositional groups that can be compared to different types of South Asian glass. Although several drawn, cut and rounded monochrome glass beads were found in Quseir, only one bead, a semi-translucent yellow bead (OIM E45095) dating to Islamic period, revealed a mineral soda high-alumina (m-Na-Al) composition (Fig. 2.1). Interestingly, it belongs to the m-Na-Al 1 type that has already been recognized from sites in Southern India and Sri Lanka, dating from the fourth century BCE to the fifth century CE. The bead found in the Islamic context should be seen as a residual or reused item from the Roman period (Then-Obłuska & Dussubieux, 2016).

2.2 Berenike Berenike lies about 300 km south of Quseir. It is located south of Ras (Cape) Banas. The site was excavated by the American–Dutch Berenike Project between 1994 and 2001 and then by the Polish-American Berenike Project since 2009 (e.g. Sidebotham, 2011; Sidebotham & Zych, 2011, 2017). Berenike is an excellent example of a cosmopolitan harbour. At this port, people of many ethnicities, religions and cultures met and lived in close proximity. In his characterization of the urban population of the port, the excavator Sidebotham (2011: 69) wrote: ‘Those who dwelt in Berenike, either briefly or as more or less permanent settlers, came from throughout the ancient world, including Egypt, the Mediterranean, Axum, sub-Saharan Africa, and the kingdoms of southern Arabia, Nabataea, and Palmyra. Indian sailors or

Indian Glass Beads in Northeast Africa …

537

Fig. 2 Indian/Sri Lankan bead found at Quseir (photograph by J. Then-Obłuska) and South Asian bead types from Ptolemaic and Early Roman Berenike (photographs by J. Then-Obłuska, K. Brauli´nska and The Berenike Project): 2.1. Quseir, OIM E45095; 2.2. Berenike, BE1177/001/PB 03; 2.3. Berenike, BE11-77/001/PB 01; 2.4. Berenike, BE11-77/001/PB 03; 2.5. Berenike, BE11-77/001/PB 03; 2.6. Berenike, BE11-77/001/PB 02; 2.7. Berenike, BE11-74/003/PB 04; 2.8. Berenike, BE11-76/999/PB 27; 2.9. Berenike, BE14-96/005/PB 07; 2.10. Berenike, BE14-96/006/PB 08; 2.11. Berenike, BE15/18–107-044/PB071; 2.12. Berenike, BE01/18–048128/PB130

538

J. Then-Obłuska

merchants, and likely their Sinhalese contemporaries, visited Berenike and either stayed for a few months, arriving in early summer and catching the monsoon back to India in August, or resided there on a more permanent basis’. Multicultural contacts would have involved both elites and non-elites, e.g. crews of merchant ships, parties passing through the desert, craftsmen and nearby dwellers providing supplies and services to residents of this port, which laid at the edge of the desert. During the work by the American–Dutch Project, Francis (2002, 2007) identified IP beads at the site, whereas only a few beads were recorded from Ptolemaic and Early Roman contexts (Francis, 2002: 48). Excavations by the Polish-American team in 2011 documented 11 additional beads from Ptolemaic and Early Roman Berenike contexts (Fig. 2.2–8). They exhibited the following colours: transparent or translucent purple, and translucent dark blue. According to Francis (2002: 228, Note 21; 2004: 450), these colours would indicate an Arikamedu provenance of Early Roman drawn and rounded beads. However, lack of chemical composition analysis makes it impossible to ascribe Early Roman drawn and rounded glass beads to this region of production in South Asia. Additionally, the 2014 excavations of the Early Roman rubbish dump in Berenike have brought to light two outstanding finds in the form of threaded drawn glass beads (Then-Obłuska, 2018b: Fig. 5: 1, 3). These were yellow and dark purple to black beads on a string fragment (Fig. 2.9) as well as red beads still on their string too (Fig. 2.10). A few examples of collared glass beads of South Asian origin and ca. 10 mm in length were recently found at the Early Roman animal cemetery (Fig. 2.11–12). IP beads constituted 41% (396 examples) in the Late Roman Berenike during the excavations by American–Dutch team (Francis, 2002: 48). Drawn and rounded monochrome beads were also the most common type in the late Berenike phase during the 2010 season (Then-Obłuska, 2015), constituting 54% (ca. 640 objects) of all glass beads and 45% of all beads found (Fig. 3.1–23). Semi-translucent and translucent green (48% of all drawn and rounded beads), translucent light blue-green (23%) and semi-translucent yellow and opaque yellow (15%) dominated the colour palette. Other colours were: opaque orange (6%), black (3%), opaque dark red (3%), translucent dark blue (2%), white (1) and translucent amber (1). The blue/green and orange varieties were usually very tiny beads, up to 2 mm in diameter. The remaining ones ranged from approximately 2 mm to 5 mm in diameter. Almost all the glass beads recorded between 2011 and 2013 from the Harbor Temple were IP ones, and many IP beads were found during 2015 seasons in the upper layers of the Isis Temple (Then-Obłuska, 2017a, 2018b). A few beads were made of drawn and rounded sections of glass tube covered with longitudinal stripes. The beads were dark purple (but apparently black) with longitudinally applied white stripes (Fig. 3.24–25), registered from late Berenike context, i.e. second half of the fourth–fifth century CE (Then-Obłuska, 2015: Fig. 4: 31–32). Beads made of white glass covered with colourful stripes were also found at Berenike trench 18 in the 1998 season and trench 31 in the 1999 season (Fig. 3.26– 28). While trench 31 is an Early Roman (mid- to late first century CE) trash dump (Sidebotham, 2007: 44), the pottery dates for trench 18 (BE98-18–192/032) range

Indian Glass Beads in Northeast Africa …

539

from the fourth to fifth century CE (Sidebotham, 2000: 88). These beads are said to have originated from Sri Lanka or—more precisely—Mantai (Francis, 2002: 44, Pl. 32, 2013: Pl. 13.4.3: 13). The drawn beads in the form of an opaque orange layer atop a translucent red one are among the rare finds from old Berenike excavations, unearthed in season 1995 (BE95-005–013#72, BE95-005-TS#97) (personal observation) (Fig. 3.29–30).

