Sasanian and Islamic Settlement and Ceramics in Southern Iran (4th to 17th Century AD): The Williamson Survey

By Seth M. N. Priestman and Derek Kennet

This monograph comprises the final publication of a study supported by the British Institute of Persian Studies and undertaken by Seth Priestman and Derek Kennet at the University of Durham. The work presents and analyses an assemblage of just under 17,000 sherds of pottery and associated paper archives resulting from one of the largest and most comprehensive surveys ever undertaken on the historic archaeology of southern Iran. The survey was undertaken by Andrew George Williamson (1945–1975), a doctoral student at Oxford University between 1968 and 1971, at a time of great progress and rapid advance in the archaeological exploration of Iran.

The monograph provides new archaeological evidence on the long-term development of settlement in Southern Iran, in particular the coastal region, from the Sasanian period to around the 17th century. The work provides new insights into regional settlement patterns and changing ceramic distribution, trade and use. A large amount of primary data is presented covering an extensive area from Minab to Bushehr along the coast and inland as far as Sirjan. This includes information on a number of previously undocumented archaeological sites, as well as a detailed description and analysis of the ceramic finds, which underpin the settlement evidence and provide a wider source of reference.

By collecting carefully controlled archaeological evidence related to the size, distribution and period of occupation of urban and rural settlements distributed across southern Iran, Williamson aimed to reconstruct the broader historical development of the region. Due to his early death the work was never completed. The key aims of the authors of this volume were to do justice to Williamson’s remarkable vision and efforts on the one hand, and at the same time to bring this important new evidence to ongoing discussions about the development of southern Iran through the Sasanian and Islamic periods.

• Language: English
• Publisher: Oxbow Books
• Release date: Oct 31, 2023
• ISBN: 9798888570531

Seth M. N. Priestman

Seth Priestman is an archaeologist and ceramic specialist with a track record of research and publication across Arabia, Iran, the Caucasus, East Africa, and Japan from the post-Iron Age to the 20th century. His main research focuses on long term changes in the maritime economy of the Indian Ocean region and the Gulf using ceramic finds as a proxy for wider historic reconstruction. He has previously held research positions with the British Museum, and the Universities of Durham, Edinburgh, and Southampton. He has also worked widely in private consultancy for museum development and heritage protection projects across the GCC. He currently holds the post of Honorary Research Fellow in the Department of Archaeology at Durham University.

Book preview

Published in the United Kingdom in 2023 by

OXBOW BOOKS

The Old Music Hall, 106–108 Cowley Road, Oxford, OX4 1JE

and in the United States by

OXBOW BOOKS

1950 Lawrence Road, Havertown, PA 19083

© Oxbow Books and the authors 2023

Hardback Edition: ISBN 979-8-88857-052-4

Digital Edition: ISBN 979-8-88857-053-1 (epub)

A CIP record for this book is available from the British Library

Library of Congress Control Number: 2023942348

All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording or by any information storage and retrieval system, without permission from the publisher in writing.

Printed in Malta by Melita Press

Typeset in India by Lapiz Digital Services, Chennai.

For a complete list of Oxbow titles, please contact:

UNITED KINGDOM

Oxbow Books

Telephone [phone]

Email: [e-mail]

www.oxbowbooks.com

UNITED STATES OF AMERICA

Oxbow Books

Telephone [phone], Fax [phone]

Email: [e-mail]

www.casemateacademic.com/oxbow

Oxbow Books is part of the Casemate Group

Front cover: Kuh-i Istakhr from Kuh-i Qalah, Fars province in 1969. Photograph by Andrew Williamson (WA-IN0329).

Contents

List of figures

List of tables

Preface

Conventions and explanatory notes

Acknowledgements

1. Introduction

2. The Williamson Collection

2.1 An archaeological career

Williamson’s early years (1945–67)

First prospective survey (1968)

The detailed survey (1969–71)

Williamson and Prickett

An overview of the survey

Writing up the results

Distractions and new opportunities (1972–75)

Posthumous work on the Collection

The challenge of a comprehensive study

2.2 Survey methodology and concept

Survey methodology

Site identification

Vehicular survey

Site recording and sampling

Value of the data

2.3 Processing the Collection

The Collection as it stood

Sherd numbering

Initial catalogue

Site location

Ceramic classification

Fabric description

The refined classification

3. Testing the reliability of the survey

3.1 Issues of interpreting the data

Discoveries made through the work

Sampling and selection

3.2 Site card analysis

Sherd comparison

Site comparison

Discussion

3.3 Stein/Williamson comparison

The Stein Collection

Comparison of sites

Discussion

3.4 The Bushehr Hinterland Survey (2004)

Revisiting Williamson’s sites

Background to the investigation

Discussion

3.5 The Hormuzgan Survey (2005)

Site and area selection

Comparison of assemblages

Secure identifications

Casual identifications

Non-Williamson sites

Discussion

3.6 The Williamson Collection in Tehran

The Williamson Collection in Iran

The Tehran Collection

Object labels

Finds selection

Tehran Collection sites

Discussion

3.7 Reliability of the survey results

4. Regional settlement patterns

4.1 Regional surveys

Settlement dynamics

Shah Maran-Daulatabad Survey

Eastern Bardsir Survey

The Bushehr Hinterland Survey

Diyala Basin Survey

The question of recent settlement change

Marv-Dasht Survey

Ras al-Khaimah Survey

4.2 Analysis of settlement change

Dating of pottery and sites

The initial model

Improving the site periodisation

Regional distribution of settlement

Coastal areas

Inland areas

Trends and limitations

5. Patterns of ceramic distribution

5.1 Regional reorientation of ceramic production

Alkaline-glazed ware

Distribution of alkaline-glazed ware

Honeycomb ware and torpedo jars

Systemic change

The Samarra horizon

The Samarra horizon in Iran

Glazed wares of the 11th to 13th century

5.2 Eastern trade

South Asian pottery

East Asian pottery

Regional distribution of East Asian pottery

Emporia ‘migration’

Shifting economic zones

6. Conclusion

Appendix I – Class catalogue

Introduction

Record explanation

Class index

Class code identification

Class descriptions

Appendix II – Fabric catalogue

Introduction

Record explanation

Fabric index

Fabric description

Appendix III – Site catalogue

Introduction

Abbreviations

Region prefixes

Site catalogue

Region A – Rud-i Shur to Bandar-i Lengeh

Region AA – Hormuz Island

Region AE – Kish Island

Region B – Lengeh to Qalat-i Abdl Rahman

Region D – Neran to Naband

Region DD – Tepe Dasht-i Deh

Region F – Tahiri to Rud-i Mond

Region H – Bushehr peninsula

Region J – Rud-i Shur to Minab

Region K – Minab plain

Region L – Minab to Jask

Region P – Halil-Rud

Region Q – Rudan and Bulak

Region R – Betweek Jiroft and Fars

Region S – Sirjan

Region V – Shiraz to Region B

Region Z – Bam/Rayen

Marv-Dasht Survey

Hormuzgan Survey

Appendix IV – Pottery forms

Published sources

Williamson Collection Project archive contents

Colour plates

List of figures

Frontispiece. Survey by Land Rover near to Lake Bakhtegan, Fars province in 1968. Photograph by Andrew Williamson (WA-IN0238).

Fig. 2.1. Andrew George Williamson in his early 20s. Image reproduced with kind permission of the Williamson family.

Fig. 2.2. A sketch map showing the routes taken by Williamson during his initial survey undertaken in 1968–69 and projected routes he intended to cover (see WA-R03: 2).

Fig. 2.3. Coastal survey near to Jask in 1969. Photograph by Andrew Williamson (WA-IN0276).

Fig. 2.4. Group photograph taken at Tepe Yahya in 1971. Andrew Williamson and Martha Prickett sitting together in far left of the second row. Reproduced with kind permission of Karl Lamberg-Karlovsky (after Lamberg-Karlovsky & Potts 2001: fig. F.9).

Fig. 2.5. A map prepared by Williamson showing the areas surveyed ‘intensively’ and ‘non-intensively’ between 1968 and 1971 as well as the areas that he intended to survey if he had taken up the Randall MacIver Junior Research Fellowship at Queen’s College, Oxford, to research the history of pearling in the southern Persian Gulf.

Fig. 2.6. A map showing the location of individual site codes, major rivers and ground above 500 ft ( c .152 m; shaded in grey). The sites lettered A–M in the Minab area are: A = K7–8; B = K6, K70; C = K33; D = K9, K13; E = K19; F = K14–15, K66, K169; G = K143, K145; H = K20–25, K27–28, K62–63; I = K29–30; J = K40–43, K54, K67–69; K = K26; L = K1–2, K51; M = K162, 170; N = K103; O = K102, K130–31.

Fig. 2.7. Williamson’s survey regions based on sites which have been relocated from the toponyms that appear on the card index of sites.

Fig. 3.1. A map showing the location of sites that have been used for comparison. The marked site codes are those used by Williamson.

Fig. 3.2. The proportion (above) and number (below) of sherds by period from Shahr-i Daquianus in the Stein and Williamson Collections.

Fig. 3.3. The proportion (above) and number (below) of sherds by period from Gust-i Burjan in the Stein and Williamson Collections.

Fig. 3.4. The proportion (above) and number (below) of sherds by period from Tump-i Kharg in the Stein and Williamson Collections.

Fig. 3.5. The proportion (above) and number (below) of sherds by period from Tump-i Hazar Mardi in the Stein and Williamson Collections.

Fig. 3.6. The proportion (above) and number (below) of sherds by period from Ziarat-i Danial Gulashgird in the Stein and Williamson Collections.

Fig. 3.7. The proportion (above) and number (below) of sherds by period from Kung in the Stein and Williamson Collections.

Fig. 3.8. The proportion (above) and number (below) of sherds by period from Siraf in the Stein and Williamson Collections.

Fig. 3.9. The location of survey projects conducted in southern Iran in the Bushehr hinterland in 2004 and the Minab plain in 2005 in relation to the Williamson Survey coverage.

Fig. 3.10. Distribution of sites recorded during the 2004 Bushehr Hinterland Survey.

Fig. 3.11. Left: Williamson’s map of later Sasanian-period settlement on the Bushehr peninsula (after Whitehouse & Williamson 1973: fig. 4); top right: the rock-cut ditch along the southern side of Rishahr fort; bottom right: the eroding shoreline section in front of Rishahr fort.

Fig. 3.12. Comparison of the total number of sites with more than 2% of the assemblage falling within a particular period, as a percentage of the total number of registered periods of occupation in Williamson’s Region H and the Bushehr Hinterland Survey.

Fig. 3.13. Location map of the Minab plain and Rudan area showing the sites that were visited during the survey. Hormuz Island (PK12) is just off the map.

Fig. 3.14. View of Gishnau looking south-west from the direction of the village. The dense surface debris visible in the foreground stretches as far as the line of trees in the distance.

Fig. 3.15. Comparison of Williamson and Hormuzgan Survey assemblages from Gishnau (K143, K145 and PK1.1–2) showing the adjusted total (below) and percentage of sherds (above) by period.

Fig. 3.16. A detailed map of Nakhl Ibrahim showing the extent of the area covered by field walking (inside the broken line) and the position of the individual sites referred to below.

Fig. 3.17. New houses being constructed on elevated rock and gravel foundation platforms on the sabkha plain next to Tiab, in a similar situation to those at Nakhl Ibrahim.

Fig. 3.18. A palm-garden wall at the back of Nakhl Ibrahim that has eroded away leaving only a slightly elevated strip and a thick concentration of potsherds scattered along its length.

Fig. 3.19. Comparison of Williamson and Hormuzgan Survey assemblages from the main site at Nakhl Ibrahim (K36/PK2.9) showing the adjusted total (below) and percentage of sherds (above) by period.

Fig. 3.20. Comparison of Williamson and Hormuzgan Survey assemblages from all the sites at Nakhl Ibrahim (K30, K32–36 and PK2.1–9) showing the adjusted total (below) and percentage of sherds (above) by period.

Fig. 3.21. Comparison of Williamson and Hormuzgan Survey assemblages from the Middle Islamic sites mid-way between Dehu and Nakhl Ibrahim (K29, K31 and PK7) showing the adjusted total (below) and percentage of sherds (above) by period.

Fig. 3.22. Boats of the same design as those constructed in the Tiab boatyard today docked in a line along the tidal creek that forms Tiab’s harbour.

Fig. 3.23. Comparison of Williamson and Hormuzgan Survey assemblages from the main site at Tiab (K70/PK10.3) showing the adjusted total (below) and percentage of sherds (above) by period.

Fig. 3.24. View of the prominent tell site of Tump-i Mehta (PK13) with gullied erosion channels down its sides.

Fig. 3.25. View of the low industrial tell site of Tump-i Ali (PK15) with onion cultivation in the foreground.

Fig. 3.26. Earth extraction pits on the north-west side of Tump-i Ali (PK15) where the fabric of the mound on this side is shown to be composed almost entirely of broken pottery.

Fig. 3.27. The south-east side of Tump-i Ali (PK15) showing the ground surface made up of a dense accumulation of slag probably associated with metalworking.

Fig. 3.28. Comparison of Williamson and Hormuzgan Survey assemblages from Tepe Mauru (Q17/PK16) showing the adjusted total (below) and percentage of sherds (above) by period.

Fig. 3.29. View of the fertile Rudan plain looking out from the fort of Qalat-i Khumis (PK17).

Fig. 3.30. Impenetrable undergrowth and palm-garden cultivation surrounding the village of Saravan.

Fig. 3.31. A detailed map of Hakemi showing the five blocks (indicated by broken lines) covered intensively by field walking. The five transects were numbered PK11.1 starting at the bottom up to PK11.4 next to the road and PK11.5 to the right.

Fig. 3.32. Comparison of Williamson and Hormuzgan Survey assemblages from all sites at Hakemi (K40, 42–43, K54, K65, K67–69 and PK11.1–5) showing the adjusted total (below) and percentage of sherds (above) by period.

Fig. 3.33. Comparison of Williamson and Hormuzgan Survey assemblages from all sites at Hakemi (K40, 42–43, K54, K65, K67–69 and PK11.1–5) showing the percentage of sherds by period considering only the 3rd-to 17th-century assemblage.

Fig. 3.34. The narrow inhospitable strip between the coast and the Zindan hills on the road south of the Minab plain towards Jask.

Fig. 3.35. Small fishing boats pulled up along the beachfront at Kuhestak.

Fig. 3.36. Comparison of Williamson and Hormuzgan Survey assemblages from Jarun Island (AA1–3, AAP, AAC and PK12) showing the adjusted total (below) and percentage of sherds (above) by period.

Fig. 3.37. A partially exposed jar burial at PK6 with the top half of the vessel missing. Small flecks of human bone can be seen surrounding the jar.

Fig. 3.38. A map showing the location of site toponyms represented in the Tehran portion of the Williamson Collection in relation to the areas of ‘intensive’ and ‘non-intensive’ survey.

Fig. 4.1. A map showing the location of selected regional surveys in Iraq, southern Iran and the UAE in relation to the Williamson Survey. The area of the Williamson Survey is a conservative representation based only on the sites that have been relocated.

Fig. 4.2. Number of sites by period in the Shah Maran-Daulatabad basin (based on Prickett 1976: 176; 1979: 52). Dates are derived from the following periods: pre-Yahya Moradabad phase, Yahya VI, Yahya VB, Yahya VA, Iblis IV–V, Yahya IVC down to the next phase, Yahya II.