2.3 Marsa Nakari The Marsa Nakari site lies on the Red Sea Coast between Quseir to the north and Berenike to the south. It has been tentatively identified with Nechesia, the Red Sea port of likely Ptolemaic foundation (Sidebotham et al., 2008). The three archaeological seasons (1999, 2001 and 2002) by Northern Arizona University, directed by John A. Seeger, provided evidence of Early and Late Roman activities at the site as well as some activities during the Ptolemaic period (Seeger, 2001; Seeger & Sidebotham, 2005). Monochrome drawn and rounded beads are predominant within the glass assemblages at Marsa Nakari (Fig. 4.1). The assemblages include 146 monochrome beads: 107 green, 21 blue, 10 opaque yellow and eight opaque orange, all measuring between 1.6 and 4.8 mm in diameter (Then-Obłuska, 2018d). According to Francis (2002), the above-mentioned colours would associate the beads with one of the Sri Lankan production centres. This statement might to some extent be confirmed by the new evidence in the form of two outstanding types: a striped bead and a stupa bead. The first bead is a drawn and rounded section of a white glass tube covered with longitudinal stripes (Fig. 4.2) (Then-Obluska, 2018e: Fig. 3.14, Find number: MN02-007–012#14/0013, D 7.1, Th. 4.9, HD 1.4.) of the type already mentioned from Berenike. The second outstanding glass object, a so-called stupa bead, was found in the same context as the striped bead (Fig. 4.3) (Then-Obłuska, 2018d: Fig. 3.15); Find number: MN02-007–012#14/0013, W 10.7, Th. 8.6, L 10.4, HD 2.5). It is a moulded round tablet with a raised edge and centre. This type was called stupa by Francis (2002: 138), since it resembles a Buddhist stupa with its surrounding wall and two entrances. Although stupa beads are known from Thailand and Malaysia, so far, the earliest specimens have been discovered at the Sri Lankan sites of Ridiyagama, Tissamaharama, Jethavana and Mantai (Hannibal-Deraniyagala, 2001: 203–226, 2005: 23, 2013: 370– 373; Francis, 2013: 359, plate 13.4.4: 30; Dussubieux, 2001). Interestingly, at Jethavana, a few of the stupa beads were made of amethyst (Hannibal-Deraniyagala, 2005: 23). The glass beads can be blue, green, red and yellow in colour. Chemical composition analysis reveals that they were produced from various types of glass: Syro-Palestinian, Middle Eastern and South Asian. The hypothesis is that those glass beads were manufactured in Sri Lanka but with recycled glass (Dussubieux, 2001 and personal communication). Two of the stupa beads from Ridiyagama were found within archaeological contexts dated from the fourth century BCE to the first

540

J. Then-Obłuska

Fig. 3 Indian/Sri Lankan bead types found at Late Roman Berenike (photographs by J. ThenObłuska and The Berenike Project): 3.1. BE10-59/001/PB 06; 3.2. BE10-59/001/PB 06; 3.3. BE1061/003/PB 35; 3.4. BE10-59/001/PB 06; 3.5. BE10-61/003/PB 31; 3.6. BE10-61/013/PB 42; 3.7. BE10-59/001/PB 17; 3.8. BE10-59/001/PB 15; 3.9. BE10-59/001/PB 13; 3.10. BE10-59/001/PB 06; 3.11. BE10-59/001/PB 06; 3.12. BE10-61/005/PB 37; 3.13. BE10-59/001/PB 06; 3.14. BE1059/001/PB 07; 3.15. BE10-59/001/PB 07; 3.16. BE11-79/004/PB 06; 3.17. BE10-59/001/PB 06; 3.18. BE10-59/001/PB 06; 3.19. BE10-59/001/PB 06; 3.20. BE10-59/001/PB 06; 3.21. BE1059/001/PB 02; 3.22. BE10-59/001/PB 02; 3.23. BE10-59/001/PB 02; 3.24. BE10-59/001/PB 17; 3.25. BE10-59/001/PB 05; 3.26. BE99-10/436/PB 01; 3.27. BE99-31/018/PB 51; 3.28. BE9818/192/PB 32; 3.29. BE95-05/013/PB 72; 3.30. BE95-05/TS/PB 97

Indian Glass Beads in Northeast Africa …

541

Fig. 4 Indian/Sri Lankan bead types found at Marsa Nakari (photographs by J. Then-Obłuska and The Berenike Project): 4.1. MN 02–07-012#15/0017; 4.2. MN 2002/0013; 4.3. MN 2002/0013

542

J. Then-Obłuska

century CE (Dussubieux, 2001 and personal communication). In Tissamaharama, they occurred for the first time in the early fourth century CE with a few in the later phases (sixth–seventh to seventh–eighth century CE) (Hannibal-Deraniyagala, 2013). In the excavations of the Jethavana Stupa, they were found in a context dated to the fourth century CE (Hannibal-Deraniyagala, 2005: 23). This fourth-century dating would also be consistent with Marsa Nakari examples, as the stupa and striped beads were found together with other IP beads, usually dated to the Late Roman period.