Fig. 4.3. Number of sites by period in the eastern Bardsir area (based on the catalogue of sites in Sajjadi & Wright 1990: 31–32).

Fig. 4.4. Plot of the number of new, old and abandoned sites within the Dashtestan region from the Chalcolithic to Late Islamic periods, based on the results of the Bushehr Hinterland Survey (after Carter et al. 2006: table 3).

Fig. 4.5. Plot of the number of new, old and abandoned sites within the Diyala basin from periods ranging from the 4th millennium BC to the 15th century AD (based on Adams 1965: 34, 37, tables 10–20, 23, 24).

Fig. 4.6. Site counts by period from the Marv-Dasht Survey (based on Williamson’s period divisions presented in Sumner & Whitcomb 1999: 314, table 2).

Fig. 4.7. Site count by period based on the 1994 survey of the Sir and Jiri plains in northern Ras al-Khaimah (taken from Kennet 2002: fig. 9).

Fig. 4.8. Number of sites by period for the whole of the Williamson Survey dataset.

Fig. 4.9. Number of sherds by period across the whole of the Williamson Survey dataset.

Fig. 4.10. Number of sherds by period based on the ‘whole periodisation’, where classes are assigned to a single period, and the ‘fractional periodisation’, where fractional values are assigned to classes that cross-cut period divisions.

Fig. 4.11. Number of sites by period based on the ‘whole’ and ‘fractional’ periodisation and number of sites with more than 10% of the assemblage relating to each period (for the last value see below).

Fig. 4.12. Number of sherds in the Williamson Collection broken down by region. For region locations see Fig. 2.4 .

Fig. 4.13. Number of sites with sherds represented within the Williamson Collection broken down by survey region. For region locations see Fig. 2.4 .

Fig. 4.14. Region H, the Bushehr peninsula, showing the number of sites as well as the number of sites with more than 10% of the sample from each period.

Fig. 4.15. Region K, the Minab plain, showing the number of sites as well as the number of sites with more than 10% of the sample from each period.

Fig. 4.16. Regions F and D, the closest available to the Bushehr peninsula, showing the number of sites as well as the number of sites with more than 10% of the sample from each period.

Fig. 4.17. Regions A, AA, J & L, all adjacent to the Minab plain, showing the number of sites as well as the number of sites with more than 10% of the sample from each period.

Fig. 4.18. Regions B and AE, showing the number of sites as well as the number of sites with more than 10% of the sample from each period.

Fig. 4.19. Inland Regions P, Q, R, S, V and Z, showing the number of sites as well as the number of sites with more than 10% of the sample from each period.

Fig. 4.20. A selection of the most important regions presented in geographical order from west to east and inland, showing periods in which sites were abandoned, founded in new locations or continued to be occupied from the preceding period.

Fig. 5.1. Sites marked in red are those that have been relocated with TURQ in their assemblages.

Fig. 5.2. Percentage of sites dated to between the pre-3rd and 11th centuries with assemblages containing TURQ broken down by region.

Fig. 5.3. Percentage of 3rd-to 9th-century sites by region with TORP and HONEY, and with TORP, HONEY and TRC combined.

Fig. 5.4. Sites marked in black are those that have produced finds of 9th-to 10th-century ‘Samarra horizon’ wares. Regions outlined and labelled are referred to below.

Fig. 5.5. Number of sites by region with early, middle and late Samarra horizon classes.

Fig. 5.6. Sites with early, middle and late Samarra horizon classes as a percentage of the total number of 9th-to 10th-century sites in each region with five or more sites belonging to the period.

Fig. 5.7. Comparison of the percentage of 9th- to 10th-century sites with Samarra horizon classes to the percentage of 11th-to 13th-century sites with sgraffiato and sgraffiato-related classes.

Fig. 5.8. Sites marked in black are those that have produced finds of the four main South Asian classes in the Williamson Collection: IRPW, HARMIC, IRAB and SBBW. Regions outlined and labelled are referred to below.

Fig. 5.9. Count of the number of sites within each region producing assemblages containing sherds of each of the main South Asian classes: IRPW, HARMIC, IRAB and SBBW.

Fig. 5.10. Sites with the four main classes of South Asian pottery expressed as a percentage of the total number of sites within each Region.

Fig. 5.11. East Asian pottery as a proportion of the assemblage through the combined sequences of Kush and al-Mataf, Kush Periods V–VIII and al-Mataf Periods I–REC (after Kennet 2004: table 32).

Fig. 5.12. Comparison of the Kush/al-Mataf sequence in the Persian Gulf and the Shanga sequence in East Africa in terms of the percentage of East Asian imports during equivalent periods (after Kennet 2004: table 31).

Fig. 5.13. Comparison of the number of coastal and inland sites with assemblages containing East Asian pottery by period.

Fig. 5.14. Sites receiving East Asian imports displayed as a percentage of the total number of sites sampled across the survey area.

Fig. 5.15. Percentage of sites in the Upper, Middle and Lower Persian Gulf and inland areas with each of the main East Asian wares. Tang = CHANG, DUSUN & GWSW; WW = all WW from East Asian Period 2 ( see Table 5.1 ); LQC = LQC.1–4 & BW = CBW.1–45, VBW.1–2. Upper = Regions H, F & D; Middle = AE & B; Lower = Regions AA, A, J, K & L; Inland = Regions P, Q, S, V & Z.

Fig. 5.16. Number of CBW and VBW sherds in the Williamson Collection assemblage broken down by period.

Fig. 5.17. Number of East Asian sherds by period from the islands of Kish and Hormuz represented in the Williamson Collection.

Fig. IV.1. a–b = CLINKY/Form 1; c–d = CLINKY/Form 2; e–i = SMAG.A/ Form 2; j–l = SMAG.B/Form 4; m–p = SMAG.B/Form 5; q–t = SMAG.B/Form 6, scale 1:3.

Fig. IV.2. a–m = SMAG.A/Form 3, scale 1:3.

Fig. IV.3. a–k = SMAG.A/Form 1, scale 1:3.

Fig. IV.4. a–e = SMAG.B/Form 1; f–h = SMAG.B/Form 2; i–k = SMAG.B/Form 3, scale 1:3.

Fig. IV.5. a–e = SMAG.B/Form 7; f = SMAG.B/Form 8; g–h = SMAG.B/Form 9; i–j = SMAG.B/ Form 10; k–m = SMAG.B/Form 11; n–o = SMAG.A/Form 4, scale 1:3.

Fig. IV.6. a–d = SMAG.C/Form 1; e–j = SMAG.C/Form 2; k–m = SMAG.C/Form 3, scale 1:3.

Fig. IV.7. a–f = SMAG.RC/Form 1, scale 1:3.

Fig. IV.8. a–d = SM.RC: 01, scale 1:3.

Fig. IV.9. a–q = SMAG.RC/Form 2, scale 1:3.

Fig. IV.10. a–i = SMAG.RC/Form 3, scale 1:3.

Fig. IV.11. a–h = SMAG.RC/Form 3, scale 1:3.

Fig. IV.12. a–k = GRIT/Form 1, scale 1:3.

Fig. IV.13. a–h = GRIT/Form 2, scale 1:3.

Fig. IV.14. a–g = GRIT/Form 3; h–l = GRIT/Form 4, scale 1:3.

Fig. IV.15. a–g = FIGRIT/Form 1; h–n = FIGRIT/Form 2, scale 1:3.

Fig. IV.16. a–d = FIGRIT/Form 3; e–x = FIGRIT/Form 4, scale 1:3.

Fig. IV.17. a–k = FIGRIT/Form 5, scale 1:3.

Fig. IV.18. a–g = FIGRIT/Form 5, scale 1:3.

Fig. IV.19. a–d = FIGRIT/Form 6; e–i = FIGRIT/Form 7, scale 1:3.