2.4 Shenshef In 1997, a few trenches were excavated at Wadi Shenshef (also known as Hitan Shenshef) by the American–Dutch Berenike Project team. This site is located in the Eastern Desert in the extreme south-east part of Egypt, 12 km west of the Red Sea and 21.3 km south–south-west of the site of Berenike (Gould, 1999: 371). Stretching 800 m (east–west) and 300 m (north–south), the site includes approximately 300 structures of various sizes and functions, and at least 500 tumulus tombs (Gould, 1999; Aldsworth, 1999; Sidebotham et al., 2008: 360; Sidebotham, 2011: 275–276). The purpose of this large and well-built settlement has yet to be established. Excavation of seven trenches, BE97-Sch.1 to BE97-Sch.7, revealed refuse middens near the houses have demonstrated close contacts between Shenshef, Berenike and many distant lands dating from the fourth/fifth to sixth century CE. Walnuts, olives, almonds and umbrella pine were imported from the Mediterranean; amphorae came from the Eastern Mediterranean, mainly from Cyprus and Cilicia (Tomber, 1998: 170–179); black pepper, sorghum and teak originated from South Asia; a sapphire from Sri Lanka was also found. The site also yielded an important collection of 200 beads, pendants and their fragments that are housed in the Supreme Council of Antiquities (SCA) storage room at Quft. The bead and pendant collections at Shenshef are made predominantly of glass (n = 180). One-third of this glass assemblage (n = 69) is comprised of monochrome drawn and rounded beads (Fig. 5.1–3) (Then-Obłuska, 2017b: Figs. 1:2,4–12,13, 21–29; 2:7–9,13; 3:4,5,10–16; 4:6,10,11,14,15,30–33; 5:6,8; 7:3,4,10–13, 8:3,5,7,14,16,17,19–21; 9:3,6,10,12,14,20–24; 10:1,2). These measure from 1.5 to 7.0 mm in diameter. Green (n = 29) and blue-green (n = 15) beads dominate this type, but yellow (n = 8), orange (n = 7), red (n = 6), white (n = 2) and black (n = 1) are also present in smaller quantities. Additionally, one black bead is decorated with stripes in white and red (Fig. 5.4) (Then-Obłuska, 2017b: Fig. 1:14). In sum, IP beads constitute 35% of whole bead assemblage and 39% of all glass beads at Shenshef.

Indian Glass Beads in Northeast Africa …

543

Fig. 5 Indian/Sri Lankan bead types found at Shenshef (photographs by J. Then-Obłuska and The Berenike Project): 5.1. BE97-SH.01/005/PB 05; 5.2. BE97-SH.01/baulk clean/PB 08; 5.3. BE97-SH.02/003/PB 03; 5.4. BE97-SH.01/004/PB 04

2.5 Sikait More than 2550 beads, pendants and their fragments were collected during excavations and field surveys conducted in May–June 2002 and December 2002–February 2003 in the environs of Sikait (ancient Senskis/Senskete), a Roman period mining

544

J. Then-Obłuska

settlement in the Eastern Desert of Egypt (Sidebotham et al., 2004). An almost eight-km-long route links the settlement at Sikait to Wadi Nugrus/Wadi Gemal, the juncture with the main road between the Ptolemaic–Roman Red Sea emporium of Berenike and the Nile (at Edfu and later at Koptos). Located about 120 km northwest of Berenike and 45 km west of the Red Sea, Sikait was one of at least nine beryl/emerald-mining settlements that comprised Mons Smaragdus, the only source of these gemstones within the boundaries of the Roman Empire. The floruit of Sikait and its surroundings was from the first to the sixth century CE, though there is evidence of earlier Ptolemaic and later Islamic activity. The beads and pendants were found in all 11 excavated trenches. The overwhelming majority was made of glass. Monochrome drawn and rounded beads constitute 23% of whole bead assemblage in Sikait and 30% of all glass beads (Fig. 6.1–4) (Then-Obłuska forthcoming). They are almost absent from trenches excavated in the eastern part of the site in 2002. Their percentage share varies in the western trenches, peaking at 54% of all glass beads in trench SK03-10. In general, green beads predominate the IP collection, with blue and yellow as the next most common colour. The remaining beads are orange, blue-green, red, black and purple. The assemblage also features two black- and white-striped beads (Then-Obłuska forthcoming. While one bead was a part of a white-striped black glass tube (Fig. 6.5), the other was a section of black tube covered with black and white stripes (Fig. 6.6).

3 Nubia Nubia is a geographic region located in the north-east corner of Africa. It encompasses the southernmost part of Egypt and northern Sudan, and it is divided into Lower Nubia in the north and Upper Nubia in the south. Different regions within Nubia are separated by a series of cataracts, from the First Cataract located south of Aswan to the Sixth Cataract north of present-day Khartoum. Three entities emerged in Nubia between the fourth and sixth century CE after the fall of the Kingdom of Meroe: Nobadia in Lower Nubia, Early Makuria in Upper Nubia and Alwa (Alodia) in the region up from the Fifth Cataract. Once the Romans had withdrawn from Lower Nubia around 298 CE, the Nobadians, possibly from the Western Desert, and the Blemmyes from the Eastern Desert encroached on the area (Ricke, 1967; Strouhal, 1984). The Meroitic pyramid and mastaba superstructures disappeared then, to be replaced by the widespread use of tumuli with single- or multi-chambered substructures to contain the burials and abundant grave goods. During the interdisciplinary project on Nubian glass beads run by the University of Warsaw, the chemical composition of almost two hundred beads and pendants was examined using LA-ICP-MS (Then-Obłuska & Wagner, 2017, 2019a, 2019b). The glass bead samples from Nubia were collected across a broad chronological and geographical spectrum from the first to the sixth century CE and from areas between the First and Sixth Nile Cataract.

Indian Glass Beads in Northeast Africa …

545

Fig. 6 Indian/Sri Lankan bead types found at Sikait (photographs by J. Then-Obłuska and The Berenike Project): 6.1. SK03-10–148#343; 6.2. SK03-10–148#343; 6.3. SK03-10–148#343; 6.4. SK03-10–026#049; 6.5. SK02-03–016#026; 6.6. SK02-03–024#032