Fig. IV.20. a–i = VITFIG/Form 1, scale 1:3.

Fig. IV.21. a–j = VITFIG/Form 2; k–q = FINT/Form 1, scale 1:3.

Fig. IV.22. a–e = VITFIG/Form 3; f–j = FINT/Form 3, scale 1:3.

Fig. IV.23. a–t = FINT/Form 2, scale 1:3.

Fig. IV.24. a–m = FINT/Form 4, scale 1:3.

Fig. IV.25. a–h = FINT/Form 4; i–k = FINT.B/Form 1; l–o = FINT.B/Form 2, scale 1:3.

Fig. IV.26. a–h = GROG/Form 1, scale 1:3.

Fig. IV.27. a–i = GROG/Form 2, scale 1:3.

Fig. IV.28. a–g = ORG.H/Form 1; h–l = ORG.H/Form 2, scale 1:3.

Fig. IV.29. a–h = ORG.H/Form 2, scale 1:3.

Fig. IV.30. a–f = ORG.H/Form 3, scale 1:3.

Fig. IV.31. a–e = ORG.H/Form 4, scale 1:3.

Fig. IV.32. a–b = ORG.H/Form 4; c–d = ORG.H/Form 5; e–g = ORG.I/Form 1; h = ORG.I/Form 5, scale 1:3.

Fig. IV.33. a–c = ORG.I/Form 2; d–g = ORG.I/Form 3, scale 1:3.

Fig. IV.34. a–c = ORG.I/Form 4; d–f = ORG.S/Form 1, scale 1:3.

Fig. IV.35. a–c = ORG.S/Form 1; d–f = ORG.S/Form 2; h–i = ORG.S/Form 3, scale 1:3.

Fig. IV.36. a–d = ORG.S/Form 3; e–g = LAG/Form 1, scale 1:3.

Fig. IV.37. a–c = SWIS/Form 1; d–e = SWIS/Form 2; f = SWIS/Form 3; g–k = CHAM.1/Form 1, scale 1:3.

Fig. IV.38. a–d = CHAM.1/Form 2; e–h = CHAM.1/Form 3; i–k = CHAM.1/Form 4, scale 1:3.

Fig. IV.39. a–b = CHAM.1/Form 5; c–j = CHAM.2/Form 1, scale 1:3.

Fig. IV.40. a–e = CHAM.2/Form 2; f–i = CHAM.2/Form 4; j–n = CHAM.3/Form 4, scale 1:3.

Fig. IV.41. a–h = CHAM.3/Form 1; i–j = CHAM.3/Form 3, scale 1:3.

Fig. IV.42. a–l = CHAM.3/Form 2, scale 1:3.

Fig. IV.43. a–h = SHABUR.A/Form 1; j–n = SHABUR.A/Form 2, scale 1:3.

Fig. IV.44. a–h = SHABUR.B/Form 1; i–l = SHABUR.B/Form 2, scale 1:3.

Fig. IV.45. a–d = TORP.1/Form 1; e–h = TORP.2/Form 1; i–j = TORP.3/Form 1, scale 1:3.

Fig. IV.46. a–c = TORP.4/Form 1; d = TRC.1/Form 1; e = TRC.1/Form 3; f–g = TRC.1/Form 2; h = TRC.2/Form 1; i = TRC.1/Form 4, scale 1:3.

Fig. IV.47. a–b = TRC.2/Form 2; c–g = LIME/Form 1, scale 1:3.

Fig. IV.48. a–e = IRPW/Form 1; f–l = IRAB/Form 1, scale 1:3.

Fig. IV.49. a–f = IRAB/Form 2; g–i = IRAB/Form 3; j–l = IRAB/Form 4; m = IRAB/Form 5; n–o = IRAB/Form 6; p = IRAB/Form 7, scale 1:3.

Fig. IV.50. a–i = SBBW/Form 1; j–k = SBBW/Form 0; l–t = HARMIC/Form 1, scale 1:3.

Fig. IV.51. a–h = HARMIC/Form 2; i–q = HARMIC/Form 3, scale 1:3.

Fig. IV.52. a–d = BUFRAB/Form 1; e = BUFRAB/Form 2; f = BUFRAB/Form 3; g = BUFRAB/Form 4; h–i = FIRE/Form 1; j = EACOP/Form 1; k = SLIP.PBR/Form 1; l–o = SLIP.B/Form 2, scale 1:3.

Fig. IV.53. a–k = SLIP.B/Form 1, scale 1:3.

Fig. IV.54. a–f = SLIP.TB/Form 1, scale 1:3.

Fig. IV.55. a–f = SLIP.TB/Form 2; g–h = SLIP.TB/Form 3, scale 1:3.

Fig. IV.56. a–c = SLIP.TB/Form 4; d–n = SLIP.TB/Form 5, scale 1:3.

Fig. IV.57. a–h = SLIP.R/Form 1; i = SLIP.R/Form 2; j–m = SLIP.R/Form 3; n–r = SLIP.R/Form 4; s = SLIP.R/Form 5, scale 1:3.

Fig. IV.58. a–c = SLIP.R/Form 6; d–f = SLIP.R/Form 7; g–i = SLIP.R/Form 8; j–n = SLIP.R/Form 9, scale 1:3.

Fig. IV.59. a–k = SLIP.R/Form 10; l–s = SLIP.R/Form 11, scale 1:3.

Fig. IV.60. a–e = TURQ.1/Form 1; f–h = TURQ.1/Form 2; i–m = TURQ.1/Form 3; n–r = TURQ.1/Form 4, scale 1:3.

Fig. IV.61. a–f = TURQ.1/Form 5; g–j = TURQ.1/Form 6; k–r = TURQ.1/Form 7; s–t = TURQ.1/Form 8, scale 1:3.

Fig. IV.62. a = TURQ.1/Form 9; b–d = TURQ.2/Form 10; e–g = TURQ.1/Form 11; h–l = TURQ.1/Form 12; m–p = TURQ.1/Form 13, scale 1:3.

Fig. IV.63. a–c = TURQ.1/Form 14; d–h = TURQ.1/Form 15; i–l = TURQ.1/Form 16, scale 1:3.

Fig. IV.64. a–e = TURQ.1/Form 17; f–i = TURQ.1/Form 18, scale 1:3.

Fig. IV.65. a–d = TURQ.1/Form 19; e–i = TURQ.1/Form 20; j–n = TURQ.1/Form 21, scale 1:3.

Fig. IV.66. a–d = TURQ.1/Form 22; e–j = TURQ.1/Form 23; k–p = TURQ.1/Form 25, scale 1:3.

Fig. IV.67. a–f = TURQ.1/Form 26; g–h = TURQ.1/Form 27; i = TURQ.1/Form 28; j–m = TURQ.2/Form 29, scale 1:3.

Fig. IV.68. a–d = TURQ.2/Form 30; e–j = TURQ.3/Form 31, scale 1:3.

Fig. IV.69. a–b = TURQ.3/Form 32; c–d = TURQ.RC/Form 1; e–f = MGPAINT–1/Form 1; g–h = MGPAINT–2/Form 2, scale 1:3.

Fig. IV.70. a–d = WWSL/Form 1; e–g = CHANG/Form 1; h–k = GWSG/Form 1; l–n = DUSUN/Form 1, scale 1:3.

Fig. IV.71. a–e = WWS.1/Form 1; f–h = WWS.2/Form 1; i–j = WWS.3/Form 1; k–s = WWS.4/Form 1, scale 1:3.

Fig. IV.72. a–d = WWS.6/Form 1; e–h = WWS.7/Form 1; i–k = WWS.8/Form 1; l–r = WWS.10/Form 1, scale 1:3.

Fig. IV.73. a = WWF/Form 1; b–c = WWG.1/Form 1; d–e = WWG.2/Form 1; f = WWJ.2/Form 1; g = WWJ.4/Form 1; h–l = WW.3/Form 1; m = WW.4/Form 1, scale 1:3.

Fig. IV.74. a–m = QING.1/Form 1; n = QING.2/Form; o–r = DEHUA.1/Form 1; s = DEHUA.1/Form 2; t = DEHUA.2/Form 1, scale 1:3.

Fig. IV.75. a–c = LQC.1/Form 1; d–g = LQC.1/Form 2, scale 1:3.

Fig. IV.76. a–f = LQC.1/Form 3; g–h = LQC.2/Form 1; i–j = LQC.2/Form 3; k = LQC.1/Form 5; l = LQC.2/Form 5, scale 1:3.

Fig. IV.77. a–c = LQC.2/Form 2; d–e = LQC.2/Form 4, scale 1:3.

Fig. IV.78. a–b = LQC.2/Form 6; c–d = LQC.2/Form 7; e–f = LQC.2/Form 10; g–i = LQC.2/ Form 11; j–k = LQC.2/Form 12; l–m = LQC.2/Form 14, scale 1:3.

Fig. IV.79. a–b = LQC.2/Form 13; c–e = LQC.2/Form 15; f–g = LQC.2/Form 16; h = LQC.2/ Form 17; i = LQC.2/Form 18, scale 1:3.

Fig. IV.80. a–f = LQC.3/Form 1; g–h = LQC.3/Form 2; i–k = LQC.3/Form 3; l–m = LQC.3/ Form 5; n = LQC.3/Form 6, scale 1:3.

Fig. IV.81. a–f = LQC.3/Form 4; g–i = LQC.3/Form 7; j = LQC.3/Form 8, scale 1:3.

Fig. IV.82. a = GDC.1/Form 1; b = GDC.2/Form 1; c–f = GDC.3/Form 1; g–h = GDC.4/ Form 1; i–j = STONE-GRY/Form 1; k–m = STONE-GRY/Form 2; n–o = STONE-GRY/Form 3, scale 1:3.

Fig. IV.83. a–d = STONE-GRY/Form 4; f–g = STONE-GRY/Form 5, scale 1:3.

Fig. IV.84. a–f = STONE-BUR/Form 1, scale 1:3.

Fig. IV.85. a–c = STONE-BUR/Form 2; d–i = STONE-BUR/Form 3; j–n = STONE-THAI/ Form 1, scale 1:3.

Fig. IV.86. a–e = STONE-THAI/Form 2; f–g = STONE-THAI/Form 3; h = STONE-THAI/ Form 4; i = STONE-THAI/Form 5; j–n = LIB/Form 1, scale 1:3.

Fig. IV.87. a–b = LIB/Form 2; c–d = LIB/Form 3; e–k = DAB/Form 1, scale 1:3.

Fig. IV.88. a–j = MTB.1/Form 1; k–r = MTB.2/Form 1, scale 1:3.

List of tables

Table. 2.1. Summary of known activities during the course of Williamson’s life, and in particular his archaeological survey of southern Iran from 1968 to 1971.

Table. 2.2. Primary sorting of material as indicated by the box labels.

Table. 2.3. Categories of sorting represented by the bags and groups of sherds found loose within the boxes.

Table. 3.1. Comparison of the number of sites per survey region represented by the sherds in the Ashmolean Collection and those listed on the card index.

Table. 3.2. Comparison of the number of sherds in the Williamson Collection and the number of finds listed on the card index of sites for three sample regions. The right-hand column provides the percentage of sherds in the Collection as a proportion of the total number of finds listed on the card index.

Table. 3.3. Counts of the total number of sherds listed on the card index and those represented in the Ashmolean Collection for a selection of classes from three different regions. The bottom row provides the ratio of sherds on the cards to sherds in the Collection (rounded to one decimal point).

Table. 3.4. Isolated figures from Table 3.3 comparing the number of sherds listed on the card index with those in the Ashmolean Collection for the four classes represented in each of the three sample regions.

Table. 3.5. Comparison showing the number of sites particular classes occur in within three different regions, based either on information derived from the card index or from the sherds in the Ashmolean Collection.

Table. 3.6. Sites identified so far that were visited by both Stein and Williamson. The columns on the right give the number of sherds from the respective sites in the Ashmolean portion of the Williamson Collection and the British Museum portion of the Stein Collection.

Table. 3.7. Breakdown of the periodisation used in the analysis below including some of the main ceramic classes used as period type-fossils. For class codes see Appendix I .

Table. 3.8. The number of sites represented in the Tehran collection that have so far been relocated by sources.

Table. 4.1. Ceramic periodisation based on significant changes in the ceramic assemblage as revealed through the Kush/al-Mataf seriation, with approximate dates provided by the same excavations.

Table. 5.1. Periodisation based on the dating of East Asian pottery. Note that only those classes that fall within a single period are listed in the right-hand column.

Table. 5.2. Totals and percentages of inland sites with East Asian pottery by survey region.

Table. I.1. Ceramic classes listed in alphabetical order.

Table. I.2. Ceramic classes listed in thematic order.

Table. III.1. Region prefixes used in the site codes listed in the site catalogue below.

Preface

The Williamson Collection Project

Project origin and background

The Williamson Collection Project grew out of the gradually increasing interest in the Islamic archaeology of Eastern Arabian and the surrounding regions through the late 1980s and 1990s. This itself was inspired by an awareness amongst researchers of the potential contribution that this material – and in particular ceramics – could make to a better understanding of regional Islamic history, but also of the development of Indian Ocean trade. At the same time, such work also brought into focus some of the problems that still needed to be confronted before this material could be used as a reliable source of evidence.

From a personal perspective, I (the second author) was fortunate to be involved in a number of projects around this time that gave me an insight into wider developments. In 1988 and 1989 I participated in the French-Kuwaiti excavations, under Jean-François Salles, at the site of al-Qusur on Failaka Island. Working on the finds from what appears to have been a relatively small domestic residence at the monastic site, I was struck by the variety and quality of the ceramic wares, which included Iraqi turquoise glazed wares, Iraqi unglazed wares such as eggshell wares, stamp-impressed wares and torpedo jars, as well as Iranian wares and numerous imports from South Asia (e.g. Kennet 1991: 101–03). At that time, despite this very abundant ceramic assemblage, al-Qusur could be dated only very approximately to the ‘Sasano-Islamic’ period – about which more below.

Two years later I participated in excavations at the 14th-to 16th-century site of Julfar (al-Mataf) in the emirate of Ras al-Khaimah (UAE) under Geoffrey King. This was part of a larger project focused on the coastal town that was co-ordinated by the National Museum of Ras al-Khaimah and involved teams headed by Claire Hardy-Guilbert (CNRS), Geoffrey King (SOAS), Michael Jansen (Aachen University) and Tatsuo Sasaki (Kanazawa University). As part of the same project, I carried out a surface survey of Jazirat al-Hulaylah, an island to the north of Julfar, where successive phases of occupation from the 8th century AD (not the 1st/3rd AD as then believed) until the 20th century covered much of the island’s surface (Kennet 1994). Dating the changing pattern of settlement was based entirely on the ceramics. In the following year (winter 1992–93) I took on responsibility for the study of the more than 46,000 sherds from Geoffrey King’s excavations at Julfar, which were studied as part of my own PhD thesis (Kennet 2001; 2004: 19–24). This assemblage contained Iranian and Iraqi wares, Chinese Longquan celadons and Jingdezhen porcelains, as well as glazed and unglazed wares from around much of the Indian Ocean.

From these early years of personal initiation into the Islamic archaeology of Eastern Arabia, and in particular into the ceramics, a number of ideas began to emerge that I am sure I shared with other researchers.