546

J. Then-Obłuska

According to the results, the low-alumina glass used in the production of Nubian beads was of Egyptian and Levantine origin, while the high-alumina glass, with level of alumina above 5 wt%, characterized South Asian glass compositions. Among South Asian high-alumina glasses, Dussubieux et al. (2010) distinguished a few subtypes based on the concentrations of five elements: strontium (Sr), zirconium (Zr), barium (Ba), uranium (U) and caesium (Cs). The compositions of highalumina glass beads from Nubia were compared against these subtypes. They show average concentrations of trace elements that match subtype 1 glass (Then-Obłuska & Wagner, 2019a, 2019b), which is usually found at sites in Sri Lanka and South India dating between the turn of the fourth century BCE and the fifth century CE (and in Southeast Asia between the fifth century BCE and the tenth century CE), and thus were most probably manufactured at one of Sri Lankan sites. The subtype 1 glass beads found in Nubia were parts of drawn glass tubes that were heat-rounded. They are green, yellow, blue, blue-green, orange, black, red and orange-on-red, and measure between two and seven mm in diameter. Looking for a more comprehensive picture of the IP glass bead import to Northeast Africa, the author studied macroscopically several European, American, Egyptian and Sudanese bead collections from Nubia, representing assemblages in all Nubian regions (Then-Obłuska & Wagner, 2019b). As a result, IP beads were identified in the Nile Valley belt of about 900 km that run between the First Nile Cataract and the confluence of the Blue and White Nile, where they appeared at least in 30 sites and 60 graves. Altogether, more than 7300 IP beads were found. The beads are monochrome and bichrome, orange-on-red including (Fig. 7.1–8). In Lower Nubia, a few IP beads were already present at the mid-fourth-century CE Blemmyan sites of Bab Kalabsha and Wadi Qitna (Then-Obłuska, 2016a). Their numbers were much higher in the late fourth-century sites at Qustul and Gammai, and at the beginning of the fifth century at Ballana (Then-Obłuska 2018c; ThenObłuska & Wagner, 2019b). Upriver from the Third Cataract, they were excavated in Early Makuria tombs, including those at El-Zuma and El-Detti, dated between the mid-fifth and mid-sixth century CE (Then-Obłuska, 2016b, 2016c, 2017c). Meroe in Butana and Khour Shambat at the confluence of the Blue and White Nile are the southernmost sites where these beads could be macroscopically recognized in the Nile Valley (Then-Obłuska & Wagner, 2019b). Although very rare, a few sites with typical post-Meroitic materials were recently discovered in Eastern Sudan. Among other places, sites with typical post-Meroitic materials were recently recorded north of the Khor Umm Sitebah and at the foot of the Jebel Abu Gamal (Manzo, 2017: 58–59, Fig. 52, site JAG 1). The latter site is characterized by the occurrence of a large soil tumulus, exceeding 10 m in diameter, delimited by a ring of granite rocks and originally topped with white quartz pebbles covering funerary pits filled with stones and a lateral niche containing skeletons in contracted position. Among the objects collected, there were IP green, yellow and orange-on-red glass beads, as can be observed in the published photograph (Manzo, 2017: Fig. 52 c-d). Nevertheless, the isolation of the sites may confirm that they are intrusive and possibly represent an occasional presence in the region of groups moving from the Butana.

Indian Glass Beads in Northeast Africa …

547

Fig. 7 Indian/Sri Lankan bead types found in Nubia (photographs by J. Then-Obłuska) (MAUS/SJE—Museum of Archaeology University of Stavanger/Scandinavian Joint Expedition): 7.1. Serra East, MAUS/SJE 25/19:3; 7.2. Serra East, MAUS/SJE 25/47:4; 7.3. Serra East, MAUS/SJE 25/221:1; 7.4. Bab Kalabsha, E8/a, E8/b, OIM E42042; 7.5. Qustul, Q41-1, OIM E20058.004; 7.6. Serra East, MAUS/SJE 25/221:1; 7.7. Zuma, Z12/82; 7.8. Zuma, Z4/51

548

J. Then-Obłuska

The number of IP beads in a grave varies from a single specimen to several hundred. When preserved on original string, they often form necklaces or anklets composed of beads uniform in colour. Otherwise, they were threaded together with beads made of other materials or other glass types. In one case, they formed earrings (OIM E42042). The study of the social stratification of the tombs in which the IP beads were found indicates that they were spread through the entire society. Neither is there a specific age or sex that the presence of the beads could be associated with (Then-Obłuska & Wagner, 2019b). It must be underlined that besides being widely spread at private cemeteries, IP beads were also recorded in Nobadian and Early Makuria royal and elite tombs. The import of IP beads was not a random process and was definitively based on market needs. Orange (n = 2855), black (n = 1909) and apple green (n = 1461) beads dominate the IP assemblages known from Nubia. Beads in remaining colours are recorded in small numbers. It seems that the colours dominating among IP beads allowed to fill a coloristic gap in the Nubian beadwork in the period between the fourth and sixth century CE. All the small orange, black and apple green glass beads in post-Meroitic Nubia are of IP origin, while the results suggest that there was no great need for red IP beads, as the red beads recorded in Nubian assemblages are predominantly of Mediterranean origin. With the evidence from the Eastern Desert and Red Sea coastal sites, a direct eastern direction of the IP beads’ arrival in Nubia is better demonstrated. Yet while the evidence for long-distance trade with India is much more robust in the Egyptian Red Sea ports, a southern direction for the coming of beads to Nubia cannot be excluded.

4 Aksum The Kingdom of Aksum, in modern Ethiopia and Eritrea, was a trading nation and naval power in the first–seventh century CE, exporting ivory, rhino horns, hippo hides, turtle shells, monkeys and slaves to the Roman Empire, Egypt and India. At the regional scale, the Aksumite polity, centered in the city of Aksum, emerged as a local power between the Kingdom of Meroe in the middle Nile Valley and the Kingdom of Himyar in south-western Arabia. Interaction was more intense with South Arabia, which the Aksumites occupied in the third and sixth century CE, than with Nubia, where the Aksumites conducted raids, possibly contributing to the collapse of Meroe in the late third–early fourth century CE (Fattovich, 2018; Hatke, 2013). Aksumite coins and ceramics, similar to those from Aksum and Adulis, have been found at Aqaba in Jordan, Quseir al-Qadim and Berenike in Egypt, al-Madhariba and Qana in southern Yemen, Khor Rori in Oman, perhaps Kamrej in northern India, Mangalore, Madurai and Kaur in Southern India, and Tissamaharama in Sri Lanka (Tomber, 2005, 2008).