First was the extremely wide variety of ceramic wares that were retrieved by excavations and surveys in the region from the Islamic period. These included, of course, local Eastern Arabian wares as well as a range of Iraqi, Iranian, East African, South Asian, South-east Asian and Chinese wares of all periods from the 7th to the 20th centuries. To illustrate this point, during the study of the c.200-year-occupation of Julfar, 47 ceramic wares or classes were defined from the ceramic assemblage of 46,265 sherds (Kennet 2004: 19–24). It is very difficult to compare this accurately to earlier periods because of the different methodologies employed, but it can be noted that a study of 13,910 sherds from the c.400-year-occupation of the Late Bronze Age site of Qarn al-Harf (12 km to the south-east) produced only eight ceramic wares, three locally produced and five imported (de Vreeze forthcoming), demonstrating very clearly that an important feature of the later Islamic period is a vastly increased range of ceramic production locations, and consequentially much expanded distribution systems. Such issues, rarely thought about at this scale, are potentially highly informative about the nature of demand, supply and consumption, and, indeed, of economic and cultural activity generally.

At that time, understanding of the archaeological ceramic material from this region was still relatively limited. Apart from the seminal studies by Whitcomb (1975; 1978; 1985), and Whitehouse’s evaluations of the Siraf material (e.g. Whitehouse 1970; 1973; 1979; 1983), there were only outline studies of the al-Qusur material by Pattitucci & Uggeri (1985), and of Bahrain by Larsen (1983: 271–93). The only detailed study of the later (post-13th century) material was that set out by Hansman in his book on Julfar (Hansman 1985: 25–75). Despite the relatively poorly developed state of the field, it was clear that this was a hugely exciting and informative set of material that had the potential – if properly documented and analysed – to produce significant insights into a range of questions; for example the development of the organisation of manufacturing and manufacturing technology in the Islamic world, as well as the nature and extent of trade and distribution systems, both across the Eastern Arabian/Iranian region, as well as more widely across the Indian Ocean as far as East Africa and China. In addition, it was clear that the chronology of settlement and human activity in the region depended almost entirely on the ceramics, and there were then (and still are now) many unanswered questions.

Second, it was clear that there was no coherent methodology or approach to the study, classification and analysis of the ceramics amongst researchers working in the region. This was brought home during a 1993 meeting between the ceramic specialists from the British, French, German and Japanese teams working at Julfar (mentioned above). The meeting was an attempt to co-ordinate approaches to ceramic collection, description, classification, quantification and analysis. The aim was that the results of the four excavations should be brought together into a broader, more meaningful and insightful interpretation and conclusion built on detailed (ideally quantified) comparisons between assemblages from different parts of the site and from different phases. In the end, despite a long day of detailed (and often quite heated) discussion, such a consensus proved impossible to achieve. There was a general acknowledgement that a common approach would have been a good idea, but there were fundamental differences of opinion about matters as basic as aims and methodology, as well as about specifics such as whether a particular group of sherds should be grouped as a single ware, or as two or even three separate wares. Dates, provenances and terminology also proved difficult to agree on. Some teams appreciated the value and the potential of quantified analysis, whilst others saw such an approach as meaningless, statistical sophistry. Some agreed on the creation of ware groups (without which quantification is effectively impossible), whilst others preferred to present only a selection of individual ‘diagnostic’ sherds, so as to provide an indicative chronology and idea of regional contacts. Although much was learned and many new ideas shared during this meeting, in the end each team went off to ‘plough its own furrow’ (e.g. Hardy-Guilbert 1991; Sasaki 1991; 1993, 1994; Sasaki & Sasaki 1992; Kennet 2004) and it has never been possible to bring the material together into a synthesis (although see Kennet 2003 and now Carter et al. 2020 for attempts to do so). This was a disappointing – even depressing – outcome, which vastly reduced the potential for further work and analysis. If four ceramic specialists working on the same site were not capable of agreeing approaches, then what hope could there be for co-ordination between specialists working on sites as far apart as East Africa, India, the Philippines, Thailand and China (for example)? How would it ever be possible to exploit this extensive dataset and make real progress in our understanding of the period? If basic ideas of classification and quantification could not be agreed on, how would it ever be possible to make meaningful comparisons between assemblages from different parts of the Indian Ocean (for some examples of the potential of such an approach see Kennet 2004: figs 41–57; Priestman, i.2021; ii.2021, and most recently Zhang et al. 2023, as well, of course, as the text of the present volume).

Third was the lack of chronological precision in dating many of the ceramics involved, in particular in relation to the Late Antique and Early Islamic periods. For example, the consensus was that the turquoise glazed wares found in abundance at al-Qusur and other sites were datable to the ‘Sasano-Islamic’ period, based largely on reference to the excavated sequence at Susa (e.g. Kervran 1977; Bernard & Salles 1991). This uncertainty led to major misunderstandings of the extent and nature of Sasanian-period and early Islamic settlement across Eastern Arabia. The same was true of other significant wares, for example the ‘tan chocolate-chip’ ware as defined by Potts et al. (1978: 12) as Sasanian, that is now known to be 18th century and later (Kennet 2004: 59–60), and the dating of Indian red polished ware, once dated 1st to 3rd AD, and which is now widely accepted to circulate in this region predominantly in the 7th and 8th century (e.g. Whitehouse & Williamson 1973: 39). Part of the problem was that revised chronologies produced by excavations and research in Iraq (e.g. Northedge, Bamber & Roaf 1988; Simpson 1992) were not being fed back into the region and this was preventing a more accurate appraisal of the archaeological record. At the same time, too little attention was being paid to clarifying chronology in Eastern Arabia itself.

Fourth was the obvious potential that Islamic period ceramics might have – again, if correctly studied and dated – both to our understanding of the development of settlement during this period and also to our understanding of the trade, distribution and consumption of ceramics. So far as settlement was concerned, important work was being published from Iran (e.g. Christensen 1993) and Iraq (e.g. Adams 1965; 1981), setting out broad agendas of periods of growth and decline, but work along these lines was still very much in its infancy in Eastern Arabia, especially for the earlier periods (e.g. Potts et al. 1978: 13–15; Whitcomb 1975; 1978). Research on ceramics in local trade and maritime Indian Ocean trade was also at a relative stage of under-development, although, e.g. Whitehouse’s (1970; 1973; 1979; 1983), Rougeulle’s (1991a; 1991b; 1996), and Hansman’s (1985) work had begun to develop this theme.

These considerations encouraged me to re-focus the aims of my own PhD research on trying to resolve some of these problems. Fortunately, during the Julfar excavations, Beatrice de Cardi (1916–2016), who had been personally responsible for the discovery of many of Ras al-Khaimah’s archaeological sites, pointed out the existence of a large mound, almost 100 m across, known locally as Kush, which was located close to the village of Lower Shimal (Shimāl Taḥt) about 2 km inland from Julfar. A scattering of what we termed then ‘Sasano-Islamic’ pottery was collected on the surface of the mound. At the time, the significance of this mound – which was obviously an archaeological tell – was not immediately apparent, but as knowledge of the area grew it became clear that such tells are, in fact, a great rarity in Eastern Arabia; based on surface collection, this one looked likely to contain a stratigraphic sequence that would unlock the chronology of the Sasanian and early Islamic periods. So it proved, and the limited but fortuitously located sequence that was excavated between 1994 and 2002 led to a fundamental revision of the way in which we understand these periods (Kennet 2004: 12–16; 2007).

Working on the 42,265 sherds from Geoffrey King’s Julfar excavations, along with 30,398 sherds from the excavations at Kush, as well as 13,600 or so sherds from other soundings and surface collections in Ras al-Khaimah, an attempt was made to map out a preliminary, classified ceramic sequence for the Sasanian to post-medieval period. This became a key plank of my own PhD dissertation and was followed a few years afterwards by the publication of Sasanian and Islamic Pottery from Ras al-Khaimah (Kennet 2004). The hope was that this book would provide a methodological and chronological framework within which researchers could work – or which they could at least challenge. A more precise chronology had been established, a provisional method had been set out and the potential value of a classified and quantified approach had been demonstrated. It is gratifying that the method and approach have been widely (although not universally) adopted and this has begun to generate a better understanding of all aspects of the Islamic period in this region (e.g. Nanji 2011; Priestman 2005a; 2013; Power 2015; Rougeulle 2015; Zhang 2016; Priestman ii.2021; see below).