Indian Glass Beads in Northeast Africa …

549

4.1 Adulis Adulis, the Aksumite port located near modern Massawa in Eritrea, was another place where Asian goods were unloaded. Adulis prospered between the first and the seventh century CE. Goods traded through this port were mentioned in a sixth century CE work by Cosmas Indicopleustes (2.49), who stated that Adulis was a fair-sized village that transhipped goods between the Red Sea and the Mediterranean Sea, the Indian Ocean, Arab lands, Middle and Far East, Asia as well as the African interior. A recent find of a drawn and rounded green glass bead in Adulis by the Italian Project (e.g. Carannante et al., 2015) would make a good starting point for tracing IP glass beads at this site (find no: ADU 12, B5-12, SU 1035—personal observation).

4.2 Aksum In the period between the third and the fifth century CE, Aksum was ca. 100 ha in area. Compounds, hamlets and a few villages were clustered along the slopes of Beta Giyorgis, and three large settlements were located north, east and west of the town (Fattovich, 2018 and references). An elite monumental cemetery with hewn stelae representing multi-storeyed buildings (the so-called Stelae Park) was established at the eastern base of Beta Giyorgis in the late third–early fourth century CE. Elite tombs were located both east of the main settlement, along the route to eastern Tigray, and at a large elite cemetery established west of the town (Munro-Hay, 1989: 142– 149; Fattovich et al., 2000: 73–74; Phillipson, 1997: 44–61, 93–122, 2000, 2012: 139–147). Describing beads from Aksum, Helen Morrison (1989: 170, Fig. 11: 23, 24) classifies under Type IV drawn beads in many colours (many shades of green and blue, yellow, red, purple, white, bluish-green, white and one opaque red on dark green), among which minute to small sizes (2–5 mm) outnumbered the medium (5–8 mm in diam.). These were the most numerous beads from the excavations of 1973–74; over half of them (all small) were found in the fourth-century DA I tomb, while none came from the third-century GT II tomb (Morrison, 1989). Drawn beads were also the most common type in the late fourth-century Tomb of the Brick Arches in Aksum (883 examples out of a total 1139 found beads) (Harlow, 2000: 85, Fig. 62c, 64q and t, 65a). Taken together, they make up 77.5% of the entire bead assemblage. Most of them were in green glass. They fell into two main sizes: small (1.3–2.3 mm in diam.) and large (with the largest being 3.6 mm in diam.).

550

J. Then-Obłuska

4.3 Maryam Anza Recent study of the Maryam Anza cemetery, dated to the fourth and fifth century CE, brought to light new data on Indian bead imports into Aksum. Maryam Anza lies in the northern highlands of Ethiopia’s Tigray Province, some 130 km to the east of the ancient capital at Aksum and some 250 km to the south-west of the ancient Red Sea port at Adulis. A few thousand beads were excavated by the British-Ethiopian Maryam Anza Project at the Maryam Anza Cemetery Site (MACEM) during the three seasons between 2014 and 2016 (Then-Obłuska, 2018a). The monochrome drawn and rounded beads make up about 86% of all beads and 88% of all glass beads at Maryam Anza (Fig. 8.1–8). A collection of about 2600 tiny orange beads dominates the assemblage. Blue (n = 627) and black (n = 593) are the second and third most common colours. The remaining beads are green (almost 120 beads), blue-green (51 beads), red (31 beads) and yellow (28) (Then-Obłuska, 2018a).

5 Conclusion The presence of drawn and rounded beads in the Late Ptolemaic and Early Roman period has been evidenced solely from the Egyptian Red Sea ports of Quseir and Berenike. However, in the Late Roman period they were much more common at both coastal and inland sites in Northeast Africa. Chemical composition analysis of glass samples coming from the port of Quseir and from many sites throughout Nubia indicated high-alumina glass composition, confirming the South Indian/Sri Lankan provenance of these beads. In terms of quantity and proportion (Fig. 9), IP beads in Northeast Africa are more common at Egyptian Red Sea port sites, where they make up more than 50% of glass bead assemblages. Their share is a bit smaller in the neighbouring Eastern Desert sites, with almost 40% in Shenshef and 30% in Sikait. This declining tendency continues into inland sites: IP beads constitute ca. 10% and less of glass bead assemblages in the Nubian Nile Valley. The lack of Nubian Red Sea ports in Late Antiquity as well as domination of traditional materials, i.e. ostrich eggshell and faience, in Nubian beadwork might be a reason for this small share of imported glass beads. On the other hand, in Upper Nubia, located farther away from the Mediterranean glass source, the IP beads made up 63% of glass beads at El-Zuma site. The location of Aksum as the southernmost ancient country in Northeast Africa and the proximity of its capital and other sites to the Red Sea port of Adulis were most probably the main reasons for the largest share of IP beads in this part of Africa. At Maryam Anza, they actually dominated the bead and glass bead assemblage (ca. 88%). Some interesting observations can be drawn from looking at the north–south distribution of colours of the IP beads in Northeast Africa (Fig. 10). Orange beads

Indian Glass Beads in Northeast Africa …

551

Fig. 8 Indian/Sri Lankan bead types found at Maryam Anza (photographs by J. Then-Obłuska): 8.1. MACEM ‘15-T5/6 (136)-FN25-5; 8.2. MACEM ‘15-T5/6 (136)-FN25-4; 8.3. MACEM ‘15-T5/6 (136)-FN25-5; 8.4. MACEM ‘15-T5/6 (136)-FN25-1b; 8.5. MACEM ‘15-T5/6 (136)-FN25-1; 8.6. MACEM ‘15-T5/6 (136)-FN25-2; 8.7. MACEM ‘16-T14 (202)-FN220-15; 8.8. MACEM ‘15-T5/6 (136)-FN25-3

552

J. Then-Obłuska

Fig. 9 Percentage share of Indo-Pacific beads at Northeast African sites (map drawing by Szymon Ma´slak): 9.1. Sikait (after Then-Obłuska forthcoming); 9.2. Berenike (after Then-Obłuska, 2015, 2017a); 9.3. Shenshef (after Then-Obłuska, 2017c); 9.4. Lower Nubia (Scandinavian Joint Expedition) (after Then-Obłuska & Wagner, 2019b); 9.5. El-Zuma (after Then-Obłuska, 2016c, 2017c); 9.6. Maryam Anza (after Then-Obłuska, 2018a)

displayed a rising tendency in the north–south direction, while green beads saw a reduction. This might have been caused by difference in fashion and market needs between Egypt on the one hand and Nubia and Aksum on the other. However, it was green beads that prevailed in the drawn glass bead assemblage in the Tomb of the Brick Arches in Aksum (Harlow, 2000: 85, Fig. 62c, 64q and t, 65a). IP beads as found in Late Antique Northeast Africa were items of the global market in a time of intensive maritime trade contacts in the Indian Ocean region. Outside Southeast Asia (e.g. Carter, 2016) and South Arabia (Lischi, 2018), they have recently been identified in Early Merovingian graves dated to the fifth–sixth century CE as well as in the present territories of the Netherlands, Germany, Switzerland, Spain and Serbia (cf., e.g., Gratuze et al., 2021; Pion & Gratuze, 2016: Fig. 11). Among other things, they were used in rich decoration of Byzantine clothing: in embroidery of the textiles or as elements of headgear. It is also interesting that in both Nubia and

Indian Glass Beads in Northeast Africa …

553

Fig. 10 Percentage share of colours of the Indo-Pacific beads as found in Northeast Africa (map drawing by Szymon Ma´slak): 10.1. Marsa Nakari (after Then-Obłuska, 2018d); 10.2. Sikait (after Then-Obłuska forthcoming); 10.3. Berenike (after Then-Obłuska, 2015, 2017a); 10.4. Shenshef (after Then-Obłuska, 2017c); 10.5. Nubia (after Then-Obłuska & Wagner, 2019b); 10.6. Maryam Anza (after Then-Obłuska, 2018a)

Aksum the largest collections of orange beads were found in rich women’s graves: as anklet in the queen tomb at Ballana (Then-Obłuska & Wagner, 2019b) and in a grave of a rich young woman at Maryam Anza (Then-Obłuska, 2018a: Grave 12). Due to lack of data, the presence of IP beads at the Egyptian Nile Valley sites remains an open subject to be studied. Acknowledgements I am grateful to Dr. Laure Dussubieux for introducing me to LA-ICP-MS analysis for glass beads. I would like to thank Prof. Barbara Wagner and Luiza K˛epa for running laboratory analysis of glass beads from Nubia. Many thanks go to Szymon Ma´slak for drawing all the maps.

554

J. Then-Obłuska

References Aldsworth, F. (1999). The buildings at Shenshef, In S. E. Sidebotham, & W. Z. Wendrich (Eds.), Berenike 1997: Report of the 1997 excavations at Berenike and the survey of the Egyptian Eastern Desert, including excavations at Shenshef (pp. 385–418). Research School of Asian, African, and Amerindian Studies (CNWS), Universiteit Leiden. Carannante, A., Flaux, C., Morhange, C., & Zazzaro, C. (2015). Adulis (Eritrea) in its regional maritime context. A preliminary report of the 2015 field season. Newsletter Di Archeologia CISA, 6, 279–294. Carter, A. K. (2016). The production and exchange of glass and stone beads in Southeast Asia from 500 BCE to the early second millennium CE: An assessment of the work of Peter Francis in light of recent research. Archaeological Research in Asia, 6, 16–29. Dussubieux, L. (2001). L’Apport de l’ablation laser couplée à l’ICP-MS à l’étude du verre archéologique de l’Océan Indien (PhD thesis). Université d’Orléans. Dussubieux, L., Gratuze, B., & Blet-Lemarquand, M. (2010). Mineral soda alumina glass: Occurrence and meaning. Journal of Archaeological Science, 37, 1645–1655. Francis, P. Jr. (2000). Human ornaments. In S. E. Sidebotham, & W. Z. Wendrich (Eds.), Berenike 1998: Report of the 1998 excavations at Berenike and the survey of the Egyptian eastern desert, including excavations at Wadi Kalalat (pp. 211–225). Research School of Asian, African, and Amerindian Studies (CNWS). Francis. P. Jr. (2002). Asia’s Maritime Bead Trade 300 B.C. to the Present. University of Hawai’i Press. Francis, P., Jr. (2007). Personal adornments. In S. E. Sidebotham & W. Z. Wendrich (Eds.), Berenike 1999/2000: Report on the excavations at Berenike, including excavations in Wadi Kalalat and Siket, and the survey of the Mons Smaragdus Region (pp. 251–257). Cotsen Institute of Archaeology, University of California. Francis, P. Jr. (2013). The Beads. In J. Carswell, S. Deraniyagala, & A. Graham (Eds.), Mantai, City by the Sea (pp. 349–373). Linden Soft. Fattovich, R. (2018). From community to state: The development of the Aksumite polity (Northern Ethiopia and Eritrea), c. 400 BC–AD 800. Journal of Archaeological Research, 27, 249–285. https://doi.org/10.1007/s10814-018-9122-x Fattovich, R., Bard, K. A., Petrassi, L., & Pisano, L. (2000). The archaeological area of Aksum: A preliminary assessment. Istituto Universitario Orientale. Gould, D. A. (1999). The excavations at Shenshef. In S.E. Sidebotham, & W. Z. Wendrich (Eds.), Berenike 1997: Report of the 1997 excavations at Berenike and the survey of the Egyptian Eastern Desert, including excavations at Shenshef (pp. 371–384). Research School of Asian, African, and Amerindian Studies (CNWS), Universiteit Leiden. Gratuze, B., Pion, C., & Sode, T. (2021). Indian glass beads in Western and North Europe in Early Middle Age. In A. K. Kanungo, & L. Dussubieux (Eds.), Ancient Glass of South Asia— Archaeology, Ethnography and Global Connection. Springer Nature / IIT Gandhinagar. Hannibal-Deraniyagala, A. S. (2001). Beads from Tissamaharama—A typology of Sri Lankan glass and semi-precious stone beads. In H.-J. Weisshaar, H. Roth, & W. Wijeyapala (Eds.), Ancient Ruhuna. Sri Lankan-German Archaeological Project in the Southern Province 1, Materialen zur Allgemeinen und Vergleichenden Archäologie 58 (pp. 203–226). Philipp von Zabern. Hannibal-Deraniyagala, A. S. (2005). Beads from Anuradhapura and Tissamaharama Sri Lanka: Trade contacts in the early historic period. Journal of Indian Ocean Archaeology, 2, 21–24. Hannibal-Deraniyagala, A. S. (2013). Addendum 2. In J. Carswell, S. Deraniyagala, & A. Graham (Eds.), Mantai, City by the Sea (pp. 370–373). Linden Soft. Harlow, M. (2000). Glass, beads and pendants. In D. W. Phillipson (Ed.), Archaeology at Aksum, Ethiopia, 1993–7 (Vol. I, pp. 82–86, Vol. II: pp. 337–342, 458–460). British Institute in Eastern Africa. Hatke, G. (2013). Aksum and Nubia: Warfare, commerce, and political fictions in Ancient Northeast Africa. New York University Press.