Nonetheless, the weaknesses and failings of the approach and classification set out in Sasanian and Islamic Pottery from Ras al-Khaimah were clear for all to see. Many ‘classes’ that had been defined were never seen again, whilst other classes, which have since become well established, were not mentioned. In particular, the later part of the period, the period after the abandonment of Julfar in the late 16th century, was poorly represented, despite the fact such material makes up at least 70% of surface collections from the area. It was clear that the analysis had to be taken forwards and the method and classification had to be improved if real progress was to be made.

Having used the Ras al-Khaimah sequences and material to establish the chronological and methodological framework described above, the obvious next step was to turn to the Williamson Collection, on which the present volume is based. This unique collection of 16,909 sherds from 500 to 600 rural sites in southern Iran provided the perfect opportunity with which to test and develop the method, classification, and chronology, but also to apply it at broader geographical scale where questions such as changing settlement density and location, production and distribution systems, engagement in long-distance maritime trade and consumption of imported ‘luxury’ ceramics could be investigated.

I was first introduced to the Williamson Collection on a cold winter’s morning in 1992, thanks to the kindness of Peter Morgan. He showed me through the many of the 150 or so boxes of material and impressed upon me the unique nature of this Collection, which had been assembled from a wide range of sites, many of which probably no longer existed. In the early 2000s the collection was housed at the Ashmolean Museum in Oxford, where it had been since shortly after Andrew Williamson’s untimely death in 1975. Andrew Williamson was a rising star in the Islamic archaeology of this region and would, to judge by what he had already achieved by such a young age, no doubt have become one of the leading academic lights in the field, had his life not been cut so tragically short.

The British Institute for Persian Studies (BIPS), under which Williamson had worked, had some responsibility for the collection and it was to this institution that a request for funding was made to take the next step. Fortuitously, James Allen was at the time the Ashmolean curator responsible, and also had a senior position in BIPS. His kind and enthusiastic support and his appreciation of the value of the material led to BIPS providing funding to support Seth Priestman through an MA by Research focused on the Collection in the Department of Archaeology at Durham University from 2001. A large van was rented and the Williamson Collection and its associated archive was moved to Durham University by Seth Priest-man and me, and work then began on the ‘Williamson Collection Project’.

Project aims

In approaching the Williamson Collection, it was decided that the key aim was to produce a comprehensive catalogue/database and analysis of the whole Collection. Up to that time a number of smaller studies had examined aspects of the material (e.g. Morgan & Leatherby 1987; P. Morgan 1991; Rougeulle 1991a; 1991b; 1996). However, it was keenly appreciated that the collection needed to be approached as a coherent whole and analysed and interpreted as such, rather than as a series of sites, areas, or specific questions. Part of the rationale behind this was the desire to produce a database for general consultation, but the key thinking was to investigate broader patterns of settlement development, location, size and density as well as the degree to which both small inland and small coastal sites engaged in trade and exchange with the mercantile maritime economy of the Indian Ocean throughout the period of study. It was strongly appreciated that the size and scope of Williamson’s work – and its unrepeatable nature – gave it significant and unique academic value when considered as a whole.

Project organisation and funding

The project was kindly and generously funded largely by BIPS, with support from Durham University. I initially approached BIPS about the possibility of their funding a research assistant for a year at Durham in order to assist me in cataloguing and classifying the entire Williamson Collection. The projected cost of around £27,000 was, at that time, too large for BIPS’s budget. An agreement was therefore reached by which BIPS contributed £13,500 to cover a part-time salary for a research assistant whilst Durham University contributed the tuition fees for an MA by Research and Seth Priestman was recruited to the two joint posts – which were in reality a funded MA course. This arrangement complicated the intellectual ownership of the work. Seth Priestman was required to present the results in fulfilment of his MA by Research and this partly explains the complicated authorship arrangement in this book.

Seth Priestman worked tirelessly and brilliantly for well over a year in the Department of Archaeology at Durham and managed to mark, classify and database all 16,909 sherds, to link them to the sites where they were picked up and to analyse the chronological and geographical patterns that they suggested. With the kind help of St John Simpson, this was supported by a Sackler Fellowship at the British Museum in 2003 (Priestman 2003). Seth Priestman submitted his MA by Research thesis in 2005, and his MA was awarded that same year.

BIPS generously provided further funding at later stages, most notably in 2005 to begin the process of editing and adapting the MA in order to make it suitable for publication, and also to cover the cost of a brief fieldtrip to Hormuzgan and Tehran to make certain that the assemblages Williamson had brought back to the UK accurately reflected what was on the ground at a sample of sites. During that visit, the chance was taken by Seth Priestman to inspect the part of Williamson’s material that had been deposited in the Iran National Museum in Tehran.

Two further grants from BIPS in 2021 and 2022, totalling £9500, have allowed the text and images to be finalised and copy-edited, in order to make the book ready for publication. In total, BIPS has kindly contributed £32,000 to this project. Their support is warmly acknowledged.

Project contributions

The contributions of the two authors to the research project and the final production of this book were as follows:

Seth Priestman undertook the day-to-day work, including numbering sherds, error checking, classifying sherds, writing class descriptions, creation and maintenance of the database. He also undertook literature search, further research into various aspects, background reading and analysis of the database. He wrote the book text and produced the figures and plans.

Derek Kennet conceived and set up the project, arranged funding and permissions, provided intellectual direction and strategy, set and devised methods, set aims and structure and provided supervision of all stages, including ceramic classifications, descriptions and data analysis.

Derek Kennet

Conventions and explanatory notes

The present monograph is based on a study of the archaeological survey of Sasanian and Islamic period settlement in southern Iran undertaken by Andrew Williamson (1945–75), with the support of the British Institute of Persian Studies between 1968 and 1971. The volume presents work undertaken on surviving portions of the surface collection material, comprising a collection of almost 17,000 ceramic sherds and associated archival records donated to the Ashmolean Museum by Andrew Williamson’s parents following their son’s untimely death in 1975. These materials were transferred to the Department of Archaeology, Durham University in 2001 and form the basis of a study carried out by Seth Priestman under the supervision of Derek Kennet, with most of the work undertaken between 2001 and 2004 (Priestman 2005a).¹ Further associated work on Williamson’s sites and an additional portion of the Williamson Collection in the National Museum of Iran was undertaken in 2005 (Priestman et al. 2005), the results of which are reported below. All aspects of the current study have been generously supported by the British Institute of Persian Studies.

Pottery classification

An important element of the current monograph (Appendices I, II, IV & Colour plates) concerns the presentation of a classification of the ceramics in the Williamson Collection developed on the basis of the sizeable assemblage of material in the UK, which represents one of the largest collections of its kind outside Iran. There is some complexity in the manner in which the current classification has developed and evolved since 2001. Underpinning the investigation is work undertaken on the classification of pottery from the sites of Kush and al-Mataf in the United Arab Emirates (Kennet 2004). Wherever possible, clear reference is given to this foundational study. The first author’s work has subsequently evolved to take on new ideas and information offered by the wider body of source material under investigation. An initial presentation of this work appears in an earlier text (Priestman 2005a). Further work that builds on the study was subsequently carried out by Priestman on the excavated ceramics from Siraf in the British Museum (Priestman 2007; Priestman & Simpson forthcoming). More recently, material from Kush/al-Mataf, Williamson, Siraf and wider studies across the Indian Ocean have been synthesised and presented in the form of the integrated ‘Indian Ocean Pottery Classification’ or IOPC (Priestman 2013) and in a further update to the work in Ceramic Exchange and the Indian Ocean Economy (Priestman i.2021; ii.2021). As far as possible, the current monograph has been edited and updated to bring the Williamson Collection pottery classification in line with the IOPC. This volume therefore supersedes earlier work on the Williamson Collection (Priestman 2005a). At the same time, certain features, particularly regarding the form classification, are based on older, less well-developed work than that associated with the 2021 publication. The current work is intended as a stand-alone presentation of the Williamson Collection but should also be understood as a body of data that adds to a wider programme of ceramic research that feeds into the IOPC.