Indian Glass Beads in Northeast Africa …

555

Lischi, S. (2018). Macroscopic analysis of the bead assemblage from the South Arabian port of Sumhuram, Oman (seasons 2000–2013). Arabian Archaeology and Epigraphy, 29, 65–92. Manzo, A. (2017). Eastern Sudan in its Setting, The archaeology of a region far from the Nile Valley, [=Cambridge Monographs in African Archaeology 94]. Archeopress Access Archaeology. Meyer, C. (1982). Roman Glass. In D. Whitcomb, & J. Johnson (Eds.), Quseir Al-Qadim 1980 (pp. 215–239). The American Research Center in Egypt, INC. Meyer, C. (1992). Glass from Quseir Al-Qadim and the Indian Ocean Trade. Oriental Institute of the University of Chicago. Morrison, H. M. (1989). The Beads. In S. C. Munro-Hay (Ed.), Excavations at Aksum (pp. 168–178). British Institute in Eastern Africa. Munro-Hay, S. C. (Ed.). (1989). Excavations at Aksum. Thames and Hudson. Peacock, D. (2011). Beads. In D. Peacock, & L. Blue (Eds.), Myos Hormos—Quseir Al-Qadim: Roman and Islamic Ports on the Red Sea. Vol. 2: Finds from Excavations 1999–2003 (pp. 57–78). BAR International Series 2286. Peacock, D., & Blue, L. (2001). Myos Hormos—Quseir al-Qadim Roman and Islamic Ports on the Red Sea: Roman and Islamic Ports on the Red Sea. Volume 2: Finds from the excavations 1999–2003. Archaeological Reports International Series 2286. Peacock, D., & Blue, L. (2006). Myos Hormos—Quseir Al-Qadim, Roman and Islamic Ports on the Red Sea: v. 1: The Survey and Excavations 1999–2003. Oxbow Books. Phillipson, D. W. (1997). The Monuments of Aksum. Addis Ababa University Press. Pion, C., & Gratuze, B. (2016). Indo-Pacific glass beads from the Indian subcontinent in Early Merovingian graves (5th–6th century AD). Archaeological Research in Asia, 6, 51–64. https:// doi.org/10.1016/j.ara.2016.02.005 Ricke, H. (1967). Ausgrabungen von Khor-Dehmit bis Bet El-Wali. Oriental Institute Nubian Expedition 2. Oriental Institute University of Chicago Press. Seeger, J. A. (2001). A preliminary report on the 1999 field season at Marsa Nakari. Journal of the American Research Center in Egypt, 38, 77–88. Seeger, J. A., & Sidebotham, S. E. (2005). Marsa Nakari: An ancient port on the Red Sea. Egyptian Archaeology, 26, 18–20. Sidebotham, S. E. (2000). The Excavations. In S. E. Sidebotham & W. Wendrich (Eds.), Berenike 1998, Report of the 1998 Excavations at Berenike and the Survey of the Egyptian Eastern Desert, including Excavations in Wadi Kalalat (pp. 3–148). Research School of Asian, African, and American Studies, CNWS, Universiteit Leiden. Sidebotham, S. E. (2007). Excavations. In W. Wendrich & S. E. Sidebotham (Eds.), Berenike 1999/2000, report on the excavations at Berenike, including excavations in Wadi Kalalat and Siket, and the survey of the Mons Smaragdus region (pp. 30–165). Cotsen Institute of Archaeology, University of California. Sidebotham, S. E. (2011). Berenike and the Ancient Maritime Spice Route. University of California Press. Sidebotham, S. E., Nouwens, H. M., Hense, A. M., & Harrell, J. A. (2004). Preliminary report on archaeological fieldwork at Sikait (Eastern Desert of Egypt), and environs: 2002–2003. SAHARA, 15, 7–30. Sidebotham, S. E., Hense, M., & Nouwens, H. M. (2008). The Red Land: The illustrated archaeology of Egypt’s Eastern Desert. American University in Cairo Press. Sidebotham, S. E., & Zych, I. (Eds.) (2011). Berenike 2008–2009: Report on the excavations at Berenike, including a survey in the Eastern Desert [=PCMA Excavation Series 1], (pp. 117–162). PCMA UW. Sidebotham, S. E., & Zych, I. (Eds.). (2017). Berenike 2010–2011. Report on two seasons of excavations at Berenike, including survey in the Eastern Desert and reports on earlier work [=PCMA Excavation Series 4]. PCMA UW. Strouhal, E. (1984). Wadi Qitna and Kalabsha-South: Late Roman–early Byzantine tumuli cemeteries in Egyptian Nubia I. Archaeology. Charles University.