Class codes

Within the text, frequent reference is made to the ceramic classes listed in the catalogue (Appendix I), each of which has been assigned a class code made up of an acronym of the full class title. Some of the class codes contain a string composed of two parts separated by a full point, with the first part designating the class group and the second part a particular sub-category. For example, FRIT.IW, where FRIT represents the broad ceramic family and IW stands for ‘incised white’, i.e. incised white frit. Within the text, either the full class title is given or just the first part of the string when the discussion is of the group as a whole. A hash symbol is used to indicate where all sub-class categories are being referred to; for example, GRAF# means all sgraffiato sub-categories. A list of all of the class codes contained within the catalogue and referred to in the text is given at the start of Appendix I.

Referencing

Extensive use is made throughout the study of unpublished documents contained within the Williamson Collection archive (https://zenodo.org/communities/williamson_archive). To guide the reader, all unpublished documents contained within the Williamson Archive are listed at the end of the bibliography. They have first been placed into one of four categories: notebooks, documents, reports and letters. Within each sub-category, items have been arranged by known or inferred date. All items have been assigned a numbered code in order to provide a stable reference. The item numbers are prefixed with WA = Williamson Archive, followed by a letter designating their sub-section allocation (e.g. B = notebook; D = document, R = report; L = letter). Where information from archive sources is referred to throughout the text, the item reference is given, for example, WA-L34 refers to Williamson Archive Letter Number 34. All other citations, including Williamson’s published sources, are given using the standard Harvard referencing system.

Note

1A complete database of all of the pottery contained within the UK portion of the Williamson Collection together with the sites documented during the survey is available for access at: https://doi.org/10.5281/zenodo.7813046

Acknowledgements

First and foremost, I (the first author) wish to thank Derek Kennet who – as explained in the preface – conceived the project, raised support for the study and supervised every stage. I am indebted to Derek for teaching me so much of what I know in archaeology and being such a generous colleague and friend. The Williamson Collection Project could not have been undertaken without the support of the British Institute of Persian Studies. I would like to thank, in particular, James Allen and Vesta Curtis for their generous encouragement. Equally, the project has benefited from the support of the Department of Archaeology, Durham University. Other individuals to whom I am extremely grateful include St John Simpson for reading and suggesting corrections to the text below, as well as for much stimulating discussion and support; Robert Hillenbrand and Rémy Boucharlat for inspiration and encouragement when the project first began; Beatrice de Cardi (1914–2016), Chris Gerrard, David Hinton, James Howard-Johnston, John Robert Jackson, Sarah Jennings (1947–2009), Karl Lamberg-Karlovsky, Peter Morgan, Valeria Fiorani Piacentini, Axelle Rougeulle, Jennifer Scarce, William Sumner (1928–2011) and David Whitehouse (1941–2013) for the important information they supplied as colleagues of Andrew Williamson or on their own work on the Williamson Collection; Paul Luft for his invaluable assistance relocating sites; Regina Krahl and John Guy for help with the classification and spot dating of the East Asian pottery; Alireza Khosrowzadeh and Abofazl Aali for assistance with the Hormuzgan Survey and Robert Carter and Hossein Tofighian for inviting me to participate in the Bushehr Hinterland Survey and the work that we did together.

Other experts who have offered knowledge on the pottery and other aspects of the project include Søren Andersen, Kasturi Bai, José Carvajal López, David Hill, Mark Horton, Peter Magee, Justin Morris, Carl Phillips, Tim Power, Janumala Varaprasada Rao (1955–2006), Andrew Sage, Hanae Sasaki, Tatsuo Sasaki, Roberta Tomber (1954–2022), Oliver Watson, Lloyd Weeks and Michael Willis. In carrying out the work on the Williamson Collection, the project has benefited from the assistance of the following volunteers and paid assistants: Rory Ambrose, Rosey Priestman, Rebecca Tyson, Krysia Truscoe, Heather Stephenson, Maria Chaloglou and Brendan Morton who assisted with the sherd marking; Jess Brown, Jessica Philips and Rachel Cubit who assisted with the sherd re-entry phase; Rachel Cubit who also carried out useful work on the site cards; Graham Galloway who helped to set up a GIS map for the project, and Chris Leach who helped with site relocation. More recently, it has been possible to enhance the available information on the site catalogue, mapping and archives through the support of the Maritime Endangered Archaeology Project (MarEA) at the University of Southampton and especially the contributions of Lucy Blue, Crystal El Safadi and Friederike Juercke. During the production of the book, we have been generously and ably supported by Cameron Petrie, the series editor, and Helen Knox as copy editor. I am grateful to Indy Priestman and Maria Gajewska for corrections on earlier drafts and to the two anonymous referees for their generous feedback and encouragement. The kindness and interest of Andrew Williamson’s family is very much appreciated. Finally, I would like to acknowledge a number of individuals who have been of particular and direct importance during many years of work on the project since 2001. They include Motoko, Nuno, Noi, Andy and Rosey Priestman and in Durham Paula, Sasha and Jaz Lewthwaite, Charlotte Buswell, Bashak Arda, Mark Andrews and Ben Duckworth.

Frontispiece: Survey by Land Rover near to Lake Bakhtegan, Fars province in 1968. Photograph by Andrew Williamson (WA-IN0238).

Chapter 1

Introduction

Between 1968 and 1971, Andrew George Williamson, a doctoral student from Pembroke College, Oxford, carried out one of the most extensive and ambitious archaeological surveys undertaken in southern Iran or the Persian Gulf. Using simple equipment and working on his own or with one other companion, he was able to record over 1200 archaeological sites distributed widely through southern Iran in the provinces of Fars, Kerman, Bushehr and Hormuzgan (WA-R21: 1). Williamson’s survey represents the single most detailed and sustained attempt to investigate the archaeological settlement record across the region of southern Iran and is one of the few examples of large-scale research dealing with Sasanian and Islamic settlement in the region specifically. Williamson’s untimely death in Oman in 1975 prevented the work from being completed and, despite the massive scope and potential of the research, his work has remained relatively unknown since.

A sizeable portion of the finds that Williamson collected during the survey (comprising ceramics as well as smaller quantities of glass, stone and metal) was exported to the UK and deposited with the Ashmolean Museum in Oxford, along with much of the documentation associated with the project. Although various aspects of the Collection have been previously investigated since it was deposited in the Museum (Morgan & Leatherby 1987; P. Morgan 1991; Rougeulle 1991a; 1991b; 1996), a full synthesis of the material has never previously been attempted. As one of the single-largest bodies of Iranian ceramics available for study internationally, and as a substantial and significant piece of archaeological research, the potential of the Williamson Collection is evident. A detailed study of the Williamson Survey and associated finds Collection was initiated at the University of Durham in October 2001 with the support of the British Institute of Persian Studies, which also provided the majority of the funds for the initial research.¹

The principal aims of the Williamson Collection Project were to:

•develop an independent classification of the Williamson Collection with clear links – where they could be established – to the stratified and fully quantified sequences from Kush and al-Mataf in Ras al-Khaimah in the UAE (Kennet 2004);

•register and record all the ceramic finds associated with the survey;

•establish a list of the sites that Williamson recorded and relocate as many of Williamson’s sites as possible;

•use