556

J. Then-Obłuska

Then-Obłuska, J. (2015). Cross-cultural bead encounters at the Red Sea port site of Berenike, Egypt, Preliminary assessment (seasons 2009–2012). Polish Archaeology in the Mediterranean, 24(1), 735–777. https://doi.org/10.5604/01.3001.0010.0126 Then-Obłuska, J. (2016a). Beads and pendants from the tumuli cemeteries at Wadi Qitna and Kalabsha-South, Nubia. Beads: Journal of the Society of Bead Researchers, 28, 38–49. Then-Obłuska, J. (2016b). Early Makuria research project. Late antique beads and a Napatan amulet from the Early Makuria Phase II tumuli cemetery at El-Detti (about AD 450–550), season 2015. Der Antike Sudan, Mitteilungen Der Sudanarchäologischen Gesellschaft Zu Berlin E.v. Mittsag, 27, 139–145. Then-Obłuska, J. (2016c). Trade and faith in Nubian Early Makuria (AD 450–550): Macroscopic examination of personal adornments from el-Zuma in Nubia. Polish Archaeology in the Mediterranean, 25, 741–760. https://doi.org/10.5604/01.3001.0010.3009 Then-Obłuska, J. (2017a). Beads and pendants in the late Harbor Temple and harbor temenos in the Red Sea port of Berenike (seasons 2010–2013): Materials, techniques, functions and affiliations. Polish Archaeology in the Mediterranean, 26(2), 193–210. https://doi.org/10.5604/01.3001.0012. 1826 Then-Obłuska, J. (2017b). Between the Nile and the Ocean. The bead assemblage from Shenshef in the Eastern Desert (4th-6th centuries AD). Polish Archaeology in the Mediterranean, 26(1), 725–752. https://doi.org/10.5604/01.3001.0012.1816 Then-Obłuska, J. (2017c). Royal ornaments of a late Antique African Kingdom, Early Makuria, Nubia (AD 450–550) Early Makuria Research Project. Polish Archaeology in the Mediterranean, 26(1), 693–724. https://doi.org/10.5604/01.3001.0012.1817 Then-Obłuska, J. (2018a). Report on the Bead and Pendant Finds from Excavations at Maryam Anza, Seasons 2014–2016 (unpublished report). Then-Obłuska, J. (2018b). Beads and pendants from the Hellenistic to Early Byzantine Red Sea port of Berenike, Egypt, seasons 2014 and 2015. Polish Archaeology in the Mediterranean, 27(1), 203–233. https://doi.org/10.5604/01.3001.0013.2001 Then-Obłuska, J. (2018c). Short and long distance contacts of Late Antique Nubia: A view through the bead hole, In M. Honegger (Ed.),Nubian archaeology in the XXIst century: Proceedings of the Thirteenth International Conference for Nubian Studies, Neuchâtel, 1st–6th September 2014 [=Orientalia lovanensia analecta 273] (pp. 587–595). Peeters. Then-Obłuska, J. (2018d). Bead trade in Roman ports—a view from the Red Sea port Marsa Nakari. In A. Manzo, Ch. Zazzaro, & D. J. de Falco (Eds.), Stories of Globalisation: The Red Sea and the Persian Gulf from late prehistory to early modernity. Selected papers of Red Sea Project VII (pp. 264–280). Brill Then-Obłuska, J. (n.d.). Beyond the emerald, Beads and pendants from Mons Smaragdus mining settlement at Sikait, Egypt, Seasons 2002–2003. Then-Obłuska, J., & Dussubieux, L. (2016). Glass bead trade in the Early Roman and Mamluk Quseir ports—A view from the Oriental Institute Museum assemblage. Archaeological Research in Asia, 6, 81–103. https://doi.org/10.1016/j.ara.2016.02.008 Then-Obłuska, J., & Ple¸sa, A. D. (2019). Roman to Islamic beads and pendants from Matmar and Mostagedda Middle Egypt. Beads: Journal of the Society of Bead Researchers, 31, 50–74. Then-Obłuska, J., & Wagner, B. (2017). Glass bead trade in Northeast Africa in the Roman period— a view according to The Museum of Archaeology University of Stavanger assemblage. In S. Wolf, & A. de Pury-Gysel (Eds.), Annales du 20e Congrès de l’Association Internationale pour l’Histoire du Verre (AIHV 20), Fribourg/Romont 7–11 septembre 2015 (pp. 248–256). Then-Obłuska, J., & Wagner, B. (2019a). Glass beads and pendants from Meroitic and Nobadian Lower Nubia, Sudan: Chemical compositional analysis using laser ablation-inductively coupled plasma-mass spectrometry. Archaeometry, 61(4), 856–873. https://doi.org/10.1111/arcm.12465 Then-Obłuska, J., & Wagner, B. (2019). Glass bead trade in Northeast Africa, the evidence from Meroitic and Post-Meroitic Nubia [=PAM Monograph Series 10]. University of Warsaw Press.

Indian Glass Beads in Northeast Africa …

557

Tomber, R. (1998). The pottery. In S. E. Sidebotham, & W. Z. Wendrich (Eds.), Berenike 1996: Report of the 1996 excavations at Berenike (Egyptian Red Sea Coast) and the survey of the Eastern Desert (pp. 163–180). CNWS. Tomber, R. (2005). Troglodites and trogodites: Exploring interaction on the Red Sea during the Roman period. In J. C. Starkey (Ed.), People of the Red Sea: Proceedings of Red Sea Project II (pp. 41–49). Archaeopress, BAR International Series 1395. Tomber, R. (2008). Indo-Roman trade: From pots to pepper. Duckworth. Whitcomb, D. (1979). Small Finds. In D. Whitcomb, & J. H. Johnson (Eds.), Quseir Al-Qadim 1978: Preliminary Report American Research Centre in Egypt—Reports 1.